The money is somewhere else. It is in the percentage that comes off every gift the site collects, and that percentage runs from 2.15% to 3.00% depending on which company you sign with. On a church receiving $250,000 a year online, the gap between the cheapest and the most expensive published route is $2,465 a year, which is more than ten years of that $19 website. This article is built from the prices and rates these companies print on their own pages, read on 31 August 2026, plus the arithmetic those numbers force. We build websites for churches and nonprofits ourselves, so our own figures are in here alongside everyone else's.

The 30-second answer

A church website costs between nothing and about $190 a month, and the number is smaller than almost every guide claims. Planning Center includes an app and website at no charge. Tithely charges $19 a month, SolaSites $50 plus $500 once, Church Web Works from $58, Ekklesia 360 from $70, ChurchSpring from $71, REACHRIGHT $97, and Nucleus $99. A fully custom build is quoted per project and REACHRIGHT publishes a $2,500 deposit as the entry point. But the subscription is the small number. The giving platform attached to that site takes 2.15% to 3.00% of every card gift plus 30 to 45 cents per transaction, and on $250,000 of online giving that is between $3,763 and $6,228 a year. Choose the rate before you choose the template.

The guides answer in a range, their own pricing pages do not

Every guide currently ranking for this question answers with a range so wide it cannot be wrong, and three of them print an exact price elsewhere on their own sites. On 31 August 2026 we opened five of the pages returned for this query. One of the five, Kingdom Web Services, returned nothing our reader could parse. The other four all lead with a span.

SolaSites, in a post dated 14 April 2026, opens with "$0 to $40,000+" and then calls that range unhelpfully broad itself. REACHRIGHT says "$0 to $20,000+". ChurchCreation also says $0 to $20,000, then usefully splits it into three routes at $0 to $300 a year, $500 to $1,800 a year, and $3,000 to $20,000 up front. It is also the only one of the four to put a number on the giving fee, quoting roughly 2.9% plus 30 cents per card gift and about $3,000 a year on $100,000 of online giving, though it uses one generic rate rather than any vendor's published one. Ekklesia 360's guide takes a different shape and walks through four imagined churches instead.

Now open the pricing page on two of those same domains. SolaSites charges $500 once and $50 a month. REACHRIGHT charges $97 a month for church web design, with a $2,000 setup fee waived if you commit to twelve months, and quotes fully custom work from a $2,500 deposit. Ekklesia 360, whose guide is on a marketing subdomain, publishes $70, $95 and $190 a month on its main site. Three companies that answered "somewhere between nothing and twenty or forty thousand dollars" were all printing their real numbers two clicks away.

Nobody is hiding anything here. Those pricing pages are public and one click from the guides. What the pattern actually shows is smaller and more useful than a gotcha: a span running from zero to forty thousand dollars is the question asked again in a longer form, and the company writing it already knew its own answer.

Ten published prices for the same thing

Ten companies print a monthly price for a church website. The whole published range fits between zero and $190, which is a narrower band than any of the guides in the previous section prepares you for.

Read it top to bottom and the shape of the market appears. Planning Center gives away the app and website that its members actually use, called Church Center, and sells a content upgrade for $15 or $32 a month if you want custom pages, themes, sermon notes or media channels. Tithely lists a custom church website at $19 a month standalone, or folds it into an All Access bundle at $119 a month with church management, a mobile app, worship planning and text giving. Servant Keeper never prices a website at all: its core tier is $79.99 a month and its complete tier is $159.99, and a website, an app, a content library and live streaming are the things that appear when you cross between them, so the published marginal price of its website is $80 a month.

At the top, Ekklesia 360 runs $70, $95 and $190 a month with a fourth tier quoted on a call, plus $500 once for onboarding, training and setup on a stock theme. Its feature grid is worth reading closely for a reason we will come back to: online giving and church management software are both marked as an additional cost on every single tier, including the $190 one.

Two of these prices carry conditions that change what they mean. SolaSites is $50 a month but $500 to start, so a congregation that stays two years pays $1,700 rather than $1,200. REACHRIGHT is $97 a month with a $2,000 setup fee that disappears if you sign for twelve months, and twelve months of $97 comes to $1,164. The fee being forgiven is larger than the entire year of subscription it is forgiven for, which tells you the $2,000 was never a measurement of the work.

Why the subscription is the small number

The subscription is not the product these companies are selling, and you can tell because most of them will hand you the website for nothing. Planning Center charges zero for Church Center. Tithely's giving tools are $0 a month with no contract. Subsplash Giving is $0 a month. Vanco has a plan called Grow at $0 a month. Givelify has no monthly fee.

Nobody gives away software for the love of it. What all five of those companies are selling is the payment rail underneath the giving button, and they are paid a percentage of everything that crosses it. That percentage is the real price of a church website, because the website exists in order to collect the offering, and the offering is the only large number in the whole arrangement.

This is worth stating plainly because it inverts the usual advice. In most industries the recurring software fee is the thing to negotiate. Here the recurring fee is often zero and the negotiable number is a rate you may never have looked at, applied to a total you already know down to the dollar.

Every rate a church can be charged for a card gift

Ten published card rates sit between 2.15% and 3.00%, and the gift being processed is identical in every case.

The cheapest published rate belongs to Planning Center Giving, at 2.15% plus 30 cents for credit and debit, and its page says every church pays the same rate with no volume tiers, no contracts, no setup fees and no cancellation charges. The most expensive belongs to easyTithe's entry plan at a flat 3.00%. Between them sit Subsplash, Vanco, Tithely and Givelify, all clustered around 2.9%.

Eighty-five basis points does not sound like an argument. Applied to an offering it is one. A congregation receiving $175,000 a year in card gifts pays $3,763 at 2.15% and $5,250 at 3.00%, a difference of $1,488 for a service that is, at the point of the transaction, indistinguishable.

Two vendors will sell you a lower percentage

Two companies publish something unusual and genuinely useful: a ladder where paying more each month buys a smaller cut of every gift. It is the clearest evidence available that the monthly fee and the rate are two prices for one product.

easyTithe prints three plans. Giving Small Church is $19 a month at 3.00% plus 42 cents per card transaction. Giving Small to Mid-Size Church is $49 a month at 2.6% plus 42 cents. easyTithe Pro, which adds church management and a mobile app, is $75 a month at 2.75% plus 42 cents and is capped at 500 records. Bank transfer is 1.00% plus 42 cents on all three. Note the third rung: the most expensive plan does not carry the best card rate, because you are buying software with it rather than a discount.

Vanco does the same thing with two plans. Grow is $0 a month at 2.90% plus 45 cents for cards and 1.00% plus 45 cents for bank transfer. Thrive is $54 a month at 2.65% plus 39 cents for cards and 0.90% plus 39 cents for bank transfer. The $0 plan carries a footnote worth reading: a $10 monthly fee applies if an organization does not complete its first meeting with a Vanco giving specialist within 15 days and get its giving page live within 30. American Express is 3.99% on both plans, a chargeback costs $25, and a returned bank transfer costs $5. Add-ons are priced separately again: text-to-give $10 a month, giving insights $75 a month, email outreach $19.95 a month.

easyTithe attaches a floor to all of its plans as well. If your processing fees come to less than $10 in a month, you pay $10 anyway, which its page explains is roughly $300 of transactions. At 3% that floor is reached at about $333 of card gifts, so a small congregation on the $19 plan is really on a $29 plan until it crosses it.

Where each ladder actually pays off

Both ladders have an exact crossing point, and both are computable from the vendor's own numbers in about a minute.

On easyTithe the sums are unusually clean, because the 42 cents is identical on both plans and cancels out. The $49 plan costs $30 a month more and saves 0.40% on card gifts. Divide $30 by 0.004 and the answer is $7,500. Below $7,500 a month in card giving the $19 plan is genuinely cheaper. Above it the $19 plan is the more expensive of the two, and the further above it you go the worse the gap gets. In annual terms that is $90,000 of card giving.

Vanco is a little messier because the per-transaction charge moves too. Thrive costs $648 a year and saves 0.25% on cards, 0.10% on bank transfers and 6 cents on every transaction of either kind. Taking the 70% card and 30% bank transfer split that Tithely uses in its own calculator footnote, the percentage savings alone come to 0.205% of everything, so $54 a month is repaid at about $26,300 a month of online giving. Add the 6 cents at an average gift of $150 and the threshold falls to roughly $22,000 a month. Either way, Vanco's paid plan is a large-church product wearing a small-church price.

A discount priced on exclusivity

Subsplash publishes two card rates for the same giving product, and the difference between them is not volume, it is loyalty. The standard rate is 2.99% plus 30 cents. The rate labeled exclusive is 2.3% plus 30 cents, and the banner across the top of the page ties it to going exclusive with Subsplash. Bank transfer is 1% with no per-transaction charge on both, and a program it calls GrowCurve is described as dropping card as low as 1.9% as a church grows.

Sixty-nine basis points is the published price of not going exclusive. On $175,000 of annual card giving that is $1,208 a year, which is what it costs a church to keep a second option open. That is a legitimate thing for a vendor to charge for and a legitimate thing for a church to buy. It is not a legitimate thing to sign without noticing, because the cost of leaving is being collected in advance, in installments, out of the offering.

The rate that is a feature bullet with no number

Servant Keeper sells the same idea and does not tell you the price. Its comparison grid lists integrated giving on both tiers, and against the $79.99 core plan the entry reads Standard Processing Rate while against the $159.99 complete plan it reads Preferred Processing Rate. Neither rate appears anywhere on the page. A church reading that grid can see that $80 a month buys a better rate. It cannot see how much better, which means it cannot work out whether the upgrade pays for itself, which is the only question the grid was built to help with.

Pushpay, one of the largest names in church giving, publishes no price of any kind on its pricing page and asks you to book a demo for answers and pricing options. Subsplash publishes rates but quotes its Subsplash One and Enterprise packages on church size and usage. That is three well-known vendors, on three pricing pages, where the number that decides the decision is absent.

Bank transfer is the line nobody markets

The single largest saving available to a church in this article is not a discount, it is a payment method. Card rates cluster between 2.15% and 3.00%. Bank transfer, the same gift moved by ACH instead of a card, is priced at 0% plus 30 cents at Planning Center, 0.90% plus 39 cents on Vanco's paid plan, and 1% at Tithely, Subsplash, easyTithe and Vanco's free one.

Put a number on that. A church receiving $75,000 a year by bank transfer pays $750 at 1% and nothing at all at 0%, before the flat per-gift charge. Move another $50,000 of card giving over to bank transfer at Planning Center's rates and the saving is $1,075 a year. There is no plan to upgrade to and no contract to sign. Somebody has to put the option in front of givers and explain it, which is a design job on the giving page and a communication job from the front.

Tithely publishes the effect of the same idea in a different form. Its rate page works an example all the way through: a $100 gift by card deposits $96.80, by American Express $96.20, and by bank transfer $98.70. Nearly two dollars in every hundred, decided by which button somebody taps.

One offering, eight routes, one year

Here is what a single year costs a single church, priced eight ways, using nothing but published numbers.

The assumption on the split is not ours. Tithely's own savings calculator footnote describes a typical church as sending 30% of online donations by bank transfer and 70% by card, so that is the mix we used. The subscription column is the plan price times twelve. The processing column is each vendor's published percentage applied to that mix. Nobody's flat per-transaction charge is in there, because it depends on how many separate gifts the $250,000 arrives in, and that is a fact about your congregation rather than about the vendor.

Three things fall out of the chart. The first is the spread: $3,763 at the bottom and $6,228 at the top, a difference of $2,465 a year on identical giving. The second is that the subscription is barely visible. The largest black segment in the chart is Vanco's $648, and it is dwarfed by every processing bar next to it. The third is that Vanco's paid plan, at this volume, costs more in total than Vanco's free plan, which is exactly what the break-even said would happen below $26,300 a month.

None of these are identical products and the chart would be dishonest if it implied they were. Planning Center's row buys the app, the website and a giving platform with no subscription at all, which is remarkable value and also means you are living inside one company's ecosystem. Tithely's row includes a website, and both easyTithe rows include a giving platform with no website attached. What the chart does show is that whatever else you are choosing between, you are choosing a rate, and the rate outweighs everything else on the invoice.

The fee can be moved onto the giver

The large platforms all now offer to ask the giver to cover the processing fee, and enough givers say yes that it changes the arithmetic materially. Tithely calls the feature Cover the Fees and states that churches typically see around 60% of donors cover their giving fees, and that the feature saved churches over $5 million in the last year. Subsplash's calculator uses a more conservative 30% as its average and defaults to $40,000 a month of donations, with a slider running from $5,000 to $150,000.

If 60% of givers cover the fee, a 2.9% headline rate behaves like roughly 1.16% for the church, which would move our $250,000 congregation from $6,053 to about $2,558. That is a large enough effect that it belongs in the decision, and it is also the reason to be careful with vendor savings calculators: a platform quoting you an effective rate has usually applied an adoption assumption you did not make.

There is a judgment underneath the arithmetic that is not ours to make. Adding a checkbox that asks a person to pay $2.90 on top of a $100 gift is a pastoral decision before it is a financial one, and some congregations will decide the friction is not worth it. What we would say is that the choice should be made deliberately rather than inherited from a default, and that whichever way it goes, the giving page should say plainly what the fee is and who is paying it.

What the IRS requires your giving page to send

Two federal rules turn your donation receipt from a nicety into a document with required contents, and both of them are build specifications rather than legal theory. They apply to the automated email your giving platform sends, which means somebody has to check that it actually says the right things.

The first covers any single gift of $250 or more. The IRS states that the written acknowledgment substantiating it must carry the name of the organization, the amount of the cash contribution, a description but not a value of any non-cash contribution, a statement that no goods or services were provided if that is the case, a description and good faith estimate of the value of any that were, and a statement that any goods or services consisted entirely of intangible religious benefits if that was the case. That last clause is written for churches specifically. On the timing, the IRS says the acknowledgment is contemporaneous if the donor obtains it no later than the date they file their return for the year of the gift, and it notes that the acknowledgment and the bank-record communication can be the same document.

The second covers what the IRS calls a quid pro quo contribution, meaning a payment made partly as a gift and partly for goods or services. Above $75 the organization must provide a written disclosure statement, and the threshold is on the whole payment rather than the deductible part. The example the IRS gives is a $100 payment that returns a $40 concert ticket: the deductible portion is $60, which is under $75, and the disclosure is still required because the payment was over it.

The disclosure must tell the donor that the deductible amount is limited to the excess of what they paid over the fair market value of what they received, and give a good faith estimate of that value. It can be sent with the solicitation or with the receipt. There are exceptions, and one of them matters here: no disclosure is required where the only thing provided is an intangible religious benefit, which the IRS defines as something generally not sold commercially outside the donative context. Admission to a religious ceremony that has no admission charge is the example given. Tuition, travel services and consumer goods are explicitly not intangible religious benefits.

The penalty is small but real, and it scales with the mailing rather than the gift. The IRS imposes $10 per contribution for failing to make a required quid pro quo disclosure, capped at $5,000 per fundraising event or mailing, with an out for reasonable cause. A church running a banquet, a concert or an auction through its website is squarely inside this, and the fix is a receipt template that can carry a fair market value line rather than a blanket thank-you.

What that gap buys in staff time

A $2,465 annual difference is abstract until you set it against a wage. The Bureau of Labor Statistics county employment series for the fourth quarter of 2025 covers 26,371 religious organizations that report to the unemployment insurance system, employing 206,753 people, at an average weekly wage of $797. That is about $41,444 a year for a full-time role.

Measured that way, the $2,465 spread between the cheapest and most expensive published route is a little over three weeks of one staff wage. Ekklesia 360's $190 a month, the top of the website range in this article, comes to $2,280 a year, or just under three weeks of the same wage. Both are real money to a congregation that counts it. Only one of the two is the number every guide points at.

It is also worth saying what these figures are not. That series counts organizations that file unemployment insurance reports, so it is a floor on the sector rather than a census of American congregations. We are using it for the wage, which is the part it measures well.

Where church giving actually sits

Religion is the largest single destination for American charitable giving and it is not growing. Giving USA 2026, published by the Lilly Family School of Philanthropy at Indiana University, puts total US charitable giving at $617.20 billion in 2025, up 5.7% in current dollars and 3.0% after inflation. Religion took $151.58 billion of that, just under a quarter of everything given, and more than any other category.

The growth line is the part that matters for this decision. Giving to religion rose 2.4% in current dollars and fell 0.2% once inflation is taken out, while giving to education, human services, health, arts and the environment all reached new inflation-adjusted highs. In a flat sector, a percentage point of processing is not a rounding error. It is one of the few levers a church can pull that does not require anybody to give more.

What we charge, and when a platform beats us

Having spent an article inside other people's rate sheets, here are our numbers. A build with us is $1,499, billed a single time, as a flat package rather than an opening position, while a design and development partnership runs at $2,499 a month for as long as it is earning its place. Two weeks to launch is the pattern we see on a focused build across 200 plus shipped projects, offered as a typical rather than a guarantee, and anyone asking about partnership hears back within 48 hours. Scope gets agreed on a call, and the offer page sets out what sits inside each tier.

The giving platform sits outside that price, deliberately, and the reason is worth stating precisely. None of the companies named in this article pays us anything, and we neither sell nor resell a giving product. Online giving wired into a church management database is bought from Planning Center, Tithely, Subsplash, Vanco or easyTithe at the rates printed above, on an invoice with our name nowhere on it. Keeping those two purchases apart is what lets our number stay flat, because a builder who is paid a share of the offering has an interest in which button a giver taps.

Plenty of churches should not hire us, and it is cheaper for everyone if we say which. A congregation whose entire annual technology budget is a few hundred dollars should take Planning Center's free app and website, or Tithely at $19 a month, and spend nothing with us. A church that wants the website, the app, the database, the check-in kiosk and the giving on one invoice from one company should buy a platform, because that integration is genuinely valuable and we do not build church management software. A church in the middle of a capital campaign should fix the giving rate first and come back to the website afterwards. Where we earn our place is a congregation that has outgrown a template, needs the site to do a specific job well, and wants to own what gets built. Our work in this area sits alongside the rest of our nonprofit practice.

Reprice last month

Every worksheet in this genre asks you to forecast. This one asks you to look backwards, because you already have the only data that settles the question and it is more reliable than any projection.

Pull last month's giving report out of whatever you currently use. You need five numbers from it.

Now price that same month at every vendor you are considering, using the published rates in this article. Card amount times the card rate, plus bank transfer amount times the bank transfer rate, plus the gift count times the per-transaction charge, plus one month of the subscription. You will get a real number for a real month rather than an estimate for a hypothetical one.

Four rules keep the exercise honest. Use the rate you would actually get, not the exclusive or preferred one, unless the condition attached to it is written down and you intend to meet it. Put the flat per-gift charge in, because a congregation making many small gifts is affected by 30 cents far more than by 15 basis points. Do not credit yourself with fees covered by givers unless you already have that feature switched on and know your real adoption rate. And run line 5 against your own calculation on your current vendor first: if the two do not match, the rate you think you are paying is not the rate you are paying, and that is the finding.

If you only change one thing this month

Make bank transfer easy and obvious on the giving page. It is the only change on this list that costs nothing, requires no new contract, and can be worth four figures a year on its own. At Planning Center it is 0% plus 30 cents. Almost everywhere else it is 1%, against roughly 2.9% for a card. If your own giving form asks for a card number before it offers a bank option, that default is quietly routing your largest recurring gifts down the most expensive road.

After that, in order: work out your real effective rate using the exercise above; get any preferred or exclusive rate written into the agreement with its conditions stated; check that your receipt email can carry a fair market value line before the next fundraising event; and only then argue about the template. If you are also weighing who should hold the domain and the accounts when all this is set up, we wrote about that in who owns your website, and the mechanics of the donation page itself are covered in our piece on nonprofit donation pages.

The website is not the expensive part. It has not been the expensive part for years. The expensive part is the percentage attached to it, and unlike the template, it is a number you can change this week.

That figure does not appear on any page ranking for this question. What appears instead is a range, usually $3,000 to $8,000, sometimes $500 to $25,000, always uncited. This article is built the other way round: from the prices vendors print on their own pages, read on 29 August 2026, plus the arithmetic those prices imply. We sell websites too, so our numbers are in here with everyone else's.

The 30-second answer

Almost nobody in this market sells an auto repair shop a website as a project. Every vendor we found that publishes a price sells a subscription, and those published prices run from $49 a month to $699 a month. At the bottom, Cannone Marketing charges $199 to set up and $49 a month with no contract. In the middle, Tamarack Media Group starts at $100 a month billed month to month, and Shop Boss folds a website into the third tier of its shop management software, $100 a month above the tier below it. At the top, Shop Marketing Pros charges $499, $599 or $699 a month on an 18 month agreement, then $199 a month once the site is paid off. Back the build price out of that last one and it is $5,400, $7,200 or $9,000 depending on page count. Over three years the published routes cost between $1,963 and $16,164. Our own price is $1,499 for a build or $2,499 a month for partnership.

The numbers at the top of Google

They disagree by a factor of fifty, and not one of them cites a vendor's published price. On 29 August 2026 we opened the guides returned for this query and wrote down the headline number each one leads with. Qrolic splits the market into three bands, $500 to $2,500 for do it yourself, $3,000 to $8,000 for a professional build and $10,000 to $25,000 and up for multiple locations, then summarizes the whole thing as $500 to $25,000. XSOne Consultants puts the average at $2,000 to $5,000 and itemizes it down to $300 to $500 for making the site work on a phone. FreshySites says $2,500 to $7,000.

A shop owner who reads all three learns that a website costs somewhere between five hundred dollars and twenty five thousand. That is the question restated, not answered. And the shape of the answer is wrong in a way the size of the range hides: every one of those figures prices a project, a thing you buy once. In this trade the working parts of a site, the booking, the inspection links, the review requests, the payment page, are billed monthly by whoever owns your repair orders. Stop counting at the build and you have stopped before the bill that never ends.

Everyone who publishes a price sells a subscription

We could not find a single company selling auto repair websites that publishes a one time build price. We did find five that publish a monthly one, and the spread between them is fourteen to one.

Cannone Marketing is the floor: $199 to set up and $49 a month after that, cancel any time, with a custom homepage demo built before you pay. Its own page explains the price by pointing at what it does not carry, no office, no sales floor, no account managers, and it names the going rate it is undercutting as $3,000 to $5,000 up front plus $150 to $400 a month. Tamarack Media Group starts at $100 a month, month to month, no long term contract, covering design, hosting, updates and local search work, and also builds a working demo before payment. Two vendors, both offering to show you the site before you spend anything, which tells you something about how much trust there is to win in this market.

The three larger numbers all come from companies that sell auto repair shops something else as well. That is the pattern worth understanding, and it is where the real money is.

The build price you can back out of a published rate

Shop Marketing Pros publishes three website tiers and, next to each one, the price after the agreement ends. Essential is $499 a month for 10 custom written pages. Pro is $599 for 20 pages. Top-Tier is $699 for 30. All three carry an 18 month agreement, and all three fall to $199 a month afterwards. The feature list above the cards says the site is 100% yours after completing 18 months of payments.

Subtract and the structure falls out. The part of the bill that stops after 18 months is $300, $400 and $500 a month. Multiply by 18 and you get $5,400, $7,200 and $9,000. That is the build, financed. The $199 that continues is hosting, maintenance, monthly edits and analytics reporting, which is roughly what the rest of the industry charges for the same job and close to what we found when we looked at what website maintenance actually pays for.

The same $500 a month appears in the company's four bundled packages, which all include what it calls a Top-Tier website. Website with a Plan is $1,598 a month and then $1,098. Car Count Booster is $2,497 and then $1,997. The Top-Tier is $3,396 and then $2,896. Legacy Builder is $5,295 and then $4,795. Four packages, four different totals, one identical $500 drop at month 19. The page is not perfectly consistent about the after-term figures, printing $1,098 in one block and $1,099 in another for the same package, and $2,896 and $2,899 for another, which is the sort of thing that happens when a number is maintained by hand in two places.

Two things follow from this that matter more than the price itself. The first is that a monthly website bill with a fixed term is a financing arrangement, and financing has a term you can be inside. The second is the ownership line. If the site is yours after 18 months of payments, the natural reading is that it is not yours in month nine, which is the exact question we spent a whole article on in who owns your website. Ask what happens if you leave early, and get the answer in writing before you sign, not after.

At two vendors the website is a software tier

The cheapest website in this market with a real published price is not sold by a web company at all. It is a $100 a month step up inside shop management software.

Shop Boss sells four tiers. Gold is $199 a month, Platinum is $349, Premier is $449 and Premier Plus is $649, all month to month, with annual billing at $179, $314, $404 and $584. What Premier adds over Platinum, and the only thing it adds beyond one to one training, is a responsive website, website design, website maintenance, monthly website reporting and a quarterly website review. So a Shop Boss customer already paying $349 can have a website designed, built, hosted and maintained for another $100 a month, or another $90 on annual billing. Whether that site is any good is a separate question. The price is not a range and it is not a quote.

Tekmetric does the same thing one layer out. Its plans are $199, $349 and $439 a month with no website in any of them, and its Marketing add-on is $345 a month per shop on top of a plan. That add-on is where the site lives: Tekmetric's own websites page describes a search optimized site with online booking built in, appointments landing straight in the shop calendar, plus reminders, campaigns and review collection. The cheapest published Tekmetric configuration that includes a website is therefore $199 plus $345, or $544 a month.

Neither AutoLeap nor Shopmonkey lists a website in any tier of its published comparison. AutoLeap does sell an AI receptionist for $99 a month, which answers the phone rather than the site. Torque360 sells a $999 appointment scheduler add-on that puts booking on a site you already own.

Three vendors, one price list

Once you start reading shop management pricing pages, the striking thing is not that they hide numbers. It is how closely the ones that publish agree.

Tekmetric, AutoLeap and Shop Boss all charge $199 a month for their entry tier and $349 for the tier above it, to the dollar. At the top they part company by ten dollars: Tekmetric's Scale is $439 and AutoLeap's Elite and Shop Boss's Premier are both $449. All three discount annual billing to $179 and $309 or $314. Shopmonkey sits about twenty percent above the pack at $239, $399 and $499. R.O. Writer publishes $199 a month for its Essentials tier and $299 for the tier above, both described as first year prices.

We are not suggesting anything improper. Three companies converging on the same round numbers is what usually happens when a market is competitive, visible and mature, and every one of these firms can see the others' pages as easily as we can. The useful consequence for a shop owner is that the monthly software number is close to fixed. If three vendors want $349 for the middle tier, negotiating that line is not where your effort belongs. Where they differ is in what the tier contains, and that is where a website quote either makes sense or does not.

What the ladder does not include

Three things sit outside the headline price at most of these vendors, and all three are the sort of number that shows up on an invoice months after the decision.

Seats. Shop Boss, Tekmetric and AutoLeap all advertise unlimited users and unlimited repair orders. Shopmonkey does not: Basic includes three user licenses, Clever four and Genius five, with each additional user $20 a month. A seven person shop on Genius pays $499 plus $40, so $539 a month against Shop Boss's $449 for the equivalent tier. The gap widens every time you hire.

Add-ons. Tekmetric charges $70 a month per shop for multi-shop management and $39 a month per shop for its tire suite, both on top of a plan. Shop Boss sells its CRM and online service scheduler as a $249 a month add-on, or you take Premier Plus at $649 and get them bundled. Shopmonkey's card reader is $249 plus $10 a month per device, and its online and in person payment rates are not published at all.

Getting your history out. Shop Boss publishes what it costs to bring data in: $399 to transfer customers and vehicles, $899 to transfer customers, vehicles, repair history and inventory, against more than 130 built in transfer routes. Read that as the price of leaving as well as arriving, because the shop you switch to will quote you the same kind of number. It is also the reason the software choice and the website choice should not be made on the same afternoon by the same salesperson.

What three years actually costs

Monthly prices are hard to compare because the terms are different. Here is every published route in this article priced over 36 months, using each vendor's own monthly figures and nothing else.

These are not equivalent products and the table would be dishonest if it pretended otherwise. The Tekmetric and Shop Marketing Pros rows carry search work, campaigns, reporting and, in Tekmetric's case, online booking wired into the appointment book. The $49 and $100 rows are a site, hosting and edits. Our $1,499 row is a build and does not include hosting, which we do not bill for and which a shop can buy for a few dollars a month.

What the table does show is that the decision is not four figures against five figures, it is roughly $2,000 against roughly $16,000 over the same three years, and the difference is mostly marketing services rather than the website. That is a completely reasonable thing to buy. It is not a reasonable thing to buy by accident because it arrived inside a website quote.

The word free on your website is regulated

Nearly every auto repair site in America offers something free: a brake inspection, a check engine light scan, a courtesy shuttle, a quote. The Federal Trade Commission has a published guide about exactly that word, it has existed since 1971, and it contains hard numbers almost nobody in this trade has read.

16 CFR 251.1 sets out how a free offer has to work. Subsection (c) says every term, condition and obligation attached to the free item should appear clearly and conspicuously at the outset of the offer, in close conjunction with it, and it names the workaround it will not accept: putting the conditions in a footnote reached by an asterisk is not disclosure at the outset. That is a design instruction. A hero band reading FREE BRAKE INSPECTION with an asterisk pointing at small print in the footer is the exact pattern the guide rules out.

Subsection (h) is the one that surprises people.

A single kind of service should not be advertised with a free offer in the same trade area for more than six months in any twelve, at least thirty days should pass before the next one, and there should be no more than three such offers in a year in that area. A permanent free inspection banner, sitting on the same page for three years, is not what the guide contemplates. Subsection (i) closes the obvious escape route: swapping in gift, bonus or given without charge does not fix an offer that does not meet the rest of the section. Subsection (b)(2) adds that a regular price only counts as regular if you have actually been selling at it, openly and actively, for something like thirty days.

Two honest caveats. This is a guide, not a rule carrying its own penalties, and it interprets the deception provisions of the FTC Act rather than standing alone. And enforcement in this area lands far more often on national advertisers than on a five bay shop. But the guide is what a regulator, a competitor or a state attorney general reaches for, and the fix is free: put the conditions next to the offer, in the same type, on the same screen, and rotate the offer instead of leaving it up forever.

The warranty sentence on almost every shop site

Somewhere on most independent shop websites is a line saying that using them will not void your factory warranty. The claim is broadly right, it is worth making, and the usual wording overstates it.

The law behind it is the Magnuson-Moss Warranty Act. Section 2302(c) stops a warrantor conditioning a written or implied warranty on the consumer using a branded article or service, unless that article or service is provided free under the warranty or the FTC has granted a waiver. The FTC's own guide for businesses gives worked examples of both. Requiring a specific brand of filter to keep a warranty alive is prohibited. Disclaiming coverage for damage that a non-authorized third party actually caused is permitted, and the guide prints a model sentence doing exactly that.

The consumer facing version on the FTC site is blunter: a dealer cannot deny warranty coverage because someone else did the routine maintenance or repairs, and using aftermarket or recycled parts does not end the warranty. But the same page carries the exceptions. If the warranty says the work will be done at no charge, the manufacturer can choose the facility. If it says the parts are free, the manufacturer can choose the parts.

So the accurate sentence for a shop website is narrower than the one most shops use. Servicing here does not void a factory warranty, and we document every job so the record stands up if a claim is ever questioned. That is defensible. We guarantee your warranty stays intact no matter what is not, because the shop does not control the warrantor's response to damage it believes the shop caused. While you are rewriting that paragraph, note the FTC's flat statement elsewhere that there is no standard warranty on repairs, which means your own repair warranty is a genuine differentiator and belongs on the page in writing.

What an online estimate legally is

If your site has a get a quote or instant estimate widget, find out what your state calls the thing it produces before you build it. In California it is a regulated document.

Business and Professions Code 9884.9 requires a registered automotive repair dealer to give the customer a written estimated price for the labor and parts a specific job needs. No work may be done and no charges may accrue before the customer authorizes it, and nothing above the estimate may be charged without consent obtained after the shortfall is discovered and before the extra work happens. Consent to an increase can arrive by email, which is the section quietly acknowledging that this conversation now happens on a phone. Body and collision work has to be itemized further, with each crash part identified as original equipment or aftermarket.

There is an exception worth designing around. Under subsection (e) a written estimate is not required for defined preventative maintenance if the service is free or if its total price is displayed conspicuously to the customer at the facility. A published, honest price list is the thing that removes paperwork, not the thing that creates it.

The advertising side sits in 9884.7, which lets the director discipline a registration for making any statement, written or spoken, that is untrue or misleading, or that reasonable care should have shown to be untrue or misleading, and separately for false promises likely to induce a customer to authorize work. Your website is written, it is a statement, and everything on it is covered. Section 9884.6 makes operating without a valid registration unlawful in the first place, which is why the registration number belongs in the footer rather than nowhere.

California is one state, and we are not going to publish law we have not read for the other forty nine. The transferable point is that in this trade the estimate is a legal instrument, so the estimate flow is the highest risk thing on the site and the one to design last, carefully, with your own board's current rules open.

Set against a shop's payroll

The website bill is small compared with the payroll it is supposed to keep busy, and the arithmetic is worth doing once so the argument moves off the quote.

In the fourth quarter of 2025 there were 86,552 private establishments in general automotive repair in the United States, employing 397,377 people at an average weekly wage of $1,227, according to the Quarterly Census of Employment and Wages. Widen to automotive repair and maintenance as a whole and it is 176,403 establishments, 1,023,816 employees and $15.0 billion of wages in that single quarter. The average general repair shop in those figures has between four and five people on the payroll.

At $1,227 a week the average worker in that industry costs about $5,317 a month in wages alone. The most expensive published website subscription in this article, $699 a month, is 13 percent of that. The cheapest, $49, is under one percent. The $9,000 build price we backed out of the 18 month term is a little over two months of one employee. None of that is an argument for spending carelessly. It is an argument that the interesting question is whether the site fills bays, not whether the quote says $4,000 or $6,000.

The two prices we publish

Having spent an article on other people's pricing pages, here is ours. Our build is $1,499, charged once, a flat package price and not an opening bid. Design and development partnership runs $2,499 a month for as long as it is useful. A focused build usually reaches launch around the two week mark, which is a pattern across the 200 plus projects we have shipped rather than a promise, and partnership inquiries get an answer inside 48 hours.

What is not in that number is the software. We do not sell or resell shop management systems and we take no margin on any of them. If you want online booking that writes into your repair order system, that capability is sold by Shop Boss, Tekmetric or Torque360 at their published rates and the invoice has our name nowhere on it. We think that split is the right way round, and it is also why our number can be flat.

Some shops should not hire us, and it is cheaper for everyone if we name them here. Anyone who wants the website, the ads, the reviews and the repair orders arriving on one invoice should buy from a shop management platform or a full service automotive marketing firm rather than from us. If your monthly budget for all marketing is under a hundred dollars, one of the $49 and $100 subscriptions above will serve you better than we will. Shops that need cars this month more than they need a better site should put the money into Local Services Ads and come back to the website later. Our work with trades sits on the local services side of our practice, and the full price list is on the pricing page.

Back out the build price before you sign

Every monthly website offer with a fixed term contains a hidden build price, and you can extract it in about two minutes with the numbers the vendor already gave you. Do it before you sign, because the answer changes how the deal reads.

Fill this in from the quote in front of you, not from memory.

Then ask four questions, in this order. What happens to line 5 if I cancel in month nine, do I owe the balance, and who owns the site at that moment. Is line 2 guaranteed in writing, or does the price simply revert to line 1 unless I remember to ask. What is included in line 2, hosting and edits and reporting, or hosting alone. And who holds the domain registration, the hosting login and the analytics property, because those three are the difference between owning a website and renting one.

If the vendor will not put line 2 and its start date in the agreement, line 5 is not a build price, it is just a bigger subscription with a nicer story. If they will, you now have the number every guide on this subject declined to publish, for your own quote, from your own vendor. Run the same worksheet against a flat build plus a hosting bill and the two are finally comparable.

What to pay for first

Spend on what a driver with a warning light on the dash needs at 7am, and defer the rest until those work. For a repair shop the order is unusually clear.

First, the facts, correct and structured: hours including the ones that differ, address and which entrance to use, phone number that rings in the bay area, what you do and do not work on, whether there is a shuttle or a loaner, and what happens if the car will not start. This is the cheapest thing on the list and it is what both Google and an AI assistant read when someone asks which shop near them can look at a transmission today.

Second, the one service you actually want more of, on its own page, in the words customers use rather than the ones on the repair order. Diagnostics, brakes, fleet work, European, whatever the shop is built around. Write the price policy on that page, including the estimate process, because that is the paragraph that turns a nervous first time caller into a booked appointment.

Third, a way to get in touch that matches what you can actually deliver. Real booking writes into your repair order system and is sold by your software vendor at $249 or $345 a month or bundled into a $649 tier. If that is not in the budget this year, say so and build an honest request form that a person answers within the hour. A fake booking button that quietly emails the front desk is worse than no button, because the customer thinks they have an appointment.

Fourth, your offers, written to the rules above. Conditions next to the offer in the same type, an accurate warranty sentence, an estimate flow that matches your state's law, and your registration number where an inspector could find it. Everything after that, the photography, the video of the bay, the team page, is worth real money once these four work and worth nothing while the phone number is wrong. Shops in adjacent trades face an almost identical bill structure, which we broke down for the heating and cooling business in what an HVAC company website costs.

That is not three unlucky pages. On 28 August 2026 we tried the pricing page of 39 companies that sell United States dental practices a website, practice management software or patient communication software. Nineteen returned a 404 or a 403. Of the 20 that opened, five named a price for the product the page was selling, and not one of the five was a website company. Every cost guide ranking for this question is therefore built on numbers nobody published. This one is built on the numbers somebody did, and they turn out to belong to the software your website has to plug into. We sell websites as well, so our own figures are in here with everyone else's.

The 30-second answer

A dentist website in 2026 is two purchases that arrive as one bill, and only one of them has a public price. The build is quoted privately: of the eight dental website companies whose pricing pages we could open, not one named a price for a website, and the guides ranking for this query fill that silence with ranges running from $499 to $50,000. The recurring half is published and checkable. Open Dental charges $199 a month per location for the first 12 months and $149 after that, plus $75 a month for the online scheduling that makes a booking button work. Practice-Web charges $179 then $129, with online scheduling at $119 and online forms at $49. tab32 is $125 a month in year one and $225 from year two. CareStack folds online appointments, forms and the patient portal into $829 a month. We charge $1,499 for a build or $2,499 a month for partnership. Budget for the build once, then for the subscriptions that make it function, every month, for as long as the practice exists.

What the pages ranking for this question say

They say seven different things, and the disagreement is not small. On 28 August 2026 we opened seven guides returned for this query and recorded the headline range each one leads with. Rosemont Media says $3,500 to $30,000. Emizentech says $2,500 to $20,000. Flamingo Agency says $4,000 to $8,000 for most small practices, rising to $8,000 to $20,000. The Dental Signal says $3,000 to $15,000. Bemo Design splits into $2,500 to $6,000, $6,000 to $15,000 and $15,000 to $40,000. Dentive lands on $5,000 to $10,000. Top Dental Websites opens with $499 to $50,000.

Line those up and the low ends vary by a factor of ten and the high ends by a factor of six. A reader who visits all seven learns that a dental website costs somewhere between five hundred dollars and fifty thousand, which is the question restated rather than answered. None of the seven cites a dental website company's published price, for the good reason that there are almost none to cite.

There is a second problem underneath the first. Every one of those ranges prices a project, as though the website were a thing you buy once and own. For a dental practice the working parts of the site, the booking, the forms, the reminders, the payment link, are almost always billed by a different company on a monthly subscription tied to the practice management software. A range that stops at the build has stopped counting before the recurring bill starts.

We tried 39 pricing pages

The result was lopsided enough to be the answer to the whole question. We took 39 companies that sell dental practices a website, practice management software or patient communication software, and opened the page each one calls Pricing. Nineteen of the 39 never returned a readable page. Thirteen gave us Page Not Found: Smile Marketing, WEO Media, Delmain, Wonderist Agency, Studio 8e8, Pain-Free Dental Marketing, Doctor Genius, Great Dental Websites, Sesame Communications, Dentist Identity, RevenueWell, Dental Intelligence and Dentrix Ascend. Four blocked us with a 403: PBHS, DentiMax, Firegang and Titan Web Agency. Two, TNT Dental and Rosemont Media, timed out.

Of the 20 pages that did open, five stated a price for the product on sale. All five sell software. None sells websites.

Some of the near misses are worth naming, because they show how the silence is dressed. Officite's page does carry a price table, for phone lines, at $24.72 a line a month falling to $18.54 at higher line counts, with no figure anywhere for the website the page is selling. Practice Cafe's pricing page contains two dollar amounts, both rhetorical, in a passage explaining that it cannot compete with cheap logos. Curve Dental's and Podium's pages price nothing and quote customers describing what they saved or earned. Adit's Pricing Plans page carries three dollar amounts and all three are sign-up incentives, including a gift card. Dental Marketo is the one page that does print figures, for advertising, search, social and reputation retainers, though it shows what look like annual totals in a per month slot and names no price for a website at all.

ProSites is the clearest example, because the page is genuinely useful right up to the point where a number would go. It sets out four tiers, Get Started, Competitive Edge, Market Leader and Market Owner, and compares them feature by feature so you can see exactly what moving up a tier buys. Thirteen questions at the foot of the page help you choose between them. What none of it tells you is what any tier costs.

We are not claiming anything sinister. Agency pricing genuinely varies with scope, and a quote-only model is a normal way to sell custom work. The consequence is still real: when nobody in a market publishes, every guide written about that market has to invent, and the numbers at the top of Google for this query are inventions.

The prices you can check belong to the software

Two dental software companies publish complete, itemized fee schedules, and between them they answer most of what a practice actually wants to know. Open Dental lists every fee it charges on one page, from the software licence to a single text message. Practice-Web does the same, and says out loud why its competitors do not. Its pricing page asks: why is it that other vendors don't even disclose pricing on their websites?

Open Dental's licence and support is $199 a month per location for the first 12 months, dropping to $149 a month from the thirteenth. That covers every computer in the office for up to three providers, with each dentist beyond three adding $20 a month, hygienists not counted. Canada is $164 falling to $137, other countries are $89, and dental schools pay nothing. Practices that signed up before 2 February 2026 keep the old $179 initial rate for their first year, which is how you can tell the price moved this year.

Open Dental also publishes its own price history, which almost nobody does. The 2003 prices were $139 for the first year and $99 from the thirteenth month, with increases in 2008, 2015, 2018, 2022 and 2026. The company notes that the 2003 prices adjusted for inflation to December 2025 would be $247 and $176, so the 2026 prices are about 19 and 15 percent lower in real terms than the ones it charged twenty-three years ago.

Practice-Web sits $20 below Open Dental at both stages, at $179 a month falling to $129, and it uses the same shape: 12 month initial term, same $20 for each dentist past three. Its list prices, the ones it calls regular rather than special, are $199 and $149, which are Open Dental's numbers exactly. Two independent vendors, one structure, one twenty-dollar gap.

Above them sit the all-in cloud platforms. tab32 publishes $125 a month for Alpine in year one, described as a start-up rate for up to three providers, and $225 from year two for up to five, with claims at 20 cents each, attachments at 50 cents and eligibility checks at $1.25. CareStack publishes $829 a month for Essentials and $1,299 for Intelligence, and its Essentials tier includes online appointments, the digital patient portal and forms in the price. Weave, which sells the messaging layer rather than the record system, says its plans start from $199 a month. Everything else we opened, including Denticon, Oryx and NexHealth, asks you to book a call.

What a booking button costs every month

Between $75 and $119 a month at the two vendors that will tell you, on top of the software itself, and it is the single line item most often missing from a website quote. The button lives on your site, so it feels like part of the website. It is sold by the company that owns your appointment book, because that is the only place a real appointment can be written.

Open Dental prices it plainly. Web Sched New Patient and Web Sched Existing Patient together are $75 a month per location. Web Sched Recall, the one that texts patients who are due, is another $75. Web Sched ASAP is $20, or free if you already have Recall. Directly above those rows, Web Forms is free for customers on support, as are the Patient Portal, the Payment Portal, eReminders and automated thank-you messages.

Practice-Web charges more for the same job and less for others, which is why comparing whole vendors rather than line items is misleading. Its Online Scheduling is $119 a month, with online forms broken out at $49, patient reviews at $69, tablet patient forms at $65, teledentistry at $49 and two-way texting at $90 including 2,000 outgoing messages. Every one of those is an advertised special against a regular price roughly double it, and every one carries a $199 setup fee that the page says is waived.

Add the four rows most practices end up with and the picture is concrete. On Open Dental, scheduling at $75 plus texting at $5 plus e-prescribing at $57 for one provider is $137 a month on top of the $199 licence, with forms and the portal included. On Practice-Web the same four are $119 plus $49 for forms plus $90 plus $59, which is $317 on top of $179. Same category of practice, same website, a difference of $180 every month that no website quote will ever mention.

The price changes in month 13

Two of these vendors cut your price after a year and one raises it, so the cheapest quote in month one is not reliably the cheapest in month thirteen. Open Dental drops from $199 to $149 once you have been on support for 12 consecutive months. Practice-Web's special drops from $179 to $129 on the same schedule. tab32 moves the other way: its published Alpine rate is $125 for year one as a start-up price covering up to three providers, and $225 from year two as an established-practice price covering up to five.

Over 24 months, and ignoring every add-on, Open Dental is $2,388 then $1,788, so $4,176. Practice-Web is $2,148 then $1,548, so $3,696. tab32 is $1,500 then $2,700, so $4,200. The vendor that is 37 percent cheaper in month one finishes the two years as the most expensive of the three. CareStack does not move at all, and at $829 a month it is $19,896 across the same two years, which buys a genuinely different product with the booking, forms and portal inside it.

Open Dental also publishes discounts almost nobody asks about: prepay six months for 5 percent off, a year for 10 percent, two to three years for 15 percent. On the $199 rate a one-year prepayment saves $238.80. It also runs a 90 day money-back guarantee on support fees, and one on the add-on services, provided you stop using them.

What it costs to take your data with you

Between $1,195 and $1,450 for a typical single-database United States practice, before images, and this is the number that decides how free your next website decision really is. Open Dental publishes $1,450 per database as the typical United States conversion fee, plus $400 for each additional clinic sharing that database. Scanned documents and patient pictures are a separate $850, x-rays another $900, neither included in the data conversion. A Canadian conversion is typically $1,100. Practice-Web publishes $1,195 against a regular price of $1,295.

Test conversions are free at both, which is worth knowing before you commit. Open Dental also prices the awkward jobs it will not discuss over the phone: merging two databases runs $500 to $1,100 depending on size, splitting one runs the same, renumbering patient IDs is $200, and if you are in a jurisdiction whose privacy law forbids sending patient data to the United States, the quote adds $400 because the conversion has to run on your machine.

This matters to a website buyer for one reason. If your site's booking, forms and reminders are all provided by the record system, then changing record systems changes your website, and the price of changing record systems includes a four-figure data conversion and a scanned-image bill on top. Anything the agency has built directly against that vendor's system gets rebuilt. It is the same class of decision we set out in what to check before you sign a web design contract, applied to the software rather than the site.

Your new patient offer is regulated content

In several states the "Free Exam and X-Rays for New Patients" banner is not a design decision, it is a compliance document with required wording, and the rules are state by state. The American Dental Association publishes its own guide to the federal and state advertising rules that apply to dentists, and reading it changes what you brief an agency to build.

Florida is the strictest of the examples the ADA sets out. Under the Board of Dentistry's rule on advertising of fees and discounted services, any advertisement containing fee information must carry a disclaimer that the fee is a minimum fee only. It must state a period during which the fee applies, or the service must stay available at or below that fee for at least 90 days after the final advertisement. Each advertised service must be described using the exact wording from the ADA's Code on Dental Procedures and Nomenclature, or by its specific code number. And any advertisement of a free or discounted service must carry this, in capital letters, clearly distinguishable from the rest of the text:

That is four separate requirements on one landing page, and a designer who has never built for a Florida practice will produce none of them. Other states impose different ones.

The penalties are real rather than theoretical. New York can fine a licensed dentist up to $10,000 for each violation of its rules on false or unsubstantiated claims, alongside censure, suspension or revocation. Arkansas can fine up to $1,000 per violation and impose probation of up to 18 months per violation. Several states go further on the claims themselves: Idaho restricts advertising that a procedure is painless, Florida prohibits superiority claims whether or not you can substantiate them, and Missouri restricts claiming a procedure is curative or preventative. The word "painless" in a hero headline is a state law question in some places.

Testimonials are their own topic, governed federally as well as by state boards, and we have written about the federal side separately in what you can and cannot publish in a testimonial. The practical instruction for a dental build is short. Before anyone writes the homepage offer, get the current rule from your own state board in writing, and put the required wording in the brief rather than discovering it after launch.

The word that is not yours to use

Most dentists cannot describe themselves as a specialist in the thing they are best at, and the services menu is where practices get this wrong. The National Commission on Recognition of Dental Specialties and Certifying Boards recognizes exactly 12 dental specialties: dental anesthesiology, dental public health, endodontics, oral and maxillofacial pathology, oral and maxillofacial radiology, oral and maxillofacial surgery, oral medicine, orofacial pain, orthodontics and dentofacial orthopedics, pediatric dentistry, periodontics and prosthodontics.

Cosmetic dentistry is not on that list. Neither is implant dentistry, sedation dentistry or sleep dentistry. Those four are, in our experience, among the most common items in a dental site's navigation.

By the ADA's own workforce data, 21.2 percent of professionally active dentists reported an ADA-recognized specialty in 2024. The other four in five are general dentists, and before they advertise the ADA points them at three sections of its Code: 5.F on advertising, 5.H on announcing specialization and limitation of practice, and 5.I on general practitioner announcement of services. Read those three with your own state board's rule open beside them, because the board is the one that can act.

State boards enforce the same idea with sharper teeth. Tennessee bars a dentist from claiming to be a specialist, to be specially qualified in a branch of dentistry, or even to be giving special attention to one, until the board has certified them. Michigan bars advertising as limiting practice to, or being specially qualified in, a branch of dentistry without board certification. The ADA's guide records a Maine case in which the board found a dentist's advertisements touted professional specialties, imposed a $1,500 civil penalty, suspended his permit for 15 business days, added a year of probation with an ethics course, and required board approval in advance for every future advertisement.

None of that stops a general dentist marketing implants or veneers. It changes the words. "Implant specialist" is a risk in many states. "We place implants, and our dentists are general dentists" is not.

What these numbers sit next to

All of it is small against what a dental practice turns over, which is the real argument for getting the website right rather than cheap. National dental expenditure in the United States was $189 billion in 2024, 3.6 percent of all health spending, according to the ADA Health Policy Institute's analysis of CMS data released in January 2026. There were 202,485 professionally active dentists that year, 59.5 for every 100,000 people.

One divided by the other puts the average working dentist in front of about $933,000 of annual dental spending. That is an average across every professionally active dentist rather than any one practice's revenue, and it is a blunt instrument, but it sets the scale. Spread across roughly 250 working days it is about $3,700 a day.

Against that, a year of Open Dental's online scheduling at $75 a month is $900, or about a quarter of one working day. The $180 a month gap we found between two software vendors' add-on stacks is $2,160 a year, a little over half a day. A $6,000 website is under two days. None of which means spend carelessly. It means the question worth arguing about is whether the site fills chairs, not whether the build quote is $4,000 or $6,000.

Our two numbers

We publish both of ours, which is the least we can do having spent this article complaining that nobody else does. A focused build is $1,499, once, as a flat package price rather than an opening bid. Ongoing design and development partnership is $2,499 a month. Launch for a focused build usually lands around the two week mark, a pattern across the 200-plus projects we have shipped rather than a guarantee, and partnership requests get an answer inside 48 hours.

What that does not include is the software. We do not sell practice management systems, we do not resell Open Dental or Practice-Web or CareStack, and we take no margin on any of it. If your site needs a booking button, one of those vendors will bill you $75 or $119 a month for the scheduling behind it, and that bill has our name nowhere on it. We think that separation is the right way round, and it is also the reason our number can be flat.

We are the wrong choice for a few practices, and it is cheaper for everyone if we say so here. If you want one company to own the website, the phones, the reviews, the ads and the record system on a single invoice, buy from a dental platform, not from us. If your practice is on a system whose only integration path is a proprietary partner program, the platform's own website product will plug in more cleanly than anything we build. And if what you actually need this quarter is a state-compliant fee page and nothing else, ask your board first and a lawyer second, not a design agency. Our work in this sector sits on the healthcare side of our practice, and our full price list is on the pricing page.

The renewal calendar

Before you sign anything, build a one-page calendar of every recurring charge attached to your website, because in this market the total is a stack of subscriptions with different clocks, not a single annual fee. Most practices we talk to can name the website invoice and nothing else. The exercise takes twenty minutes and it is the closest thing to a decision aid this article can give you.

One row per recurring charge. Fill it in from invoices rather than memory.

Four rules keep it honest. Put the price you will pay in month 13 in the per month column, not the one you are quoted today, because on the evidence above it changes in both directions. Write the notice period even when you have to phone and ask, since a rate you cannot leave for two more years is a different product from the same rate month to month. Mark every row where the same company appears twice, because that is your real switching cost rather than the individual line. And put the login column in, because a renewal you cannot see is a renewal you cannot cancel.

Then total the monthly column and multiply by 24. That number, not the build quote, is what you are actually deciding. On the published figures above, a solo practice running scheduling, forms, texting and e-prescribing pays $336 a month in year one and $286 in year two on Open Dental, against $496 and $446 on Practice-Web. Identical websites, $7,464 against $11,304 over two years, a gap of $3,840 that no website quote contains.

The four things worth paying for first

Spend on the things a person in pain at 9pm needs, and defer everything else until those work. In this market the order is unusually clear, because three of the four are gated by decisions outside the website.

Booking comes first, and it comes with a decision attached. A form that emails the front desk is not booking, it is a message; real booking writes into the appointment book and needs the scheduling module your software sells. Decide which one you are buying before the site is designed, because the button, the confirmation and the reminder all hang off it. If the $75 or $119 a month is not in the budget, say so early and design an honest request form instead of a fake booking flow.

Second, the money page for whatever you actually want more of: implants, orthodontics, sedation, whatever the practice is built around. One page, written in the words patients use, with the fee language your state board requires and no specialty claim you are not certified to make. This is the page most likely to be both the highest value and the highest risk on the site.

Third, the boring facts, correct and structured: hours including the ones that differ, address, parking, which insurances you take, whether you see children, what happens in an emergency after hours. These are what a search engine and an AI assistant read to answer someone's question about you, and they are the cheapest thing on this list.

Fourth, forms, and only after you have checked who bills for them. On Open Dental they are included with support, on Practice-Web they are $49 a month, and a patient who can complete intake before arriving is worth more than most of what a redesign changes. Everything else, the photography, the video, the animation, is worth real money once these four work and worth nothing while the booking button emails the front desk. The recurring bills behind all of it are the same species of cost we broke down in what website maintenance actually pays for.

That is the shape of this whole market, and it is why the cost guides on page one cannot answer the question they are titled after. They price a project. The companies actually selling restaurant websites price a channel, and almost all of them put a percentage somewhere in the deal. Every figure below came off a seller's own pricing page or a regulator's own text, read on 27 August 2026, and the arithmetic is shown so you can redo it with your numbers. We sell websites too, so read this as a price list published by one of the firms you might be comparing.

The 30-second answer

A restaurant website is rarely sold as a website in 2026, so a flat range cannot price it. The specialist platforms bundle the site into a subscription with a percentage attached: Owner.com is $249 a month plus 5% per order or $499 a month flat, with a 5% guest fee on both, and Flipdish starts at $119 a month billed annually. The point of sale companies give you the site and charge on the card: Square is $0, $49 or $149 a month at 3.3% or 2.9% plus thirty cents online, and SpotOn is $0 or $55 a station with card rates of 2.79% or 2.45%. Toast publishes a $69 point of sale price and no website price at all. Our own numbers are $1,499 once or $2,499 a month. The right question is not what a restaurant website costs. It is what your order volume makes it cost.

Why the published ranges disagree with each other

They disagree because a range is the wrong shape of answer, and reading the pages that rank shows how wide the disagreement is. On 27 August 2026 we opened four of the guides returned for this question, and one of the four refused to serve us a page. Of the three we could read, one gives $500 to $50,000, one gives $0 to $50,000, and the third leads with $2,500 to $9,000 before splitting into tiers of $1,500 to $4,000, $4,000 to $8,000 and $8,000 to $25,000. Two ranges that run from almost nothing to fifty thousand dollars are not answers. They are the question, restated.

Credit where it is due: two of those three do account for a per-order fee, and one works through what a 5% charge costs a restaurant selling $30,000 a month. What none of the three mentions, anywhere, is a fee charged to the guest. That is the number this article was written around, and it is missing from every page currently ranking for the query.

There is a deeper reason none of them can hold. Every one of these ranges treats the price as a fact about the site you are buying. In this market it is far more often a fact about how much you sell. Square's free ordering tier costs a restaurant selling $5,000 a month online $225, and a restaurant selling $25,000 a month $1,125, on orders averaging $25 in both kitchens and identical software in both. No range expresses that, so every range hides it.

The four ways this market actually prices a website

There are four, and once you can name them the quotes stop looking random. The first is a flat subscription with no percentage, which is rare and is what most people think they are buying. The second is a subscription plus a per-order fee charged to the restaurant. The third is a percentage charged to the guest at checkout, which never appears on your invoice at all. The fourth is a card processing rate, where the software is given away and the margin is recovered on every payment, including the ones taken at the table.

Most contracts mix them. Owner.com's cheaper plan runs models two and three at once. Square runs model four with a subscription bolted on. SpotOn runs model four and adds a bundle priced in basis points on top: its Core Bundle is $50 a month plus 20 basis points of gross payment volume, capped at $200, and the product inside it that handles first-party ordering is described on the same page as commission free. Both statements are true. The commission moved rather than disappeared.

The fee your guest pays, on both plans

Owner.com's plan cards make a promise that its own frequently asked questions then qualify. The cards read: Flexible, $249 a month plus a 5% restaurant fee per order, and Flat Rate, $499 a month with no additional restaurant fees, described as best for restaurants at $5,000 or more a month in online sales.

Open the question headed what commissions do I pay as a restaurant, and the second half of the deal appears. On the Flat plan there is no additional restaurant fee. On the Flex plan there is a 5% restaurant fee. And on both plans, in the company's own words, guests pay a 5% order support fee that covers fulfillment and customer service.

Nothing here is hidden and nothing here is unusual. Owner.com states the fee, and its point that a guest still pays less ordering direct than through a marketplace charging up to 15% is fair. What matters is that the fee exists on the plan sold as fee free, which means the two plans differ by less than the cards suggest, and that on the cheaper plan a single order carries two separate five percent charges pointed at two different people.

It matters commercially as well. If the reason you are moving orders onto your own website is that the guest experience should be better and cheaper there, then the number the guest sees at checkout is part of the product you are buying. Ask what it is before you sign, and ask whether it can be turned off, because that answer decides how your direct channel compares with the marketplace you are trying to leave. We wrote up what those marketplaces charge in our piece on delivery app fees and direct ordering, and the two articles are meant to be read against each other.

What a plan priced at zero actually costs

A $0 plan is a financing arrangement, and two vendors publish the terms clearly enough to do the sums. Square offers restaurant online ordering on four tiers: Free at $0 a month, Plus at $49, Premium at $149, and a custom Pro tier for businesses processing more than $250,000 a year. The tiers differ on features, and they also differ on rate. Free is 3.3% plus thirty cents on an online order. Plus and Premium are both 2.9% plus thirty cents.

That 0.4 of a percentage point is the price of the free plan. Divide the $49 subscription by it and the volume where the two land on the same number appears: $12,250 a month in online orders. Underneath that figure a restaurant is right to stay on Free. Past it, the plan with no monthly fee is quietly the more expensive of the two, and the penalty grows with every good month. We found the identical structure in gym management software, where one vendor's free tier carried the highest card rate on its page, so treat it as a pattern in this category rather than a quirk of one company. Our breakdown of what a gym website costs has that version of the sum.

The same test applied to the tier above produces a blunter answer. Premium costs $100 a month more than Plus and buys no improvement at all on the online rate, which stays at 2.9% plus thirty cents. Its only rate advantage is in person, 2.4% against 2.5%, a tenth of a point. You would need $100,000 a month in card payments taken at the table before that tenth of a point pays the extra $100. For most independents, Premium is a feature purchase, and it should be argued for on features.

SpotOn is built the same way. Its All-In plan is $0 a station a month with hardware included, at 2.79% plus twenty cents on a card present sale, and it carries a two year minimum term. Its POS Essentials plan is $55 a station a month, month to month, at 2.45% plus fifteen cents, with a Station 15 listed at $750 and a handheld at $297. The rate gap is 0.34 of a point plus five cents a transaction. On a $25 average check that is about thirteen and a half cents an order, so roughly 408 transactions a month, near $10,200 in card volume, before the $55 plan wins on rate alone. Hardware and the two year term sit on the other side of that trade, and they are the reason the free plan exists.

What each platform puts on its own page

Here is every price we could read, with the parts each company leaves out. The gaps are as informative as the numbers, because the missing figure is almost always the one that scales with your sales.

Toast deserves a sentence of its own, because its page is the most opaque of the set. Its pricing page, marked last updated 20 August 2026, contains exactly four dollar figures across the entire document: $0, $69, $90 and $9. Payment processing appears as a feature row reading simple flat rate, with no rate. Toast Websites is listed inside a suite called Digital Storefront Pro, which appears in the comparison chart with no price anywhere on the page. You cannot cost a Toast website from Toast's website.

How long the advertised price is committed for

Compare terms before you compare monthly figures, because several of these prices are only available if you sign away one to three years. This is the column that turns a cheap quote into an expensive one, and it is rarely in anybody's comparison table.

Notice the pattern. Owner.com, whose subscription is the highest number on this page, is the one committing to nothing: month to month, no long-term contracts, no cancellation fees, in its own words. Clover's $179 is a 36 month subscription price, and the alternative it offers is $1,799 up front plus $89.95 a month. SpotOn's zero dollar plan carries a two year minimum term while its $55 plan does not. Flipdish's $119 is the annual rate and the monthly rate is $149, which is about 25% more for the right to leave.

None of that is wrong. Long terms fund hardware and onboarding, and a business that knows it is staying should take the discount. The mistake is comparing a three year price with a monthly price as though they were the same thing, which is what every quote you receive will invite you to do.

Your point of sale now picks your website

One of the most recommended restaurant website companies in the market will not sell to you at all any more unless you buy a till from its parent. BentoBox spent a decade as the design-led answer for independent restaurants. Its own home page now opens with the line BentoBox is now Clover, and its frequently asked questions are blunter than that.

The advice available to a buyer has not caught up. We searched for the best restaurant website builders of 2026 on 27 August 2026 and took the seven pages returned. Six were readable to us and the seventh returned nothing we could parse. Three of the six recommend BentoBox: one ranks it third overall, one quotes a monthly range of $119 to $479 for it, and one files it under restaurant operations specialists. Not one of the six mentions Clover anywhere on the page. A buyer following the top of these results is shopping for something that is no longer sold on the terms described.

Take the practical consequence rather than the sentiment. Choosing a restaurant website is increasingly a downstream effect of choosing a point of sale, because the site, the ordering flow, the menu data and the payment rate now ship together from one company. That is genuinely convenient. It also means the decision that sets your card rate for the next three years is being made at the moment you pick a piece of countertop hardware, usually by whoever is most persuasive in the room.

So do the two decisions separately, at least on paper. Write down what you would pay for the point of sale alone, what you would pay for the website and ordering alone, and what the bundle costs. If a vendor cannot break the bundle into those parts, that is worth knowing before you find out which half of it you dislike.

What is riding on the ordering page

The reason any of this is worth an afternoon is that the takeout counter is now most of the business. In its 2025 Off-Premises Restaurant Trends report the National Restaurant Association put nearly 75% of all restaurant traffic off premises, meaning close to three in four orders leave the building. Mobile ordering had been used recently by 57% of adults, rising to 74% of millennials and 65% of Gen Z adults. Sixty-five percent of limited-service operators offered delivery, and off-premises accounted for a larger share of sales than in 2019 for 58% of limited-service and 41% of full-service operators.

Read those numbers next to a percentage and the stakes land differently. If three quarters of your orders travel, the page that takes them is not brochure ware with a menu on it. It is the counter. A tenth of a point on a card rate, or five cents in the dollar taken from a guest, applies to the majority of what you sell, every day, without an invoice ever arriving to prompt a review.

When your menu stops being marketing

If you operate twenty or more locations, the menu on your website is a regulated document, and a surprising number of growing groups cross that line without noticing. The Food and Drug Administration's menu labeling rule at 21 CFR 101.11 defines a menu to include, in its own words, electronic menus and menus on the Internet. Your website menu is inside the definition, not adjacent to it.

The rule reaches a covered establishment, defined as a restaurant or similar retail food establishment that is part of a chain with twenty or more locations doing business under the same name and offering for sale substantially the same menu items, regardless of the type of ownership. Individually owned franchises count toward the twenty. So do the locations you opened last year while the website stayed on the plan you bought when there were six of you.

What the rule then requires is specific and it is a build detail, not a legal footnote. Calories go next to each standard menu item. Two statements have to appear as well: the succinct statement, worded 2,000 calories a day is used for general nutrition advice, but calorie needs vary, which must appear at the bottom of each page of the menu; and the statement additional nutrition information available upon request, which the regulation says must be on all forms of the menu or menu board. A content management system that cannot place two persistent lines under a paginated menu, in a type size no smaller than the calorie figures, has a real problem the day your twentieth location opens.

Two footnotes worth carrying into the vendor conversation. The rule has a voluntary registration route at paragraph (d): an operator below twenty locations can elect to be covered, and doing so replaces non-identical state and local nutrition labeling requirements with one federal standard, which is occasionally the cheaper answer for a group operating across several jurisdictions. And none of this is your platform's duty. It is yours, which is why the question to ask is not whether the vendor is compliant but whether the template lets you comply.

What we charge, and where we are the wrong call

We publish two numbers and neither has a form in front of it: $1,499 for a one-time build, and $2,499 a month when the site is treated as continuing work rather than a delivery. Both are on our pricing page, and there is no percentage attached to either, on your side or your guest's. Once your content and photography are in hand, a focused build is typically live in about two weeks, and a partnership request gets a reply inside 48 hours. Our work with restaurants and hotels sits on our restaurants and hospitality page.

Now the honest half. If you need a till that talks to a kitchen display, a menu that syncs to four delivery integrations, tableside handhelds, payroll, and one support number for all of it at two in the morning, buy a platform. Owner.com, Toast, Square, SpotOn and Clover are selling something we do not sell, and the percentage in their pricing is what funds the parts we would have to bolt together for you. At real volume those platforms are also cheap per order. We would rather say that than pretend a design studio replaces a point of sale.

Call us when the website itself is the part that is failing you. When your food photography is doing nothing for you, when the reservations page loses people on a phone, when your private dining and group booking pages sit three clicks below the fold, when your brand looks like the template it was built from. That is the kind of problem we solved for Cornerstone Healing Center, whose redesign lifted website conversions by 20%, and for Eden Digital, which grew revenue tenfold on a site we built. Past 200 projects have gone out of this studio and not one of them was a point of sale, which is the disclosure that belongs next to the rest.

The decision aid: price one order, both sides

Every quote in this market becomes comparable once you push a single order through it and write down two answers instead of one: what leaves your account, and what leaves your guest's. Take a real order at your real average check, at your real monthly volume, and fill this in for each vendor. Twenty minutes of arithmetic reorders most shortlists.

Four rules keep it honest. Use your current volume, not the volume the salesperson is projecting, because the subscription row is the one that flatters you as you grow and punishes you if you do not. Leave any row blank that the vendor will not fill in, and treat the blank as a number you will discover on a statement later. Run the table twice, once at today's volume and once at double, since the ranking often flips between the two. And put the term in the last row rather than in the footnotes, because a rate you cannot leave is not the same product as a rate you can.

Worked on the published figures above, at 400 orders a month averaging $25, Owner.com's Flexible plan costs you $249 plus $500 in restaurant fees before any card processing appears, its Flat Rate plan costs $499, and your guest is charged $1.25 an order in either case. Square's Free plan on the same volume is $450 all in and charges the guest nothing. Those three numbers describe genuinely different businesses, and no range containing all of them tells you anything.

The order we would build it in

Build the things a hungry person needs at 6pm before the things that photograph well. The order below is the one that survives contact with a real service, and it holds whether we build the site or your platform does.

The menu comes first, as real text on a real page, with prices, in the page's own HTML rather than in a PDF or a picture of a menu. It is what almost everybody came to the site for, and it is the thing most often published in a format that a search engine or an AI assistant reads badly. Then hours and location, correct and structured, including the hours that differ on holidays. Then the ordering path, tested on a phone with one thumb, timed from the menu to a confirmed order. Then reservations or waitlist, if you take them, without a second login. Then the pages that carry your bigger tickets: private dining, catering, gift cards, events. Then photography, which is worth real money once everything above works and is worth nothing while the menu is a PDF.

The decision that costs the most later is not any of those. It is ownership. Settle who holds the domain, where the menu data lives, whether the customer list can be exported, and what becomes of your ordering URLs the day you leave, which we set out in what to check before you sign a web design contract. On a deal priced as a share of every order, for a term measured in years, the cost of not being able to walk away compounds one ticket at a time.

We went looking for that number on twelve other vendors' sites on 26 August 2026 and mostly did not find it. What follows is every price we could read on a seller's own page, our own arithmetic wherever a published formula had to be applied, and the federal rules that decide what your site has to carry before anybody argues about the photography. Disclosure, since it shapes everything below: we build websites, schools included, so treat what follows as a price list written by one of the companies you might be comparing.

The 30-second answer

A private school website in 2026 is priced two incompatible ways. Specialist school platforms fold the site into an annual subscription that scales with enrollment, and the one vendor that publishes its rate charges $7,500 plus $3.80 per student a year, with a one-time $9,500 plus $700 per organization on top. Agencies and general web shops quote a project fee instead, and the largest school vendor's own guide puts a theme-based school site at $5,000 to $25,000 and a custom one at $35,000 to $75,000. Our own two numbers are a flat $1,499 for a one-time build and $2,499 a month when the site is treated as continuing work. The useful answer is a formula, not a range, and no two quotes are comparable until each is converted to a cost per enrolled student.

Why the published ranges cannot all be right

They cannot all be right because they are describing different products at different scales and presenting the result as one number. Searching for what a school website costs on 26 August 2026 returned a first page that was mostly about tuition rather than about websites. The guides that are about websites tend to give a range with no source attached, and the ranges do not agree with each other or with the prices the vendors themselves publish.

The deeper problem is that a range assumes the price is a property of the website. In this market it is frequently a property of the school. A vendor that charges per student is not quoting you for a set of pages. It is quoting you for a license whose cost is set by your enrollment figure, which means two schools can buy the same templates, the same content management system and the same mobile app and pay amounts that differ by thousands of dollars a year. A range cannot express that, so it hides it.

The one vendor that publishes a formula

Apptegy publishes an actual formula, and reading it tells you more about school website pricing than any guide we found. Its Foundations package, headed Website and Mobile Apps, is listed at $7,500 plus $3.80 per student as a yearly cost, and covers the platform, an integrated website and content management system, student information system integration, text and voice alerts, email notifications, native iOS and Android apps, push notifications, training and support. Setup and development is charged separately as a one-time cost of $9,500 plus $700 per organization.

Apply the annual half of that formula to four schools and the shape of the market appears. A school of 150 students pays $8,070 a year. At 400 students it is $9,020. At 800 it is $10,540, and at 1,500 it is $13,200. Those are our sums on the vendor's published rate, not quotes we were given.

Notice what the chart does not do. Between the smallest school and the largest, enrollment rises tenfold and the annual bill rises by about 64 percent. If you are the head of a small school reading a guide that says school websites cost between eight and thirteen thousand dollars, you will conclude you are being quoted normally. You are, and you are also paying for a price floor that was built for someone else.

What that same contract costs per student

Divided by the students it covers, the same contract costs a small school just over six times what it costs a large one. At 150 students the annual figure works out to $53.80 per student. At 400 it is $22.55, at 800 it is $13.18, and at 1,500 it is $8.80. The product does not change. The unit price collapses because the fixed $7,500 keeps being spread across more people.

Add the one-time charge and the gap widens further. For a single school buying setup and development at $9,500 plus $700 for one organization, year one is $18,270 at 150 students and $23,400 at 1,500 students, which is $121.80 and $15.60 per student respectively. The per student figure is the one to carry into a board meeting, because it is the only version of the price that survives a comparison against a competing quote. A trustee at a 400 student school who asks whether $19,220 is a lot for a website has asked an unanswerable question. The same trustee asking whether $48.05 per student in year one is a lot has asked one you can answer.

What 13 vendors will tell you before you call

Most of them will tell you nothing. On 26 August 2026 we requested the pricing page of thirteen companies that sell websites to schools: Apptegy, Blackbaud, Digistorm, Edlio, Finalsite, Foxbright, Gabbart, Juniper Education, Morweb, SchoolBlocks, SchoolStatus, Sycamore and Veracross. Two of them publish a price a visitor can read. Six have no pricing page we could find from their own navigation or at the usual addresses. Three ask you to request a quote instead of showing a number. One puts its pricing page behind a login. One returned an error to us.

The vendor behind the login is the biggest name in the category, which makes its page worth seeing. Finalsite has a pricing page at the address you would expect, and that page asks for a username and a passkey.

Blackbaud takes the opposite approach and answers the question directly in a frequently asked questions block on its own pricing page, under the heading "Why doesn't Blackbaud list pricing on the website?" Its stated answer is that each organization has different needs, so it prepares a personalized quote. Elsewhere in the same block it says pricing is based on several factors including the organization's size and its number of users. Veracross describes its pricing as customized and tailored to each school. SchoolNow and Campus Suite, two brands we started the exercise treating as separate vendors, both now redirect their pricing addresses to one SchoolStatus product page with no prices on it, which is its own small lesson about how quickly this supplier list consolidates.

We are not claiming any of this is sinister. Enterprise software is often sold this way, and a school with three campuses genuinely does need a different contract from a single-site Montessori. But it does mean the buyer carries the whole burden of comparison, and it means the two published prices in the chart above are doing an enormous amount of work for everyone else's benchmarking.

What the largest vendor's own cost guide says

Finalsite publishes a long article answering exactly the question in this headline, and it is a good deal more useful than most, with one revealing gap. It opens by saying the honest answer is that it depends, then does something better than most competing guides and explains what the answer depends on: design scope, interior page templates, campaign landing pages, integrations, hosting, security and strategic consulting.

Where it becomes interesting is the numbers it does give. A theme-based school website is put at $5,000 to $25,000. A custom site with consulting, multiple design rounds and content strategy is $35,000 to $75,000. Enterprise sites for large districts or schools running multiple microsites go up to $150,000 and beyond. It also prices its alternatives: freelancers are put at $2,000 to $50,000, do-it-yourself hosting at $100 to $500 a year, a domain and certificates at $50 to $100 a year, and an in-house developer to maintain the result at $45,000 to $75,000 a year.

Every figure in that table describes a way of buying a website. None of them describes what Finalsite charges. The company that publishes the most detailed cost guide in this category prices the freelancer, the do-it-yourself route and the in-house hire, and keeps its own number behind the login in the screenshot above. That is only a fair thing to point out if we do better, so ours are on our pricing page and in the section below.

One line in that guide is the most important sentence in this article, and it is Finalsite's, not ours. Explaining that some companies bundle design cost into the software subscription, it notes that the upfront design may be partly or fully included depending on your enrollment. The largest vendor in the category is telling you, in its own words, that the number of students you have moves the price of your website.

The bills your website generates but does not contain

The website line item is usually the smallest number a school signs, and the systems attached to it are also priced per head. This is where a build decision turns into a recurring cost that lands on families rather than on the school, which is why it belongs in the same conversation as the design brief.

Tuition payment plans are the clearest case. Holy Family Regional Catholic School publishes its options on its own site: paying in full carries no fee, a two-payment plan carries a $10 fee, and the ten-payment plan carries an annual fee of $45. That $45 is charged to each family every year and paid by the family, not the school. Two hundred families on the ten-payment plan generate $9,000 a year in plan fees, which is more than the entire per student half of the Apptegy contract at a school of 1,500.

The same rail carries the school's own charges. St. Pius X Catholic High School publishes a non-refundable enrollment fee of $400 per student billed through its FACTS account. Academy of Holy Angels publishes a fee schedule with a re-enrollment charge of $350 for upper school and $300 for middle school, a technology fee of $850 and $750, capital maintenance at $500 and $250, fundraising at $500 and $250, textbook fees at $275 and $150, plus graduation fees. None of that is website spending. All of it is money that moves because a page on a website collected a signature and an account number, which is exactly why the intake path deserves more design attention than the homepage hero.

The parent organization has its own subscription. Membership Toolkit, which sells to parent teacher organizations and booster clubs rather than to schools, publishes three annual plans at $550, $850 and $1,150, with text messaging billed at $2.50 a month plus two cents per message after a free allowance. Morweb, which counts schools among its markets, publishes its nonprofit tiers at $149 and $199 a month. Those are the realistic reference points for the smaller site that sits beside the school's, and they are useful precisely because so few school vendors will show you one.

The rule whose deadline passed in April

If any part of your site is aimed at children under 13, or you know you are collecting personal information from them, the amended Children's Online Privacy Protection Rule has applied in full since 22 April 2026. The Federal Trade Commission published the amendments on 22 April 2025 at 90 FR 16918. They took effect on 23 June 2025, and the rule text gave regulated entities until 22 April 2026 to comply, with three narrow exceptions.

Two of the amended provisions are website deliverables rather than policy documents in a drawer. 16 CFR 312.10 now requires a written data retention policy setting out why children's personal information is collected, the business need for keeping it and a timeframe for deleting it, and it requires that policy to appear in the notice on the site itself. That is a page somebody has to write, publish and keep current. 16 CFR 312.8 requires a written information security program with a named coordinator, risk assessments at least annually, safeguards designed against those risks, regular testing, and written assurances from any third party you let touch children's data. Your web vendor is a third party you let touch children's data.

There is a gap here that most summaries get wrong. In 2024 the Commission proposed adding definitions of a school and of a school-authorized educational purpose, along with a school authorization route to consent, and then declined to finalize any of it. The stated reason is that the Department of Education signaled its own plans to amend the FERPA regulations, and the Commission did not want the two to conflict. So the tidy provision letting a school consent on parents' behalf is not in the rule. If a feature on your site collects data from a child under 13, do not assume the school can wave it through, and get an answer from your own counsel before you build it.

Whether student privacy law applies to you at all

For most independent schools, FERPA does not apply, and this surprises people. The regulations set their own scope at 34 CFR 99.1: they cover an educational agency or institution to which funds have been made available under a program administered by the Secretary of Education. A private K-12 school that takes no such funds is outside that definition. Public districts are squarely inside it.

For the schools it does cover, the provision that shapes a website is 34 CFR 99.37. Publishing directory information is permitted only after the school has given public notice of three things: which categories it has designated as directory information, that a parent or eligible student may refuse to let any or all of those categories be designated, and how long they have to say so in writing. A photo gallery, an honor roll, an athletics roster and a graduation list are all directory information disclosures. The rule adds a detail worth building for: opt-outs made while a student was enrolled must still be honored after they leave, unless the former student rescinds them.

Two practical consequences follow, and they apply whether or not FERPA reaches you. First, a content management system that cannot flag an individual student as excluded from photographs is a system that will force somebody to police it by memory. Second, if you are outside FERPA, the duty does not vanish, it just moves into your enrollment contract, your state's student privacy statute and the promises your own privacy policy makes. Read all three before your vendor writes the photo release copy for you.

Accessibility, in one paragraph

Private schools are named in the statute, and religious ones are named in the exemption. The ADA's list of public accommodations at 42 U.S.C. 12181(7)(J) includes a nursery, elementary, secondary, undergraduate or postgraduate private school. 42 U.S.C. 12187 then exempts religious organizations and entities controlled by religious organizations from that subchapter, which is why a parochial school and the independent day school across town can face genuinely different exposure. The dated federal web deadline that has been in the news applies to public entities rather than to either of them, and we set out what it does and does not cover in our piece on what actually protects you from a website accessibility lawsuit. Build to WCAG 2.1 AA regardless: families using assistive technology are reading your admissions pages either way.

What we charge, and when we are the wrong choice

Our prices are $1,499 for a flat one-time build and $2,499 a month for continuing work, and both sit on our pricing page with no form in front of them. Once your content and brand assets are in hand, a focused build is typically live in about two weeks, and a partnership request gets an answer inside 48 hours. More than 200 projects have gone out of this studio, and what we do for schools sits on our education page.

Now the part that costs us work. If you need a directory of 90 faculty profiles synchronized from your student information system, a parent portal with role-based permissions, mass notification by text and voice, and a mobile app in two stores, a specialist school platform will serve you better than we will, and the per student subscription this article has been picking at is what pays for those integrations. If your school is large, that subscription is also cheap per student. Buy it.

Come to us when the site is a marketing and admissions problem rather than an operations problem: when inquiries stall, when the tuition and financial aid pages cannot be found, when the application handoff drops families, or when you already own a portal that works and simply need a front door that converts. That is the work we did for Cornerstone Healing Center, where a redesign produced 20 percent more website conversions, and for Eden Digital, which grew revenue tenfold after we rebuilt its site. Neither is a school, and we would rather say that than imply otherwise.

The decision aid: convert every quote to a price per student

Do this before you compare anything, because it is the only step that makes two school website quotes mean the same thing. Take each proposal, add up what it will actually cost you across three years including one-time charges, divide by three, then divide by your current enrollment. Write the answer in dollars per student per year. It takes about fifteen minutes and it will reorder your shortlist.

Four rules keep the arithmetic honest. Use your real enrollment, not your target, because that is the number the invoice will use. Put the one-time charges in, spread across the term you actually expect to keep the site, since a $10,200 setup fee is a different animal over three years than over eight. Include anything the contract makes mandatory, such as hosting, support and required training, and exclude anything genuinely optional. And note next to each figure which parts move when enrollment moves, because that column is your forecast for the year you add a grade.

Worked on the only published formula we have, a 150 student school lands at $76.47 per student per year across three years, and a 1,500 student school at $11.07. Those come from a one-time $10,200 plus three annual charges, added up and divided twice. Any quote you are holding can be pushed through the same two divisions. If a vendor will not give you enough detail to complete the row marked annual per student charge, that refusal is itself a data point, and you should ask a second time in writing.

What we would build first, and in what order

Build the pages that carry a decision before the pages that carry a feeling. Across the 200 plus projects we have shipped for businesses where an inquiry has a season, the ordering that survives contact with a real calendar looks like this.

Tuition and financial aid comes first, as text on a page rather than a downloadable document, because it is the page families search for by name and the one we most often find hiding in a PDF. Then the inquiry and tour booking path, tested end to end on a phone, since a form that hands over to a portal login is where enrollment funnels leak. Then the program pages, one for each grade band or subject that a family would search separately. Then faculty, with real credentials, because a skeptical parent reads it. Then events and deadlines, editable by your admissions office without a developer, because a stale open house date in October does more damage than a dated homepage. Everything else can wait for version two.

The decisions that turn out to be expensive later are the ones about ownership and portability rather than about aesthetics. Before you sign, establish who holds the domain, who holds the content and whether you can export it, which we have written about at length in what to check before you sign a web design contract. On a per-student subscription that renews every year, the cost of being unable to leave compounds quietly, one enrollment figure at a time.

That is why a price for an accounting firm website is harder to give honestly than a price for a landscaping company website, and why most published answers are useless. They price the brochure. Every figure below was read on the seller's own page on 25 August 2026, every legal citation points at the regulation rather than a summary of it, and where a number is our own arithmetic the sentence says so. Websites are what we sell, including to firms in this category, so read what follows as a price list published by one of the vendors on your shortlist.

The 30-second answer

A website for an accounting or tax firm is a one-time build in the low thousands, or a subscription: the accounting-specific platforms we priced run from $90 to $500 a month, and one agency in this niche publishes $499 for a one-time three-page build. We charge $1,499 as a flat one-time fee, or $2,499 a month when the site is treated as continuing work. The website is the small number. The platform that carries client documents is priced per person, from $19 to $149 per user per month across the five platforms that price it that way, so a five-person firm faces roughly $1,140 to $8,940 in year one on the portal alone against about $1,080 for the site itself. That split is not an accident of the market. Federal law treats an accounting firm as a financial institution, and the encryption, multi-factor authentication and vendor oversight it demands all live on the expensive side of the split.

Federal law prices your website differently from a plumber's

Your firm is a financial institution, and the regulation says so by name. The FTC Safeguards Rule defines its own scope at 16 CFR 314.2, and its list of worked examples includes this one: "An accountant or other tax preparation service that is in the business of completing income tax returns is a financial institution because tax preparation services is a financial activity." No revenue threshold, no headcount test, no distinction between a sole practitioner in a spare room and a hundred-person firm.

What follows from that is what the FTC counts as nine required elements in 16 CFR 314.4. Four of them are decisions you make when you buy or build a website. You must designate a Qualified Individual to run the program. You must "protect by encryption all customer information held or transmitted by you both in transit over external networks and at rest." You must implement multi-factor authentication for anyone accessing an information system that holds customer information. And you must oversee your service providers, which now includes whoever hosts your site, by selecting them for their safeguards, requiring those safeguards by contract, and periodically assessing them.

Small firms get less relief here than they expect. 16 CFR 314.6 exempts firms holding information on fewer than five thousand consumers from exactly four things: the written risk assessment, the continuous monitoring or penetration testing and vulnerability assessment duty, the written incident response plan, and the annual report. Encryption is not on that list. Multi-factor authentication is not on that list. Service provider oversight is not on that list. A two-person practice with four hundred clients is exempt from the paperwork and bound by the controls.

The IRS says the same thing in plainer language. Its release of 18 August 2026, IR-2026-92, opens with "Federal law requires tax and accounting professionals to create and maintain a Written Information Security Plan," names the Gramm-Leach-Bliley Act as the source, and lists the FTC's requirements to designate a coordinator, assess risks, monitor and test safeguards, and select service providers whose contracts require compliance. It also restates a reporting rule that is easy to miss, and we will come back to that at the end.

What the website itself costs

Between about $90 and $500 a month if you buy a platform built for this profession, or a low four-figure one-time fee if you buy a build. Here is what the vendors publish, read on 25 August 2026.

CPA Site Solutions prices four tiers on its pricing page: Silver at $90.00 a month, Gold at $111.00, Platinum at $137.00 and Diamond at $166.50. All four include what it calls a Secure Firm Portal, with storage rising by tier from 5GB to unlimited. CountingWorks Pro sells a bigger bundle and prices it monthly with an annual discount: Starter at $150 a month or $1,620 a year, Grow at $200 or $2,160, Premier at $300 or $3,240, and Premier Plus at $500 or $5,400. Each tier includes one website, except Premier Plus which includes two, and each caps users and contacts: one user and 250 contacts at Starter, ten users and 5,000 contacts at Premier Plus.

TaxDome takes a third approach and folds the website into a practice management seat. Its pricing page lists Essentials, Pro and Business at $800, $1,000 and $1,200 per seat per year on a one-year commitment, dropping to $700, $900 and $1,100 on a three-year commitment, all billed upfront. A "fully managed professional website for your firm" appears in the Pro tier, alongside a custom firm URL and branded emails. So the website is free in the sense that a second bedroom is free when you buy the house.

Outside the platforms, published one-time prices in this niche are rare but they exist. MITCO Tech states on its accountant page that "Pricing starts from $499 for a new 3-page, SEO optimized website, including design and build." Build Your Firm, one of the older names in accounting marketing, publishes no price anywhere on its home page. Our own build is $1,499 as a flat one-time fee, or $2,499 a month if you want the site worked on continuously rather than delivered and left alone.

Twelve months of the cheapest of those tiers, CPA Site Solutions Silver, comes to $1,080. Twelve months of its Diamond tier comes to $1,998, which buys a custom design and a certificate we will come back to.

One caution before you budget from any of these. On the same day, CPA Site Solutions' home page advertised "For as low as $76.00 a month" while its own pricing page put the cheapest tier at $90.00. That is a fourteen dollar gap between two pages on one website, and it is the reason the only price worth planning around is one a salesperson has put in an email with a date on it.

The bill nobody quotes: the portal is priced per person

Between $19 and $149 per user per month, and it scales with your staff rather than your client list. This is the line item that turns a cheap website into an expensive year, and the guides ranking for this question leave it out entirely. We pulled three of them on 25 August 2026, between 19,000 and 29,000 characters each, and searched the text: "Safeguards" appeared zero times across all three, "multi-factor" zero, "WISP" zero, "encryption" zero, "per user" zero, and "client portal" once, in one page. It does not look like a website expense until you notice that the upload button on your contact page has to point somewhere.

Six platforms in this category publish their rates. Liscio charges $19 per user per month for Intelligent Files, $49 for the Liscio Platform and $99 for Tax Team, all billed annually. Financial Cents charges $19 a month for a single-user Solo plan, then $49 per user for Team and $69 per user for Scale, and requires a minimum of five users for monthly billing. SmartVault charges $55 per user per month for Business Pro with a three-user minimum, $65 for Accounting Pro and $85 for Accounting Unlimited, both with a two-user minimum, and about a third more if you pay monthly rather than annually. Karbon charges $59 per user per month for Team and $89 for Business on annual billing. Canopy charges $74, $109 and $149 per user per month for Standard, Plus and Premium. TaxDome, as above, prices per seat per year.

Two things in that chart are worth more than the headline rates. The first is the minimum. SmartVault's cheapest plan carries a three-user minimum, so the real floor is $165 a month even for a two-person office, and Financial Cents will not bill monthly at all under five users. The second is the gap between the entry tier and the tier the vendor actually markets to tax firms, which runs from $30 a user at Liscio to $75 a user at Canopy. That gap is where the compliance features sit, which is a pattern rather than a coincidence, and the next two sections are about the two clearest examples of it.

Payment processing is the last piece, and the platforms that publish rates charge it separately. Canopy publishes 3.30% plus $0.20 on cards and 1% on ACH capped at $10. Karbon publishes 2.6% plus $0.30 on standard cards, 3.6% plus $0.30 on premium cards, 3.9% plus $0.30 on American Express and 1% plus $0.30 on ACH capped at $5. On a $236 tax return fee, the difference between taking that payment on a premium card through Karbon and taking it by ACH is $6.14, which is worth having across four hundred returns.

The line item that should not be a line item

Encryption in transit is a legal requirement for your firm, and at least one vendor in this niche sells it as an upgrade. On CPA Site Solutions' pricing page, "HTTPS (SSL) Certificate" is listed as a feature of the Diamond tier at $166.50 a month, and further down the same page "Additional HTTPS Website Security" appears in the add-on list at $9.99 a month.

Compare that to the rule the buyer is bound by. 16 CFR 314.4(c)(3) requires a firm to protect customer information by encryption "both in transit over external networks and at rest", with the only escape being a written finding by the Qualified Individual that encryption is infeasible and that compensating controls are in place instead. The FTC's own plain-language guidance puts it as encrypting customer information on your systems and in transit. A login form for a client portal, embedded in a page served without a certificate, is the exact failure that requirement describes.

We checked whether that theoretical failure is a live one. CPA Site Solutions showcases named client sites on its own website, and we tested three of them on 25 August 2026: hmwcpas.com, greysontax.com and glcpas.com. All three answer on plain HTTP with a permanent redirect to HTTPS, which is the correct behavior. So the practice looks better than the price list. That is worth stating plainly rather than implying a scandal that is not there. What remains true is that the page a prospective buyer reads before signing tells a firm with a statutory encryption duty that a certificate belongs in the most expensive tier, and the vendor's own portal feature page advertises "Sarbanes Oxley and Gramm-Leach-Bliley compliance" without mentioning multi-factor authentication anywhere on it.

Who needs multi-factor authentication, and what each platform actually says

You do, for anyone who can reach client information, and the wording your vendor uses tells you how much work is left on your desk. The requirement at 16 CFR 314.4(c)(5) is to "implement multi-factor authentication for any individual accessing any information system", and the FTC's guidance restates it as multi-factor authentication "for anyone accessing customer information on your system". The rule defines the factors: something you know, something you have, something you are, at least two of the three.

Here is what the six platforms say on their own security pages, quoted as printed on 25 August 2026. The differences matter because the obligation is yours, not theirs.

That column describes four different postures, not one feature. Enforced, which is what SmartVault claims, means the vendor has made the decision for you. Available, which is the fair reading of Canopy and TaxDome, means the capability is there across user accounts and portals without the page saying it is mandatory. Provided and recommended, which is Financial Cents' own wording, means a setting exists and somebody at your firm has to turn it on, name that decision in your plan, and check it after every new hire. Delegated, which is how Karbon describes it, means your authentication posture is whatever your Microsoft or Google tenant is doing, which may be excellent and may be a shared password from 2019. And not mentioned means ask, in writing, before you renew.

None of these are accusations. Every one of these platforms is used by real firms and several of them are visibly more careful than the average small business tool. The point is narrower: the regulation puts the duty on the firm, so the sentence on the vendor's page is the start of your work rather than the end of it.

The plan you are required to have, sold as a premium feature

You need a Written Information Security Plan, the IRS gives you a template for nothing, and at least one platform puts its version behind its top tier. SmartVault lists "Compliance Vault (WISP templates)" in the feature set of Accounting Unlimited at $85 per user per month, above Accounting Pro at $65 and Business Pro at $55.

Bundling is a legitimate way to sell software and a WISP that is wired into the tool you actually use may well be worth more than a blank document. But you should know what the alternative costs, because the IRS publishes Publication 5708, Creating a Written Information Security Plan for Your Tax and Accounting Practice, as a free template aimed specifically at smaller practices. If a tier upgrade is being justified to you on compliance grounds, the honest comparison is not template against nothing. It is template against a free template plus whatever the tool does to enforce the plan you wrote.

A plan is also not a purchase. The IRS release says tax professionals "are legally required to have a written, accessible plan and should review, test, and update it regularly", and to adjust it when operations change. Changing your website is an operational change. If you launch a new site with a document upload on it, the plan that described the old intake path is now wrong.

Your contact form stopped being a marketing widget

Treat every field on it as a decision about where regulated data will live, because the moment a prospective client types a Social Security number into a message box you have collected customer information. The Safeguards Rule does not care that you had not signed an engagement letter yet.

Three design decisions follow, and they are the ones we argue about with clients in this category. The first is what you ask for. A contact form on an accounting site should collect enough to call someone back and nothing more, because every extra field is a field somebody will paste a document number into. The second is where submissions land. A form that emails the message to a shared inbox has just made that inbox the system of record for regulated data, and the encryption duty at 314.4(c)(3) covers information at rest as well as in transit. The third is the upload button. If you accept documents on the public website rather than behind a portal login, you have created an intake channel with no authentication at all in front of it, which is the opposite of what 314.4(c)(5) asks for.

There is a fourth decision most firms never make, which is deletion. 16 CFR 314.4(c)(6)(i) requires procedures for the secure disposal of customer information no later than two years after the last date it was used to serve that customer, unless you have a business or legal reason to keep it. Form submissions sitting in a website database from 2021 are covered by that sentence. So are the ones in the inbox.

Your web designer became a service provider you have to supervise

The rule makes your website vendor part of your compliance perimeter, and it asks for three specific things rather than a general feeling of trust. Under 16 CFR 314.4(f) you must take reasonable steps to select and retain service providers capable of maintaining appropriate safeguards, require those safeguards by contract, and periodically assess them based on the risk they present.

In practice that means three documents you probably do not have. A note on file explaining why you chose this host or platform, in security terms rather than price terms. A contract clause obliging them to maintain safeguards, which most template web design agreements do not contain. And a recurring calendar entry, annual is fine for a small firm, where somebody reads the vendor's current security page and confirms nothing has quietly changed. If your website is on a subscription platform, the platform is the service provider. If it is a custom build, your host is, and so is anyone with deploy access.

This is also where the ownership question becomes a compliance question rather than a commercial one. If your vendor holds the domain, the hosting account and the form data, then your ability to assess them, or to leave them, is theoretical. We wrote a separate piece on what to check before you sign a web design contract, and for a firm in this category the answer to "who has the keys" now has a regulator attached to it.

What happens when it goes wrong

You call a list of people, and the first federal deadline is 30 days. Since 13 May 2024, 16 CFR 314.4(j) has required a covered firm to notify the FTC "as soon as possible, and no later than 30 days after discovery" of a security event involving the unencrypted customer information of at least 500 consumers. The clock starts on the first day the event is known to any employee, officer or agent other than the person who caused it. The notice has to state what kinds of information were involved, the date range if you can determine it, the number of consumers affected, and a general description of what happened.

Five hundred consumers is a low bar for a tax practice. A single-preparer office doing four hundred household returns can clear it with one compromised laptop, because a joint return carries two people.

The federal notice is also the short part. The IRS's own data theft page for tax professionals, last updated 26 March 2026, lists who else to contact: your IRS Stakeholder Liaison, who notifies IRS Criminal Investigation on your behalf and can move to block fraudulent returns filed with your clients' details, your local FBI office, the Secret Service if directed, local police for a report, the state tax agencies for every state you file in through the Federation of Tax Administrators, and the state attorney general in each of those states, because most states require it. Then a security expert to find and close the hole, your insurer, the credit bureaus, and an individual letter to every affected client.

Set that list beside the difference between a $19 portal seat and a $149 one and the arithmetic changes shape. The controls are not what protect you from the fine. They are what keeps the incident from being a notification event in the first place, because the rule's trigger is the acquisition of unencrypted customer information.

What a five-person firm actually pays in year one

Between about $2,200 and $10,000, and the website is the smallest line in every version of it. Here is the arithmetic on the published rates above, at five seats for twelve months on annual billing.

The cheapest combination in that chart, a flat website subscription at $1,080 plus five Liscio Intelligent Files seats at $1,140, comes to $2,220 for the year. The most expensive, the same website plus five Canopy Premium seats at $8,940, comes to $10,020. The spread is $7,800 and almost none of it is the website.

The two middle bars sit closer together than most firms expect: five TaxDome Pro seats on a one-year term come to $5,000, and five SmartVault Accounting Unlimited seats come to $5,100, which is a difference of $100 across the year for two quite different products. Two costs are not in the chart at all, because they are not list prices. Payment processing, at the rates above, is real money on every return you collect by card. And the person who is responsible: the Qualified Individual required by 314.4(a) is a named human being at your firm who owns this program, and their hours are the largest number on the page even though no vendor invoices for them.

Whether any of it pays for itself

It pays if the site brings in roughly three dozen returns a year, which is a target you can check against your own numbers rather than take on faith. Reporting on the National Association of Tax Professionals' 2026 survey data, CPA Trendlines put the average base charge for a Form 1040 with schedules at $236, up from $162 in the same study series two years earlier. At $236 a return, the entire $7,800 gap between the cheapest and most expensive stacks we priced is about 33 returns.

The demand side is public too. The IRS's filing season statistics through 17 April 2026 record 137,618,000 individual returns e-filed, of which 72,821,000 came from tax professionals and 64,796,000 were self-prepared. Professional filings grew 0.4% year over year while self-prepared filings grew 1.7%, and total returns received fell 0.3%, so by our arithmetic the professional share of e-filed returns slipped from 53.2% to 52.9% in a year. Meanwhile the IRS's preparer statistics, current as of 1 August 2026, count 879,698 individuals holding a preparer tax identification number for the year, among them 208,519 certified public accountants, 68,548 enrolled agents and 26,039 attorneys.

Put those together and the competitive position is clear enough. A slowly shrinking share of a flat market, contested by roughly 880,000 people holding a current preparer identification number, most of whom are found the same way you will be. That is an argument for a site that answers the questions a nervous buyer actually has, which in this profession means credentials, jurisdictions, what happens to their documents, and what the first meeting costs.

Our own prices, and the firms we are wrong for

A build from us is $1,499, charged once, and the continuing option is $2,499 a month for a site that keeps being worked on rather than handed over. Both numbers are on the pricing page because we would rather argue about scope than about price. A focused build takes about two weeks once content and access are in hand, though a firm in a regulated category should assume its own review adds time. We do not sell a client portal and we do not want to: we build the site, and we wire it to whichever portal you have chosen from the list above, which keeps the regulated data inside a platform whose whole business is holding it.

Three kinds of firm should buy something else. A solo preparer with fifty clients and no ambition to grow is well served by a template subscription, and the $90 a month tiers above will do more for them than we will. A firm that wants one vendor, one invoice and one support number for the website, the portal and the tax software should buy a platform, because a stitched-together stack means somebody at your firm owns the seams. And a firm whose real problem is that its Google Business Profile is unclaimed and its service pages do not exist should fix those first, at almost no cost, before commissioning anything.

The firms we do well for look different: a practice that has outgrown a template it cannot edit, a multi-partner firm whose specialties are invisible on a site organized around services nobody searches for, or a rebuild where the current vendor holds the domain. If that sounds like your firm, our work for finance and accounting practices is the closest thing we have to a portfolio in your category.

Map the path a W-2 takes

Before you price anything, spend twenty minutes drawing every route by which a client document can reach your firm, and mark what happens to it. Not the routes in your policy. The routes people use. This is the one exercise that turns the whole regulation into a shopping list, and it is also the input the FTC asks for: the risk assessment at 314.4(b) is exactly this drawing, written down.

Answer the middle column honestly for your own firm and the third column becomes your specification. Every vendor conversation after that is short, because you are no longer asking what a website costs. You are asking whether this vendor can carry six named routes, and what each one adds to the bill.

Do this before you shop and the numbers in this article become useful rather than alarming. Skip it and you will buy a site for $90 a month, feel sensible, and discover in March that the cheapest thing in your firm is the one every client document passes through.

That is the state of pricing information in the veterinary industry, and it is why this article works differently. Every price in it was read on the seller's own page on 24 August 2026, every industry statistic comes from the association or firm that collected it, and where a number is our arithmetic the text says so. We sell websites, so this page is written by one of the vendors you might be evaluating, and you should weigh it accordingly.

The 30-second answer

For a single-location veterinary clinic, a website you own outright is a one-time build in the low thousands: the one vet-specific agency that publishes its fee charges $1,999 to build and $129 a month to run, subscription website vendors run $99 to $299 a month, and our own package is $1,499 once or $2,499 a month for continuous work. The bigger recurring bill is the practice management software underneath, where the vendors that publish prices start between $123 and $260.50 a month and most price per veterinarian, so the bill grows as you hire doctors. Several of the biggest names publish no price at all. Judge every quote against the industry's own numbers: the average veterinary visit ran $147 in 2024 by the AVMA's count, patients average 2.39 visits a year, and industry data shows visit volume falling while prices rise, which makes the case for a site that wins new clients and recalls lapsed ones stronger than it has ever been.

The prices page one quotes are not on the vendors' pages

Before any budgeting, you should know how unreliable the published guidance is, because we measured it. We took the pricing claims in two guides ranking for veterinary software costs and checked each claim against the vendor's own website on the same day, 24 August 2026.

One guide, published by the software vendor VetSyCare in January 2026, carries a table titled 2026 Veterinary Software Pricing Comparison with the introduction, in its own words, "Here's what major platforms actually charge." It prices Digitail at $199 a month and Shepherd at $299 a month. Neither company publishes a price anywhere on its site. Digitail's pricing page lists four plans and ends every one of them in a demo request. Shepherd's site has no pricing page at all. The same table prices ezyVet at $250 and up; ezyVet's own page says $260.50. Another guide, at co.vet, says DaySmart Vet starts at $116 a month; DaySmart's page says $123.

We also ran the test a buyer in 2026 actually runs: we asked an AI-powered search tool what veterinary software costs. Its answer priced Shepherd at $299 a month, quoted a per-veterinarian rate for a platform that publishes no rates, and offered $119 a month for Hippo Manager, the product that no longer exists. AI answers are trained on the guides, the guides copy each other, and nobody in the chain reopens the vendor's page. The habit that protects you costs nothing: no price is real until you have seen it on the seller's own site, dated today.

The $119 product that does not exist

The Hippo Manager story deserves its own section because it shows how long a dead price can outlive its product. Hippo Manager was a budget practice management system that small clinics genuinely used. It was acquired by Shepherd Veterinary Software, and Shepherd's own migration page states the outcome plainly: "Veterinary practices previously using Hippo Manager have now migrated to Shepherd." There is no Hippo Manager to buy.

What happened next is the instructive part. The company's domain, hippomanager.com, was not kept. When we loaded it on 24 August 2026, both the bare domain and the www version returned a page in Indonesian advertising an online slot gambling operation. Meanwhile, buying guides published in 2026 still list Hippo Manager beside its old $119 price as a live option, and AI search tools repeat it. A practice owner who trusts the guide is now one click from a casino page, and a practice owner who trusts the AI answer is budgeting around a product that cannot be purchased.

This matters beyond one vendor. Veterinary software and veterinary website companies get acquired constantly: WhiskerCloud, the best-known vet website agency of the late 2010s, sold to PetDesk in January 2022 in what the acquisition trade publication They Got Acquired reported as a low eight-figure deal covering roughly 2,000 veterinarian clients. When your website vendor is acquired, your site, your content and sometimes your domain are part of the inventory being sold. The section at the end of this article gives you a ten-minute test for that risk before you sign anything.

What the software vendors actually publish

Now the real numbers, starting with the recurring bill, because for most clinics the practice management system costs more per year than the website. Among the major cloud platforms, exactly three publish a price a buyer can read without a sales call.

Provet Cloud is the most transparent of the group. Its Core plan is $99 per veterinarian per month plus a $249 monthly platform fee that includes the first vet, so a one-doctor clinic pays $249 a month and a two-doctor clinic pays $348. The Pro plan for multi-site groups runs $129 per vet plus a $299 platform fee. Its AI scribe is a further $40 per vet per month, and its payments product takes a share of revenue rather than a subscription. ezyVet publishes a single line: every feature included, "as little as $260.50 per month," with implementation priced separately. DaySmart Vet publishes "plans starting at $123/month" and scales by practice size from there.

The unit matters more than the number. Veterinary software prices by the doctor, not by the location. A three-vet practice on Provet Cloud Core is $447 a month before the AI scribe, which at $40 per vet adds another $120, taking the stack to $567. That is the figure to compare against any website quote: the site is usually the smaller line, and unlike the software it does not grow a seat charge every time you hire an associate.

The other half of the market has decided not to tell you. Digitail, Shepherd and Vetspire all route every plan through a demo call, and IDEXX's Cornerstone and Covetrus products have never published list prices. That is their right, but it has a consequence for you: every number you have read about those platforms came from somewhere other than the company, which is exactly how a dead product keeps a live price.

Who sells the veterinary website itself

Unlike the med spa market, where the booking platforms sell no websites, the veterinary market has an ecosystem of companies that sell the website as a subscription, usually bundled with marketing. Their published prices are the most useful benchmark you will find, because they define the floor.

iMatrix, which builds websites and marketing for veterinary, chiropractic and eye care practices, publishes a six-tier grid and prices the first three: Core at $99 a month, Trust at $199 and Build at $299, with the top marketing tiers behind a consultation. Even the $99 Core tier includes a professional website with unlimited edits, an online booking widget, lead capture forms and business listings management, which tells you how commoditized the base veterinary website has become.

Digital Empathy, a veterinary design agency, is the rare vendor that publishes both of its numbers and its competitors' typical ranges in the same place: its own development fee is $1,999 with a $129 monthly service fee, and it puts template website companies at $50 to $100 a month and professional design agencies at $200 to $400 a month with development fees "in the thousands." Those ranges match everything else we verified, so we consider them honest, and we say that as a competitor.

Do the twelve-month arithmetic before choosing a lane. iMatrix Core at $99 a month is $1,188 a year, every year, for a site you rent. Digital Empathy's model costs $3,547 in year one and $1,548 a year after. Our own flat build is $1,499 once, and what you own afterward is yours; we have written before about what ownership actually means when a vendor holds your domain and files. None of these is automatically right. A subscription with unlimited edits genuinely fits a practice with no one to manage a site. A build fits a practice that wants the asset on its own books and its costs to stop.

What a patient is worth, by the industry's own count

Website quotes only make sense against what a client is worth, and veterinary medicine is unusually well measured. The AVMA's 2024 Pet Ownership and Demographic Sourcebook, presented at its Veterinary Economic and Business Forum in October 2024, counted 89.7 million pet dogs in the United States, the highest figure it has ever recorded, with 45.5% of households owning a dog and 32.1% owning a cat.

The money numbers are the ones to pin above your desk. The average veterinary visit cost $147 in 2024 across all pets seen, down from $190 in 2023. A dog visit averaged $214 in 2024 against $265 in 2023, and a cat visit $138 against $163. Dog-owning households spent an average of $580 a year on veterinary care, cat-owning households $433. And the visits are overwhelmingly the kind a website wins: about 80% of dog and cat visits are routine checkups or preventive care, the appointment a new resident books after searching for a clinic nearby.

Two more industry findings sharpen the picture. VetSource data presented at the AVMA's 2023 forum shows the average patient visits 2.39 times a year, so at 2024 visit prices one dog on the books is worth roughly $511 a year in visit revenue alone, our arithmetic from the two published figures. And iVET360's 2026 Veterinary Industry Benchmark Report, published in April 2026, found that 2025 industry revenue grew 2.6% while transaction volume fell 4.7% and the average transaction charge rose 7.5%, with new client acquisition running below historical norms. Growth is coming from charging more per visit, not from seeing more patients. A practice that wants volume growth has exactly two levers a website touches: winning new clients, and recalling the lapsed ones the AVMA's 2023 forum flagged as a rising share of every practice's file.

The gap between having a vet and visiting one

Buried in the AVMA's 2024 numbers is the most commercially useful statistic in this article. Among dog owners, 86.8% say they have a regular veterinarian, but only 74.2% visited one in the past year. For cat owners the gap is wider: 77.1% claim a regular vet, 57.3% actually went. Roughly one dog owner in eight and one cat owner in five thinks of themselves as your client and did not show up last year.

That gap is not a marketing abstraction, it is deferred revenue sitting in your practice management system. The AVMA's economists have tracked the same pattern from the other side: practice revenue rose 5.7% between August 2021 and August 2023 while visits fell 2.7%, and the share of owners telling Cleveland Research they planned to visit the vet less to save money nearly doubled from 10% to 19% between mid 2022 and mid 2023. Price-sensitive owners defer, deferral becomes lapse, and a lapsed patient is the cheapest appointment your website can win back, because the owner already trusts you. This is why the reminder and booking layer of the site matters more in veterinary medicine than in almost any other local business, and why we treat online booking as a first-class requirement rather than an add-on when we build for clinics; the same logic drives the local service businesses we work with, where the website's job is measured in booked appointments, not visits.

What a custom build should cost, and what ours costs

Against those benchmarks, here is the honest range for the build itself. A template subscription runs $50 to $100 a month with no build fee and you own nothing at the end. A vet-specific subscription with marketing folded in runs $99 to $299 a month at published rates. A professional agency build runs a development fee in the low thousands, $1,999 at the one agency that publishes it, plus $129 to $400 a month in service fees. Ours is priced flat: $1,499, one time, the whole build, or $2,499 a month for practices that want design and development on tap rather than a single launch. A focused build usually goes live about two weeks after kickoff, a typical figure rather than a promise, and we answer partnership inquiries within 48 hours.

Two outcomes we can point to, both from adjacent industries where the same trust problem applies: a treatment center rebuild that produced 20% more website conversions for Cornerstone Healing Center, and Eden Digital, a brand we built that grew revenue tenfold. We have not built for a veterinary clinic yet, and we would rather tell you that than imply otherwise; the mechanics of a clinic site, service pages, booking paths and local search, are the same ones we ship every week, but if vet-specific pedigree matters to you, iMatrix and Digital Empathy have it and we do not.

The pricing conversation is simpler than vendors make it. As with med spas, the recurring software bill dwarfs the website over time, so the website decision is really a decision about which costs you want to stop. A $299 a month subscription is $3,588 a year forever. A $1,499 build with $30 hosting is $1,859 in year one and $360 a year after. Across five years that is $17,940 against $3,299, so the build costs less than a fifth as much; full comparison logic is on our pricing page.

Where the money should go, in order

If you are budgeting a clinic site from zero, spend in this order. The reasoning is the AVMA arithmetic above rather than preference.

• Service pages that match how owners search. One page per service you want more of: wellness exams, dental, spay and neuter, urgent care, exotics if you see them. Roughly 80% of visits are routine care, so the wellness and vaccination pages are the revenue pages, not the surgery page you are proudest of.
• The booking path. Whatever your practice management system is, its booking widget must sit on your domain, two taps from any service page on a phone. If a vendor cannot embed it, that constraint belongs in your brief before anyone quotes.
• The lapsed-client path. A visible existing-clients entry: request records, refill a prescription, book a recheck. One owner in five who considers you their vet did not visit last year; make returning frictionless.
• Local search plumbing. Accurate structured data, service-area pages if you draw from several towns, and a Google Business Profile that matches the site. For a clinic, ranking in the map pack is worth more than any blog.
• Real photography. Your building, your exam rooms, your team with animals. Stock photos of a model in a lab coat read as a franchise, and trust is the product.
• Everything else. Team bios, a blog, a careers page. Useful, rarely urgent, and the padding a page-count quote hides in.

Who should not hire us

A brand new single-doctor practice with no photography, no service list and an opening date six weeks out will get more from a $99 a month vet-specific subscription than from any custom build, ours included, because the bottleneck is content, not code. A practice whose real problem is phones going unanswered needs staffing or a phone tree before it needs a website. And if what you want is someone to run your Google and Meta ads every month, hire a veterinary marketing agency; iMatrix's upper tiers exist for exactly that, and it is not our product.

The practices that get their money's worth from us look different: a real service list that ranks for nothing, a subscription template the team cannot edit, a rebuild where the current vendor holds the domain, or a multi-location group whose clinics cannot be found separately. If that is you, the numbers are public and the first conversation costs nothing.

The ten-minute vendor-mortality test

Every price in this article could change tomorrow, and as Hippo Manager shows, so can the existence of the company charging it. Before you sign with any website or software vendor, run these four checks. They take ten minutes and they would have protected every practice caught in the stories above.

• Look up who owns your domain, right now. Run a WHOIS lookup on your own domain. If the registrant is your current website vendor rather than your practice, fixing that is the first line of any new contract. When a vendor dies or sells, domains they hold go with them, and hippomanager.com now points somewhere no client should land.
• Ask what happens on acquisition, in writing. One sentence in the contract: if the vendor is acquired or discontinues the product, you receive a full export of the site files, content and data, and the domain transfers to you within 30 days. A vendor who resists this sentence is telling you something.
• Get the price in writing, dated. Not a screenshot of a guide, not an AI answer: an email from the vendor with the number and the date. Two of the five prices we checked this week were wrong and one was for a product that no longer exists. Vendors honor their own dated emails; they do not honor co.vet.
• Test the export before you need it. Ask for a sample export of your site content or your patient data during the sales process. If they cannot produce one for a prospect, they will not produce one during a shutdown.

Do those four things and the worst outcome of any vendor's death is an inconvenient month. Skip them and you are betting your practice's front door on a company you found through a guide that quotes dead prices. In the time it took to research this article, one vendor's domain became a casino and two published prices turned out to be fiction. Your website should be built to survive its builders. Ours are, and so should everyone's.

There is a better way to answer this, and it is available because the aesthetics industry is unusually generous with numbers. The software companies your practice already runs on publish their prices. The industry association publishes what a med spa earns per visit and what a new patient costs to acquire. Put those two sets of figures next to each other and the question stops being a guess. Prices here come from the seller's own page, statutes from the statute, and the industry numbers from the association that collected them, all checked on 23 August 2026. Where a number is our arithmetic rather than somebody else's, the text says so. We sell websites, so this is a page written by one of the companies you would be paying, and you should read it in that light.

The 30-second answer

For a single-location med spa, a website you own outright is a one-time build in the low thousands, and the recurring money goes to the practice software underneath it rather than to the site. Published list prices for that software run from about $75 a month for a five-person practice on Aesthetic Record up to $520 a month for Boulevard's full aesthetics bundle, and at one vendor the entry price for a med spa is more than two and a half times the entry price for a salon. None of the major booking platforms sells you a website; they sell booking that sits on one. Agencies that publish prices put a med spa build between $2,000 and $10,000, rising past $15,000 once a site passes about fifty pages. Our own package is $1,499 once, or $2,499 a month when the work is continuous. Judge any of those against the two numbers the industry publishes: an average visit worth $527 and a new patient costing $132 to acquire.

What page one actually publishes

Before the numbers, a word on why they are so hard to find. We took the five pages ranking for med spa website cost on 23 August 2026, pulled the body text of each, and counted two things: how many distinct dollar figures each page states, and where each page sends you for evidence.

Between them the five pages state 78 distinct dollar figures. Not one of the five links to a vendor price list, a survey or a dataset. The page carrying the most numbers, 37 of them, is also the only one that links to an industry analysis at all, and that link sits beside a claim about marketing budgets rather than beside any of its prices. We have not named the five, partly because one of them publishes its own rates honestly enough to be quoted later in this article, and partly because the pattern matters more than the culprits.

Part of the reason is structural rather than lazy. The authoritative dataset for this industry is the American Med Spa Association's Medical Spa State of the Industry Report, and AmSpa sells it for $995 as a one-time purchase, free only to its Plus members. When the good numbers sit behind a four-figure paywall, the pages competing for a free search result fill the gap with estimates, and the estimates get copied.

What the software actually lists at

Start with the recurring bill, because for most practices it is larger than the website and it arrives every month forever. Three platforms dominate quotes in this space and all three publish prices. Their pricing units differ, so the chart below converts each to the same case: one location, five people who need a login.

Read across the bars and the pricing models matter as much as the totals. Aesthetic Record charges per seat, $15 on Essentials and $19 on Accelerator, so five seats come to $75 or $95 and the bill scales linearly as you hire. Mangomint splits the difference with a $120 base and $10 per user, which puts a five-person practice at $170. Boulevard prices the location rather than the people, which is why its salon Essentials plan lands at $176 for the same practice and stays there as you grow. If you are planning to double your team, the per-seat models are the ones to model out before you sign.

Two of those bars deserve their own section, because the gap between them is not about features.

The word that doubles the bill

Boulevard's pricing page has a toggle above the plan cards. Set it to Salon and Spa and you are offered three plans: Essentials, Premier and Prestige, at $176, $293 and $410 a month per location. A limited-time discount was running on the day we looked, and every price in this section is the normal one printed on Boulevard's own cards beside the offer. Set it to Medspa and the three plans disappear. You are offered two, the Aesthetics Starter Bundle and the Aesthetics Bundle, at $455 and $520 a month per location on monthly billing, or $410 and $468 if you commit to a year.

Read those two lists next to each other. The cheapest plan a med spa can buy costs more than the most expensive plan a salon can buy. The entry tier is gone entirely, and the discount for paying annually only brings the med spa floor down to the salon ceiling. Whatever the extra software does, the pricing tells you plainly that this vertical is expected to pay, and that expectation will follow you into every quote you request, including ours.

This matters for a website budget in a specific way. If your booking platform already takes $455 a month, a $399 a month website subscription on top of it is $854 a month, or $30,744 over three years, for a practice that could have bought the site once. That is the trade the rest of this article is really about.

The lines that are priced separately

Every platform in this market prices a base plan and then meters the things a med spa actually does daily. These are all published rates, not estimates, and they are the reason a quoted monthly figure and an actual monthly bill diverge.

The two largest bars are the ones practices underestimate. A phone line with calling and texting is $70 a month per line at Mangomint, and payroll starts at $50 a month plus $8 per worker, so a five-person practice running both adds $160 a month before anyone has sent a marketing message. Neither is optional in the way an integration is.

Three of those deserve a note. Client forms are not a nice extra for a med spa: intake, consent and medical history are the paperwork the treatment cannot legally happen without, and Boulevard prices them from $65 a month per location. Texting is how aesthetics practices confirm appointments, and it is metered almost everywhere; Boulevard's plans include 100, 250 and 2,500 free texts and charge a cent per email beyond the allowance, while Mangomint states plainly that unlimited text and email appointment confirmations and reminders are included at no extra cost. And card processing is a percentage, not a line item: Mangomint publishes 2.45% plus 15 cents and 2.90% plus 30 cents, while Boulevard offers a program in which the client pays a 3% fee on card transactions and the practice pays 1%.

Which of these hands you a website

None of them. That one word is worth more to a budget than any other sentence in this article, and a demo will never surface it.

Boulevard, Mangomint and Aesthetic Record all sell booking, charting, payments, memberships and marketing. Read their own feature lists and you will find online booking, a patient portal, eCommerce and gift cards. You will not find a website, because a website is not what they make. Mangomint says so directly in its own pricing FAQ: its online booking sits on your website, and you keep every appointment booked through it. That is a good arrangement, and it is also an admission that the website is your problem.

So the practice needs two purchases, and the mistake we see most often is buying one and assuming it covered both. The booking platform is the machine. The website is the shop it sits in. Whoever sells you the second one is quoting for design, content, treatment pages, the consent and intake flow, and the wiring that connects your menu to the booking widget without dumping the patient onto a different company's domain halfway through.

What agencies publish for the build

Very few will tell you before a call. Of the med spa specialists we checked, most of the pricing pages either do not exist or return a contact form. Three exceptions are worth reading.

SERPreme publishes the clearest tiering in the field. Its medical spa page is titled around a $2,000 quick launch, and its own FAQ states that medical spa website projects typically run $4,000 to $10,000 or more depending on whether it is a build or a makeover, and that a site already fifty pages or larger is closer to $15,000 to $20,000. It also publishes a monthly marketing range of $1,000 to $8,000, with most of its clients between $2,000 and $4,000.

GMR Web Team answers the cost question on its own med spa page with a single number: $399 a month buys a medical spa website with patient intake forms, hosting, reputation management and acquisition optimization, with content creation charged extra.

Growth99 will not publish a plan price, but it does something almost as useful: it publishes scope. Its Accelerator plan includes a custom website of up to 15 pages, Elite goes to 25 pages, and both specify one opportunity for layout changes and two rounds of revisions on image selection. That is the most honest description of a productized website we found in this market, and you should ask every vendor for the same three numbers.

Our $1,499 is a flat package price rather than a starting point, and it exists because a single-location med spa with a clear treatment menu does not need a fifty-page site. The $2,499 a month is a different product for practices where the work does not stop at launch. Both are published on our pricing page for the same reason we are publishing everyone else's here.

What the industry actually spends

The best free dataset on this question comes from AmSpa itself. In an article published on 9 January 2026, drawing on its own data and on Growth99's 2026 State of Aesthetic and Elective Wellness Marketing Report, the association reported that 52% of med spa practices still invest less than $2,500 a month in marketing, and that only 25% meet or exceed $5,000. It puts the working benchmark at 5% of revenue, and the average practice at $1.39 million a year, which makes the benchmark roughly $5,800 a month.

The two published bands leave 23% of practices unaccounted for, and by subtraction that is the group sitting between $2,500 and $4,999 a month. Nobody publishes that band directly, so treat it as arithmetic rather than a finding, but it does tell you the shape of the market: a large low-spending majority, a small high-spending minority, and a thin middle.

Two honest caveats. That spend covers every channel, so a practice at $2,500 a month is buying ads and content as well as anything a website costs. And the benchmark is a rule of thumb rather than a finding. What the chart is good for is scale: if half the market runs on under $2,500 a month across everything, a website subscription at $399 a month is consuming a sixth of the entire marketing budget before a single ad runs.

What the website has to be worth

The same AmSpa article publishes the three numbers that turn this into arithmetic instead of taste: an average cost per lead of $39, an average new patient acquisition cost of $132, and an average visit value of $527. It also reports average profit margins of 38%.

Take the $527 visit at a 38% margin and one additional visit contributes about $200 of profit. A $1,499 build has to produce roughly seven and a half extra visits, once, over the entire life of the site. A $399 a month subscription has to produce two extra visits every month, forever, and it has to keep producing them in the months when nothing about your site has changed.

Do the same on the acquisition side and it gets sharper. At $132 per new patient, a site that lifts booking conversion enough to save you thirty acquisitions a year has returned about $3,960 a year in avoided cost, before any of those patients books a second time. That is the honest case for spending on a website: not that it conjures demand, but that it stops you paying twice for traffic you already bought.

Your gift card page is a legal document

Now the parts of a med spa site that cost more to build than the equivalent page on a salon site, starting with the one that surprises people.

Gift cards and prepaid packages are a serious revenue line in aesthetics, and every platform in this market sells them. Aesthetic Record lists eGift cards, discounts and package creation, and monthly treatment subscriptions among its base features. The moment you sell those on your own website, the product copy on that page is governed by federal law.

Under 12 CFR 1005.20, the funds on a store gift card have to stay valid for at least five years from the date of issue or the date money was last loaded. A dormancy, inactivity or service fee is allowed only if there has been no activity for a full year, and only one such fee may be charged in a calendar month. The rule carves out loyalty, award and promotional cards, and cards not marketed to the general public, so a free birthday credit is treated differently from a card someone bought.

State law can be stricter, and in the biggest med spa market it is much stricter. California Civil Code 1749.5 makes it unlawful to sell a gift certificate to a purchaser that contains an expiration date at all, with narrow exceptions for promotional and donated certificates that must print the date in capitals of at least 10 point on the front. The same section requires a gift certificate with a cash value under $15 to be redeemable in cash.

So the line reading "valid for 12 months" that sits under most med spa gift card products is, in California, on the wrong side of a statute, and everywhere else it is at odds with a five-year federal floor. Fixing it is not expensive. Discovering it after you have sold four hundred of them is. This is what people mean when they say a med spa site costs more than a salon site: not that the pixels are harder, but that more of the copy is regulated.

Two words that need a co-star

The second regulated element is the phrase practices most want in their hero section. In California, Business and Professions Code 651 makes it unlawful for a licensee to disseminate any public communication carrying a false, fraudulent, misleading or deceptive statement, claim or image, and it then lists what an advertisement may contain. Subdivision (h)(5)(C) deals with board certification: a physician may use the term board certified only where the full name of the certifying board appears with comparable prominence, and the board has to be an American Board of Medical Specialties member or an equivalent approved before 1 January 2019.

Comparable prominence is a design instruction. It means the certifying board's full name cannot be relegated to the footer or set in gray eight point under a 64 point headline. If your hero says board certified, the board's name belongs in the same visual breath. We have written before about how testimonial and review copy carries similar constraints, and about the patient data question that sits behind any intake form. The pattern is consistent: on a medical site, the copy deck is part of the compliance surface, and it is cheaper to build it correctly than to rewrite it later.

Where the money should go, in order

If you are budgeting from scratch, this is the order we would spend in, and the reasoning is the arithmetic above rather than preference.

• The treatment pages. One page per treatment you actually want more of, each carrying what it is, who performs it, what it costs or how pricing works, and a booking link. This is the only part of the site that ranks and the only part that converts.
• The booking path. Two taps from any treatment page, on a phone, without leaving your domain. If your platform's widget cannot embed, that constraint belongs in the brief before anyone quotes.
• Intake and consent. You are paying for forms somewhere already. Decide whether they live in the platform or the site, and make sure only one of them is collecting the same medical history.
• The gift card and membership pages. Revenue, and the two places where the copy is regulated.
• Photography. Real rooms, real staff. Stock photography of a model who does not work at your clinic is the fastest way to make a good site read as a franchise.
• Everything else. Blog, team bios, an about page. Useful, rarely urgent, and the first thing a vendor will pad a page count with.

Our offer, and the practices we are wrong for

Our package is $1,499 as a flat one-time price, or $2,499 a month when a practice needs continuous work rather than a launch. A focused build usually takes about two weeks from kickoff, though that is what typically happens rather than a promise. Partnership inquiries get an answer inside 48 hours. Two outcomes we can point to: a treatment center rebuild that ended with 20% more website conversions for Cornerstone Healing Center, and Eden Digital, which we built and which grew revenue tenfold. More on how we work with clinical practices sits on our healthcare page.

Now the part that costs us work. A brand new single-room practice with one injector, no memberships and no gift cards does not need a custom site yet, and $399 a month for a hosted med spa website with intake forms folded in is a defensible purchase while you find out whether the business works. If you are opening in six weeks and have no photography, no treatment menu and no price list, a template will beat a custom build because the bottleneck is your content, not your code. And if what you want is somebody buying and managing your Google and Meta ads every month, hire a med spa marketing agency, because that is their product and it is not ours.

Where we are the right call is narrower and easier to describe. A practice with a real treatment menu that ranks for nothing. A multi-location group whose locations cannot be found separately. A rebuild where the current vendor holds the domain. A practice that has outgrown a booking-platform microsite and needs the site to carry the explaining that the booking widget cannot.

Write the spec before you ask for a price

Here is the thing nobody tells you about collecting quotes: the reason three quotes for a med spa website come back at $2,000, $8,000 and $399 a month is not that one vendor is honest and two are not. It is that all three were asked a different question, because you did not write the question down. Fix that and the quotes become comparable in an afternoon.

Write one page. Not a brief, not a strategy document, one page, and send the identical page to everybody. Six lines is enough.

• Page count, named. Not "a small site". List the pages: home, six treatments, about, contact, gift cards, memberships. Fourteen. Now every quote is pricing the same object.
• The booking platform you use, by name, and where the widget must sit. On your domain or theirs. This single line moves more money in a quote than anything else on the list.
• Who writes the words. If you are not writing the treatment pages, somebody is billing for them, and copy is the item most often quietly excluded.
• Revisions, as a number. Growth99 publishes one layout change and two image revision rounds. Ask every vendor to state theirs the same way, before you sign rather than after.
• What you are handed at the end. The domain, the hosting account, the files, the accounts. State that you keep all four.
• What happens in month two. Who edits a price, who adds a treatment, and what that costs. A build price with no answer here becomes a subscription by accident.

Send that page to four vendors including us. The quotes that come back will differ by thousands, and for the first time you will be able to see exactly which line each difference is hiding in. That is worth more than any range on any page you have read today, including this one.

Search the question and page one hands you confident ranges that do not agree with each other. Elementor's guide puts a small business site at one to two weeks if you do it yourself and four to eight weeks with a professional. WPX gives four different bands in a single article. Hostinger's tutorial breaks a WordPress business site into five phases totaling five to fourteen weeks, and an enterprise build at thirteen to twenty two. None of them cites a measurement. They are estimates, written by companies that sell hosting and page builders, presented in the tone of facts.

A better set of numbers is sitting in public, unquoted by any of them. The companies that sell finished websites publish the launch time they are willing to be held to, on the same page as their prices. Underneath those promises sit a handful of external clocks with published or measurable durations that nobody on the project can shorten. Put the two together and you get a timeline you can actually plan a launch date around. Nothing below is estimated: each figure was either quoted from the page of the company that published it or measured on this machine, on 21 August 2026. One disclosure before you start. KhanWork sells the work being timed here, so treat the section near the end about our own process as the least neutral part of the article.

The 30-second answer

Most small business websites take two to eight weeks of calendar time, and the companies selling them say so on their own pricing pages. Placester advertises a template site live in minutes. Brighter Vision and Websites for HVAC both publish two to three weeks. FMG Suite publishes four to six and names approvals as the reason. The spread has almost nothing to do with how fast anyone builds. It tracks how much of the site is written before work starts and how many people have to approve it. Under all of that sit external clocks, from a 48 hour DNS cache to a 60 day domain lock, that no agency can speed up.

Why no two published answers agree

They disagree because none of them is measuring anything. Each range is an estimate written for an audience, and once you line them up the seams show.

WPX is the clearest example, because the contradiction is internal. Its 2026 article says a professional developer can build a basic informational site in four to eight weeks. A few paragraphs later it puts a standard agency project for a small business at six to ten weeks. Under its heading for business websites, a small or medium business on WordPress is typically four to six. At the end, a moderately experienced developer using WordPress is given four to twelve. Four bands, one article, for a project most readers would describe the same way. Different delivery routes explain some of that, but not the last two, which describe the same route and differ by a factor of two.

Hostinger's tutorial is more careful, and it is the only one that shows its work by publishing phase lengths. That is also how you find the arithmetic. It puts a custom coded single page site at eight to twenty one days, then lists the phases: planning one to three days, design and prototyping two to seven, development three to ten, testing one to three, deployment one. The low ends add to eight, which matches. The high ends add to twenty four, not twenty one. It is a small thing, and it is exactly the kind of small thing that tells you the total was chosen and the phases were written to fit around it.

Elementor's table is internally consistent and worth quoting for the shape rather than the numbers. A simple brochure site of one to five pages: one to three days do it yourself, one to two weeks professional. A small business site of five to fifteen pages: one to two weeks, then four to eight. A custom web application: six to twelve months or more. All three guides agree on one thing about small sites: hiring somebody makes the project take longer, not shorter. That is worth sitting with for a second before anyone explains it away.

What the sellers put in writing

The numbers worth trusting are the ones a company has to live with. A website vendor that publishes a launch time on its pricing page has customers who will hold it to that, which makes those figures a different class of evidence from an estimate in a blog post. Six of them are plotted below, read from the vendors' own pages on the day this was written.

Placester, which sells real estate agent sites, tells buyers on its pricing page that they can "Launch your modern, IDX-ready website in minutes with AI". Brighter Vision, which sells to therapists, answers the question in its own FAQ: the whole process can be as short as two to three weeks, and longer if the client spends more time in the design phase. Websites for HVAC publishes a week by week schedule: day one you sign up and your territory is locked, weeks one to three they build, week three the site goes live, and management billing starts at launch rather than at signing. FMG Suite, selling to financial advisors, offers turnkey designs and pre-written copy and still puts launch at four to six weeks, which its page attributes to approvals and to how much the client wants personalized.

Read that list again and notice that they are not describing the same product. The minutes version is a template filled in automatically. The two to three week versions are template families with your content dropped into them. The four to six week version has a compliance step in the middle. The honest summary is that the timeline is a description of how much judgment is being applied, not of how fast anyone works.

The sentence that explains the whole spread

One vendor states the mechanism out loud, and once you have read it the rest of the market becomes legible. My Personal Trainer Website answers the question in two sentences: the framework can be ready within 24 hours, and adding content, images and integrations usually takes two to four weeks.

One day against fourteen to twenty eight of them. The software part of a website, the part people picture when they imagine somebody building it, is a day. Everything after that is deciding what the site says, finding the photographs, agreeing the wording of the service descriptions and connecting the booking or payment tools to accounts that somebody has to create. That is where the weeks live, and it is why two agencies quoting the same project can honestly say two weeks and eight weeks. They are describing different amounts of unfinished thinking.

The same page makes the point again from the other direction. Asked whether the client has to supply the content, the answer is no, because on a setup plan they will write the copy and source the images themselves. That is the fastest lever anyone has ever found for a website project, and the reason it works is that it removes the client from the critical path rather than making anyone type faster.

Building is a minority of the calendar

Even in the estimates published by companies that sell development, writing code is not where most of the time goes. Hostinger's enterprise WordPress breakdown is the useful one, because it publishes each phase rather than a single total.

Writing code accounts for 4 of the 13 weeks at the fast end of that range and 6 of the 22 at the slow end. Call it a third of the project at best. The rest is planning, design and setup at the front, testing and deployment at the back. For a small business site the proportion is usually worse, because the build shrinks while the deciding does not: a five page site still needs five pages of decisions, and those take about as long to make as fifteen pages of decisions did.

This matters when you are comparing quotes. Where one agency says six weeks and another says three, ask each of them what happens in week one. A discovery call, a content questionnaire and a wireframe round means you are buying the deciding. A staging link on Thursday means you are buying the template. Both are legitimate purchases. They are not the same purchase, and the week difference is the tell.

The part of the clock that belongs to you

Vendors are unusually specific about this when they are trying to reassure you, and the numbers give the game away. Brighter Vision tells prospective clients that most of them invest about 60 to 90 minutes of their own time in the process. That is the total, for a site the same page says takes two to three weeks.

Ninety minutes of work spread across three weeks is not a workload. It is a queue. The client's ninety minutes arrive in six or seven fragments, each one blocking everything behind it, each one waiting on somebody to open an email. Which is why the same FAQ warns that more time spent in the design phase stretches the project, and why FMG Suite hangs its four to six weeks on approvals, and why the Websites for HVAC schedule gives the client exactly one job during the three build weeks, which is to review and approve.

The practical version: if you can put one person in charge who is allowed to approve things without a meeting, you will beat the published range. If approval requires three partners and a board, you will not, and no agency selection will change that.

The clocks nobody on the project controls

Some of the elapsed time in a website launch is not work at all. It is waiting for systems that run on their own schedules, and those schedules are published or measurable. Individually they are short. What makes them expensive is that most get started on the day somebody remembers them, which is usually launch week.

Two of those bars were measured rather than quoted, and between them they cover everything people mean when they say a launch is waiting on DNS. The rest come from the documentation of the organizations that run them. Taken together they set a floor: no matter how fast the site is built, a business that also needs a verified Google listing and a domain moved to a new registrar is looking at weeks of external process running alongside the project.

What DNS propagation actually measures

DNS propagation measures a cache timer that somebody chose in advance, and you can look up the number instead of guessing at it. The folklore says 24 to 48 hours. The folklore is describing one specific case and applying it to all of them.

Every domain publishes a time to live on each record, in seconds, and that value is the longest a resolver anywhere is allowed to keep serving the old answer. We queried a handful of well known domains at their own authoritative nameservers on 21 August 2026, and separately queried the .com registry servers for the delegation records that point at those nameservers.

That last row is the whole story. Changing where a site is hosted, by pointing an A record somewhere new, is a one minute to one hour operation on every domain we sampled. Changing which nameservers are in charge of the domain is a 48 hour operation, because the .com registry stamps every delegation with a two day cache time and there is nothing you or your host can do about it. The 48 hours in the folklore is real. It belongs to one specific action, and most launches do not need to take it.

The certificate worries people more than it deserves to. The live certificate on this site, read straight off the connection on the day of writing, was issued by Let's Encrypt on 5 July 2026 and expires on 3 October 2026, a 90 day window renewed automatically. Issuing one takes minutes and no human attention. There is one way to turn it into a real delay: Let's Encrypt caps issuance at 50 certificates per registered domain every 7 days, so a launch that repeatedly fails validation and retries can lock itself out of new certificates for a week.

Launch day is not the day Google agrees

The site goes live in an afternoon. Google catches up on its own schedule, and Google says so plainly rather than leaving it to be discovered.

"Crawling can take anywhere from a few days to a few weeks," the page says, and a note directly under it adds that asking for a crawl guarantees neither speed nor inclusion at all. There is no queue to pay for and no button that helps.

Speed reporting runs on a longer clock still. The Chrome UX Report, which is the field data behind Core Web Vitals and behind the "real world" half of most speed tools, describes itself as "a 28-day rolling average of aggregated metrics". Its API documentation adds the detail that catches people out after a rebuild: the reported collection period always reads as 28 days even when the page has existed for fewer than 28, so a site that got dramatically faster last Tuesday will keep reporting a blend of the old site and the new one for four more weeks, and anyone judging the rebuild on those numbers in week two is reading mostly the site you replaced.

If the project is a rebuild of an existing site rather than a first one, the ranking mechanics deserve their own attention and we wrote them up separately in the guide to redesigning without losing your Google rankings. The timeline point stands either way: the last third of a website project happens after the launch party.

Queues you cannot jump

Several of the things a new business needs alongside a website are queues run by third parties, and each one has a published waiting time. They are worth listing because they are the ones that turn a two week build into a two month launch.

The Google Business Profile. For a local business this is often more valuable than the website itself, and it is the slowest item on the list. Google's help page says most postcard verification codes arrive within 14 days, and that once the code is entered the review itself can take up to five business days. It also warns that editing your business name, address or category while the postcard is in transit invalidates the code, which means an impatient rebrand mid-project restarts the clock.

The domain. Under the ICANN Transfer Policy a registrar is entitled to refuse to release a domain during three separate 60 day windows: within 60 days of the creation date (section 3.7.5), within 60 days of a previous transfer (3.7.6), and during the 60 day lock that follows a change of registrant (3.9.4). Register a new domain today at the wrong registrar and you may not be able to move it until well after your intended launch. Which registrar it lives at, and in whose name, is a decision that belongs at the start of a project rather than the end, and it is bound up with who actually owns the site when the work is finished.

The regulator, if you have one. Regulated firms have review clocks written into rules. A member firm in its first year of FINRA membership must file retail communications with the Advertising Regulation Department "at least 10 business days prior to first use", which for a website means the copy has to be final two weeks before the site can be published. Other regulated trades have their own advertising rules and their own reviewers, so the question to settle early is who has to see the copy before the public does.

Anything with a payout. Payment processing, appointment tools and anything holding customer data has an onboarding step with identity checks. No single step is long. The steps run one after another, though, and each one needs somebody with the legal authority to sign.

What actually makes a website project late

Projects run late for reasons the vendors themselves name in their FAQs, and none of them is technical. Brighter Vision blames extra time in the design phase. FMG Suite blames approvals and personalization. My Personal Trainer Website's fast path is the one where they write the copy for you. Read three sales pages and the same admission is in all of them.

In our own work the pattern repeats. The projects that overrun do so because the photographs were never taken, because the service descriptions turned out to be a genuine disagreement inside the business rather than a writing task, or because the one person who could approve the homepage was traveling. The build sits finished, waiting on a paragraph.

The second pattern is the queue nobody started. A business that decides in week three that it also wants a Business Profile, a new domain at a different registrar and a payment processor has added a month of external waiting to a project that was on schedule.

What genuinely compresses the calendar

Four things move the date, and they all move it before anyone opens a design tool.

Write the content first, or hand it over entirely. Every vendor quoted in this article agrees on this one, which is rare enough to be worth noticing. A site whose copy exists on day one can be built in the time the tools take.

Cut the page count honestly. Six pages that are finished beat sixteen that are half written. Pages can be added after launch at no penalty, and a smaller launch also means fewer approvals.

Name one approver. Not a committee, not a founder who reviews things on Sundays. One person with authority and a calendar.

Start every external clock in week one. Claim the Business Profile, confirm which registrar holds the domain and in whose name, lower the DNS record time to live to a minute a week before cutover, and get any regulatory filing moving while the design is still in progress. None of these depends on the site being finished, and every one of them is a week you will not spend waiting later.

How we run a build, and who should look elsewhere

Our own typical for a focused build is about two weeks from start to live, and typical is doing real work in that sentence. It holds when the content is ready or when we are writing it, when one person approves, and when the domain and the listings are already where they should be. It does not hold otherwise, and we would rather say that here than discover it with you in week five. Across 200 and more projects, the ones that hit the date were never the ones with the fastest developers on them.

What we sell is a one-time build you own outright, or a monthly partnership for businesses that keep changing the site, with the scope and the price agreed on a call rather than published as a package. The offer page lays out what each one covers. A request from a new business gets a reply from us inside 48 hours, which is the only date we will promise before we understand the project.

Now the awkward half. If your content genuinely does not exist and you do not want anybody writing it for you, a two week build is not available from us or from anyone, and paying a premium for speed you cannot use is a bad trade. Anyone who needs something live this afternoon to take bookings should buy a template from one of the platforms in the builder comparison and move later. If you are choosing between agencies on timeline alone, the checks in our guide to choosing an agency will serve you better than another quote. And if you run a service business where the Business Profile is doing most of the work, start that verification today and treat the website as the slower half of the project, which is roughly how it looks from the local services side of our work.

Everything with a clock, and when to start it

Keep this one open while you plan. Every row below is a separate clock, with a different owner, a published or measured duration, and a moment at which it should be set running.

Pick the date, then count backwards

Most website plans are built forwards, which is why they slip: the launch date is whatever is left after everything else has taken as long as it took. Try it the other way. Put a real date on the wall, one that matters to the business, then walk backwards through the clocks above and write in when each one has to have started.

Counting back from launch day, on the evidence in this article: the day before, the certificate should already exist and the record time to live should already be low. One week before, the content is final and any regulatory filing is already sitting with the reviewer. Two weeks before, the build has everything it needs and nothing is waiting on a photograph. Three weeks before, the Business Profile postcard is either in hand or in the mail. Eight weeks before, the domain question is settled, because a domain registered at the wrong registrar today cannot be moved for 60 days.

And then two dates after launch, which almost nobody writes down: two to three weeks out, when the new pages should be crawled and it is fair to ask why any of them are not, and 28 days out, which is the first morning the field speed data describes the site you actually built.

There is a better question, and the platforms answer it themselves, in public, on the same pages that advertise the monthly fee. An online store does not have one price. It has a fixed part and a part that moves with your sales, and the moving part is written in percentages. Add those percentages up on a single order and the figure is larger than most owners expect. Follow them upward and you reach published thresholds where the bill changes without anybody asking you first.

Every number below was read on 20 August 2026 from the page of the company that charges it. Where a figure is calculated, the arithmetic is shown. We build ecommerce sites for a living, so this is a page written by one of the suppliers you would be pricing, and it is fair to read it that way.

The 30-second answer

Two numbers decide what an online store costs and only one of them appears in a quote. The fixed part is the build, the platform subscription and hosting: on annual billing Shopify runs $19 to $299 a month, BigCommerce $29 to $299, Squarespace $12 to $25, and WooCommerce charges nothing at all for the software. The moving part is a set of percentages taken from every order: card processing at around 2.9%, a platform fee of up to 2% for using a payment provider the platform did not pick, and app fees stacked on top of both. On one $100 order those three can come to $6.88. The fixed part is the number in the quote. The moving part is the one that grows.

Why this matters more than it used to

Online selling is now a large enough share of retail that the percentages are attached to real money. The Census Bureau reported on 18 August 2026 that United States retail ecommerce sales for the second quarter of 2026 were $340.2 billion once adjusted for seasonal variation, up 12.2% on the same quarter of 2025 against 6.7% growth for retail as a whole. That quarter's ecommerce total was 17.1% of the $1,986.5 billion of retail sales.

The number that matters to a single store is smaller and sharper. Take a business selling $250,000 a year online. At 6.39%, which is what the layered configuration further down this page comes to, the percentages take $15,975 before anyone counts the cost of the goods. At 2.9%, which is the same sale with no platform fee and no per-transaction app, they take $7,250. The $8,725 between those two figures says nothing about the business itself. That gap is the residue of choices made before launch, while everybody was still arguing about the homepage.

What the platforms charge you to exist

Start with the part everybody quotes, because it is the smallest. Shopify's published plans on annual billing are Basic at $19 a month, Grow at $49, Advanced at $299 and Plus from $2,300. Paying monthly instead costs $25, $65 and $399 for the first three. Shopify states plainly that there are no setup fees on any plan, and its point of sale upgrade is a separate $89 a month per location.

BigCommerce publishes four self-service tiers on annual billing: Core at $29 a month, Growth at $79, Scale at $299 and Performance from $1,499. Its monthly-billing equivalents are $39, $105 and $399. Squarespace, which is a website builder that sells commerce rather than a commerce platform, publishes three plans at $12, $17 and $25 a month on annual billing, or $19, $27 and $36 monthly. Wix lists Light at $17, Core at $29, Business at $39 and Business Elite at $159 a month on yearly terms, with a note on the same page that the dollar figures are shown for reference only and that prices vary by location, which is why nothing further in this article rests on them.

One aside is worth the detour, because it shows how the ranges on page one get made. BigCommerce's guide to ecommerce cost puts platform hosting on a software-as-a-service platform at $80 to $730 a month. BigCommerce's own pricing page prices its three self-service plans at $29, $79 and $299 on annual billing. A company can be the best available source for its own prices and a poor one for the category, and the two pages sit on the same domain.

So the subscription question is answered in a paragraph, and for a small store the difference between the cheapest and the most expensive mainstream plan is a few hundred dollars a year. That is not where the money goes.

The cut nobody adds up

Every online sale is skimmed at least once, by the processor, and it can be skimmed three times. Stripe publishes 2.9% plus $0.30 for domestic cards, with no setup or monthly fee, plus 1.5% for international cards and another 1% where a currency conversion is needed. PayPal's United States rate card puts its own checkout at 3.49% plus a fixed fee, and standard credit and debit card payments at 2.99% plus a fixed fee. Those are the unavoidable ones.

The second cut is the one that surprises people, because it is charged by the platform rather than by anybody handling the money. Shopify charges a third-party transaction fee when you use a payment provider other than Shopify Payments, and the rate falls as you move up: 2% on Basic, 1% on Grow, 0.6% on Advanced and 0.2% on Plus. BigCommerce charges the same kind of fee under a different name, the Open Payment Provider fee, at 2.0% on Core, 1.0% on Growth and 0.6% on Scale, and it says in its own footnote that the fee is billed by BigCommerce to you rather than to your shoppers and appears as a separate line on the invoice, on top of whatever your provider charges.

Squarespace runs a third variant, charging an online store transaction fee of 2% on its Basic plan and 0% above it, with digital content and memberships charged at 7%, 5% and 0% across the three tiers.

The top row is a real configuration: a Shopify Basic store using an outside gateway and one subscription app. The percentage components are 2.9%, 2% and 1.49%, which is 6.39%, and the fixed components are $0.30 and $0.19, so a $100 order leaves $93.12. Drop the app and it is 4.9% plus $0.30, or $5.20. Move the same order onto BigCommerce Core with Stripe and the platform fee disappears entirely, leaving $3.20. Nothing about the store changed in those three rows. The bill more than doubled on configuration alone.

How the platform fee goes to zero

The two platforms take different routes to zero and the difference is worth knowing before you pick one. Shopify's fee exists to move you onto Shopify Payments, so the way to avoid it is to use Shopify Payments. Its card rate is not printed on the pricing page and Shopify says only that the rate depends on your plan, which means the largest recurring percentage in the whole model is the one the pricing page does not print.

BigCommerce takes the opposite approach and names its exceptions. Orders processed through what it calls Embedded Payment Providers carry no fee at all, and it publishes the list: BigCommerce Payments, Adyen, Affirm, Afterpay, Amazon Pay, Bank of New Zealand, Checkout.com, Clearpay, Fortis, Global Payments, Klarna, PayPal Braintree, PayPal Complete Payments, PayPal Wallet, Bluesnap, Bankful, Sezzle, Stripe, TD Online Mart, Worldpay and Zip Money. Twenty-one names, Stripe among them, and each order is attributed to whichever provider processed it, so a store running everything through one of those pays nothing. BigCommerce also publishes a floor for card processing itself, saying the fee for debit and credit cards begins at 2.89% plus $0.29 across all its plans, with American Express at 3.50% and no fixed fee.

That is a real and checkable difference between two platforms that get compared on subscription price alone. On one of them the 2% is a toll for choosing your own processor. On the other it disappears if you pick a provider from a published list. Settle that question before you choose the platform, not after.

The price that goes up by itself

Here is the part the cost guides leave out, and it is printed on the plan cards. BigCommerce ties each self-service tier to a sales ceiling and moves you up when you cross it. The Core card says the plan auto-upgrades to Growth. The Growth card says the same about Scale. Neither of the two guides quoted at the top of this article mentions it, and one of them is published by BigCommerce.

The thresholds are $30,000 and $100,000 of trailing twelve month gross merchandise value, and the footnote defines that value as gross order value minus 10%, so in sales terms the moves land at about $33,333 and $111,111. Scale is metered rather than capped: it covers $33,333 a month of gross merchandise value and charges 0.9% on anything above that, and once trailing twelve month value reaches $2 million the account moves to Performance. The company's own answer on how automatic upgrades work says you are notified and then upgraded.

Read that as a pricing model rather than as a complaint. It is defensible, it is disclosed, and the alternative is a hard cap that breaks your store on a good week. But it means the honest answer to what a BigCommerce store costs is not a number. It is a function of your sales, and you can plot it.

What crossing one line costs

The chart above assumes a store using a payment provider outside the embedded list all year, and it prices only the platform, not the processing. At $30,000 of sales the bill is $348 of subscription plus 2% of $27,000 in gross merchandise value, which is $540, so $888 for the year. At $60,000 the store is on Growth: $948 plus 1% of $54,000, which is $1,488. At $110,000 it is still on Growth, at $948 plus $990, or $1,938.

Then it crosses. At $120,000 of sales the store sits on Scale, and the arithmetic is $3,588 of subscription plus 0.6% of $108,000, which is $648, for a total of $4,236. Ten thousand dollars of extra sales, an increase of about 9%, has taken the annual platform bill from $1,938 to $4,236. It more than doubled.

The effect fades as the store grows into the plan. At $250,000 of sales the same store pays $4,938, and at $500,000 it pays $6,738, which includes $450 of the published 0.9% overage on the gross merchandise value above $33,333 a month. As a share of sales that is 3.5% at the crossing point and 1.3% at half a million. The bill is not unreasonable at either end. It is just that the worst moment to be on this pricing model is the month you outgrow a tier, and nobody plans for that because nobody plots it.

The apps take a percentage too

The third layer is the one that turns a tidy budget into a surprise, because store owners tend to think of apps as software subscriptions when several of them are priced like payment processors. Search the Shopify App Store for subscriptions and it returns 787 apps. On the first page of those results every listing but one showed a free entry label, free to install or free plan available, rather than a price. That label describes the cost of switching the app on, which is not the same thing as the cost of running it.

Recharge publishes two plans. Starter is $99 a month plus 1.49% and 19 cents per transaction. Plus is $499 a month plus 1.34% and 19 cents. Its own pricing FAQ states that this sits on top of anything else the store pays, naming the ecommerce platform, the payment processor and other applications. That is a vendor telling you, correctly and in writing, that its percentage stacks on the others.

The lesson generalizes past this one company. When you evaluate an app, find the transaction line before you look at the monthly line, because on a growing store the transaction line is the larger of the two. A store putting $20,000 a month through a 1.49% app pays about $298 in percentage fees, plus 19 cents on every order, against a $99 subscription. The subscription was never the price.

The route where the bill does not move

The alternative is to buy the software once and rent only the infrastructure. WooCommerce publishes its position on its own pricing page: the platform is free and open source, with no platform fee and a stated 0% revenue share. It puts hosting at $25 to $350 a month for most stores and extensions at $29 to $299 a year each, bought individually rather than bundled into a tier.

Two honest caveats belong here. Free software is not a free store: the hosting, the extensions and somebody competent to run all of it still cost money, and we have set out what that ongoing bill actually contains in the four bills behind one number. And this route hands more responsibility to you, which is a real cost even when no invoice arrives for it.

Worth noticing while you are on that page: WooCommerce's comparison table lists Shopify Basic at $29 a month and BigCommerce Core at $29 a month with online revenue capped at $50k. Checked against each company's own pricing page on the same day, Shopify Basic is $19 on annual billing or $25 monthly, and BigCommerce Core's published ceiling is $30,000 of trailing twelve month gross merchandise value. Comparison tables written about competitors go stale, including the ones published by companies large enough to know better, which is the whole argument for reading the vendor's own page and dating what you read. The same discipline applies when you are checking an agency, and we have written that method out in what to check before you call.

On the design layer, the cheapest honest number in ecommerce is a theme. The Shopify Theme Store lists 1,210 paid themes, and the first page of them runs $100 to $500, paid once. A theme is not a design and it will not solve a positioning problem, but a store with no budget and no traffic yet should almost certainly buy one and spend the difference on stock and ads.

Every percentage that can attach to one order

This is the page to keep. Each row is a separate charge with its own rate, its own invoice line and its own way of going to zero, and none of them is hidden: all seven were read off the vendors' own pages in an afternoon.

Two things follow from reading the table as a whole. Only the first row is genuinely unavoidable, because somebody has to move the money. Every other row can be taken to zero, or close to it, by a decision made in the first two weeks of the project, which is exactly when nobody is thinking about fractions of a percent.

What the money should actually buy

All of the above prices the plumbing. It says nothing about whether the store sells anything, and that is the part a design budget can move. Baymard Institute, which has run large scale checkout usability research for years, puts the average documented cart abandonment rate at 70.22%, calculated across 50 separate studies.

The reasons are more useful than the headline. In Baymard's quantitative study of United States shoppers, 42% said they had abandoned a cart because they were only browsing, which no amount of design will fix. Set that group aside and the distribution looks like this.

Extra costs being too high leads at 40%, followed by slow delivery at 20%, not trusting the site with card details at 19% and being asked to create an account at 18%. A checkout that was too long or complicated and a site that had errors or crashed tie at 17%. Then an unsatisfactory returns policy at 13%, not being able to see the total cost up front at 12%, a declined card at 10%, too few payment methods at 9%, and 7% who did not know.

Our reading of that list, offered as a judgment rather than as a finding of Baymard's: at least five of those eleven describe something on the page rather than something about the business. Extra costs arriving late in the flow, a forced account, a checkout with too many steps, a total you cannot see until the last screen, and a payment method you do not offer are all build decisions. Baymard's own measurements make the third one concrete: an ideal checkout flow can run to 12 to 14 form elements, or 7 to 8 counting only the fields, while the average United States checkout in its benchmark database shows 23.48 elements by default, 14.88 of them fields.

That is the argument for spending the money on the store rather than shaving the subscription. Cutting a checkout from 23 form elements to 14 is a fixed cost, paid once. Two extra points of fee is a percentage, paid on every order for as long as the store exists. And a shipping charge that only appears on the last screen is working against the single most cited reason on that chart.

Where we fit, and who should not call us

We build ecommerce sites, and the two shapes of work we sell are set out on our offer page: a build you own outright and keep, or a monthly partnership for businesses with continuing work. What either costs depends on what already exists, so the figure gets agreed in conversation rather than guessed on a blog. The outcome we cite most often is 20% more website conversions, measured after a treatment center rebuild for Cornerstone Healing Center, and we have shipped more than 200 projects. More on how we approach this market sits on our ecommerce page.

Several kinds of reader should not hire us, and saying so is more useful than a disclaimer. If you have not sold anything yet, buy a theme, take the entry plan, and find out whether people want the product before you commission anything. If your store already converts well and simply looks dated, the money is better spent on stock, photography or acquisition, and a rebuild will feel productive without being profitable. If the actual problem is that nobody knows the store exists, that is a traffic problem and a different specialist solves it more cheaply than we would.

Where a build earns its cost is when the checkout is the constraint: a store with traffic and a poor conversion rate, a catalog that has outgrown its template, a migration off a platform whose percentages have started to hurt, or a business that needs the site to explain something considered and expensive before anyone will buy. On ownership of what you end up with, which matters more on ecommerce than anywhere else because the store is the business, read what to check before you sign.

Price the store you expect to be

Here is the exercise, and it takes about twenty minutes with last year's sales figure in front of you. Do not model the store you are running today. Model the one you expect in two years, run the same arithmetic back to where you are now, and look hard at what happens in between.

Write down three annual sales numbers: what you did in the last twelve months, three times that, and ten times that. For each of the three, work out four lines. One, the platform subscription at the tier those sales require, remembering that BigCommerce measures the threshold on gross order value minus 10% and moves you across at about $33,333 and $111,111 of sales. Two, the platform's percentage on order value, which is zero if you use Shopify Payments or a provider from BigCommerce's embedded list, and up to 2% if you do not. Three, card processing at your processor's published rate, around 2.9% plus 30 cents unless you have negotiated. Four, every app that charges per transaction, at its own percentage.

Then look at the gaps between the three columns rather than at the totals. If the bill roughly triples when sales triple, you are on a model that scales with you and the decision is simply whether you like the product. If it jumps by more than the sales did, you have found a threshold, and you now know the sales figure at which it fires, which is enough to plan a migration a year early instead of discovering it on an invoice. Either way you will have priced the store rather than the quote, which is more than any range on page one can tell you.

There is a second kind of check that almost none of those pages mention, and it runs before you speak to anyone. Three public records will date the claims on an agency's website, and two free tools will measure the work. Google publishes its own list of questions for hiring this kind of help, and the guides competing for this search do not cite it. Every figure below was read or measured on 19 August 2026, and where a test was run on our own site the result is printed here, including the one that came out badly. Since we sell websites, this is a page written by one of the firms you would be checking, and it should be read that way.

The 30-second answer

Most of this decision can be settled before a single call, from records the agency does not control. Look up the domain registration to date the firm, pull its portfolio sites out of the Internet Archive to see whether that work is still live, measure its own site with PageSpeed Insights, and find the company registration behind the trading name. Then read Google's own published guidance on hiring, which tells you that nobody can guarantee a ranking and that you should grant read access to Search Console and nothing more while a firm is still pitching. Two checks are left for the conversation itself: what account access they ask for, and what the contract says about ownership. The whole records pass takes about half an hour.

What the advice on page one is actually measuring

The standard red-flag list measures sales performance, and sales performance is the least reliable signal available to you. An agency with no process can describe a process. A firm that has never shipped for your industry can talk fluently about your industry for forty minutes. Clear communication during a pitch predicts clear communication during a pitch.

This is not an argument that the conversation is worthless. It is an argument about order. Do the records first, take what you find into the call, and you will be asking about specific things rather than listening for a tone. A firm that has been trading for four years and a firm that says it has been trading for fifteen sound identical on a call. They do not look identical in a registry.

One more thing worth saying plainly, because it shapes everything below. Every check here is one you can run on us, and we have run them on ourselves at the end of this article. If a check only works in one direction it is marketing, not a method.

Check one, the domain record, which dates every claim on the site

The registration record is the cheapest way to test an agency's history, because it is published by the registry rather than written by the agency. Since 28 January 2025 the authoritative source for that data is the Registration Data Access Protocol, not the old WHOIS service. From that date ICANN stopped requiring generic top-level domain registries and registrars to run WHOIS at all, with .com, .name and .post carved out of the change. ICANN runs a free lookup at lookup.icann.org, and in the announcement making the change it put RDAP traffic at more than 10 billion queries a month as of December 2024.

Four fields are worth your attention. The registration date, which you compare against any claim about years in business. The registrar, which tells you where the asset actually lives. The status flags, where a set of client-side locks means somebody has deliberately made the domain harder to move or delete. And whether the zone is signed with DNSSEC, which most small business domains are not.

The company register behind the trading name is the second free record, and it is the one that tells you what you would actually be contracting with. In the United States that is the Secretary of State business search in the state of incorporation, which returns the formation date, the entity type and whether the company is in good standing. In the United Kingdom, Companies House publishes the same thing plus filed accounts and the list of officers, free, at find-and-update.company-information.service.gov.uk. A trading name with no registered entity behind it is not automatically a problem, since plenty of good freelancers are sole traders, but you should know which one you are hiring before you send a deposit.

Read the date as a question, not a verdict. Plenty of good studios rebranded and bought a new domain last year, and plenty of empty shells own fifteen-year-old domains they bought at auction. What you are testing is the gap between the record and the claim. A homepage that says it has served clients since 2011 above a domain first registered in 2024 has a story to tell, and it might be a perfectly good one. Ask for it.

Check two, the portfolio, which is a claim until you open the live sites

Treat a portfolio as a list of assertions to verify, and verify it in two passes. First, open every live URL. Screenshots with no link attached are the format of choice for work that is no longer live, was never live, or was never theirs. Second, run the ones that matter through the Internet Archive, which lets you see what a page looked like on a given date rather than what it looks like after two years of somebody else's edits.

The Archive publishes a machine-readable index, so you do not have to click through a calendar. Requesting web.archive.org/cdx/search/cdx?url=example.com&output=json returns one row per capture with the date and the HTTP status. Two patterns are worth catching. A portfolio site whose archived version from the launch year looks nothing like the case study means the client has since replaced the work, which is information, not an accusation. No captures at all before the stated delivery date mean the site did not exist when the work was supposedly finished.

The same index makes a quieter check possible on the agency's own site. Ask when their portfolio page first appeared and how much of it has survived. A studio that quietly deletes half its case studies every year is telling you something about how those relationships ended.

Check three, their own site measured, and the promise to stop accepting

If an agency promises you a performance score, ask which run. We tested our own homepage twice with Google's PageSpeed Insights on 19 August 2026, changing nothing between the two tests. At 2:55:09 PM it scored 96. Seventy-four seconds later, at 2:56:23 PM, the same URL scored 86. Largest Contentful Paint was 1.7 seconds on the first run and 4.1 seconds on the second, which is not a small wobble: those two numbers sit on opposite sides of both thresholds Google publishes, one comfortably good and one squarely poor.

The direction of the other metrics makes the point sharper. Total Blocking Time actually improved between the runs, from 130 milliseconds to zero, and Speed Index improved too. The score still fell ten points, because Largest Contentful Paint, which carries a quarter of the weight on its own, went from good to poor.

Two things follow from that chart. Total Blocking Time is the heaviest single input at 30% and it is not a Core Web Vital, while Interaction to Next Paint is a Core Web Vital and carries no weight in this score at all. So a performance score is not a measurement of the three things Google says it uses. It is a related lab test with its own recipe.

The scale is worth knowing too, because it explains why 90 sounds so impressive. Google documents that each metric's scoring curve is fitted to real HTTP Archive data, with the 25th percentile of sites mapped to a score of 50 and the 8th percentile mapped to 90. So a green metric means that page was in roughly the fastest tenth of the web on that measurement, on that run, which is a genuine achievement and also a moving target.

None of this makes the tool useless. We use it constantly, and the diagnostics under the score are some of the most useful free diagnostics available for a specific page. It makes the promise useless. A firm that guarantees a number is guaranteeing the outcome of a test that varies by ten points on an unchanged page.

Check four, the numbers Google actually uses, and why yours may not exist

The measurements that feed Google's page experience signal come from real visitors, not from a test anybody runs on demand. Three metrics matter, each has a published boundary, and a page passes only when it meets all three targets at the 75th percentile of real page loads, counted separately for mobile and desktop.

Largest Contentful Paint should arrive at 2.5 s or less and counts as poor above 4.0 s. Interaction to Next Paint should be 200 ms or less and is poor above 500 ms. Cumulative Layout Shift should be 0.1 or less and is poor above 0.25. Those are the whole standard, and any agency proposing to improve your page experience should be able to say which of the three your site is failing.

Google set those boundaries against a stated achievability rule: a good threshold has to be reachable by at least 10% of origins. When the current thresholds were fixed, 27% of phone origins were meeting the Largest Contentful Paint target, 56% were meeting the interactivity target and 60% were meeting the layout target. These are not aspirational numbers invented by an agency; they were pitched at levels a real share of the web was already hitting. Loading was the target most sites were failing then, and it is the one that failed our own homepage on the second run above.

This next part catches out most small businesses, and it caught us. Field data needs enough real traffic before Google will report it. The top row of both of our reports above says No Data, because our own site does not have the visitor volume to produce a reading. Absence is not failure here, and it means the only performance number available for a small or brand new site is the volatile lab one. Anyone showing you a Core Web Vitals pass for a site that has not launched is showing you a simulation.

Check five, the platform, which moves your odds before anyone writes code

Ask what they build on and why, because the answer shifts your starting position measurably. The Web Almanac measures Core Web Vitals across the whole crawlable web by platform, and in 2025 the spread between the best and worst mainstream systems was 40 percentage points.

Between the two ends the order runs TYPO3 at 79%, Squarespace at 69%, Drupal at 61%, Joomla at 55%, 1C-Bitrix at 54%, and Tilda and Weebly level at 47%. Duda leads the table on 85% and WordPress, the platform most agencies will propose to you, comes last on 45%.

Read this as defaults, not destiny. The tightly managed hosted builders score well because they control the hosting, the image pipeline and the theme, and their customers cannot install the plugin that breaks it. The extensible platforms score worse because their averages include every abandoned site running eleven plugins on cheap shared hosting. A carefully built WordPress site will beat a neglected Squarespace one every time, and a bad custom build will lose to both.

What the chart is genuinely good for is calibrating a promise. If a firm proposes the platform sitting at 45% and also guarantees you a fast site, they are proposing to beat their own tooling's odds, which is possible and is exactly the thing to ask them how they do. We have written separately on the wider builder versus designer decision, which is the version of this question you face before you shortlist anyone.

Check six, the questions Google publishes and nobody quotes

Google maintains a page on hiring this kind of help, last updated on 5 June 2026, and it is the single most useful document in this entire search result. It is written about hiring an SEO, but every question on it applies to anyone who will touch your site, because the risk it describes is the same one: work done on your behalf that damages your visibility.

The interview list is short and specific. Can they show previous work and success stories. Do they follow Google Search Essentials. What results do they expect and in what timeframe, and how will they measure success. What is their experience in your industry, your country and your city. How long have they been in business. How will you communicate, and will they share every change they make to your site.

The warnings are blunter than anything an agency would write. Google states that nobody can guarantee a number one ranking, and says to be wary of firms that claim guaranteed rankings, a special relationship with Google, or a priority submission. It says to be skeptical of agencies that email you out of the blue, and it compares that mail to the offers you already delete. It notes that Google does not evaluate or endorse third-party SEO tools and that those tools have no access to its internal ranking data, which is worth remembering the next time somebody shows you an authority score as though it were a fact about Google.

It also puts the liability where it lands. Google's page says the responsibility is yours: you answer for the actions of any company you hire, and deceptive content created on your behalf can get a site removed from the index entirely. That single line should change how you read a proposal offering hundreds of backlinks. Google's spam policies name buying links, excessive exchanges, automated link building, and paid guest posts or press releases carrying optimized anchor text. They also name scaled content abuse, which covers generating large volumes of pages with AI tools without adding value, and that is a fair description of several content packages being sold this year.

Check seven, the access they ask for, which is the one to get right first

Watch what permissions a firm requests and when, because the wrong grant is far harder to reverse than a bad invoice. Google's guidance is explicit on the audit stage: grant read access to Search Console at that point, and do not grant write access.

Search Console has three permission levels for people. An owner has full control and can add and remove other users. A full user sees all data and can take some actions. A restricted user gets simple view rights on most data. That ladder is not the trap. The trap is that there are two kinds of owner. A delegated owner can be removed by any owner from the user management screen. A verified owner proved ownership with a token, such as a file or a tag on the site, and removing that person means finding and removing their token from your own site. A property must keep at least one verified owner or nobody has access at all.

So the sequence matters more than the policy. Verify the property yourself, with your own token, before anyone else touches it. Add the agency as a full user, or a restricted one while they are still pitching. You can then remove them in ten seconds when the engagement ends, without going hunting through your own markup. The same reasoning applies to the registrar account, the hosting account and the analytics property, and we have set out the ownership half of that argument in detail in who owns your website.

Check eight, reviews, which are claims governed by a rule

Read testimonials as marketing copy that happens to be regulated, and check them the way Google suggests, by asking for business references and actually calling past clients. Since 21 October 2024, the FTC's rule on the use of consumer reviews and testimonials, 16 CFR Part 465, published at 89 FR 68034, has banned fake and insider reviews, undisclosed material connections, compensation conditioned on a particular sentiment, suppressing negative reviews through legal threats, and buying fake indicators of social media influence.

This is not an abstract risk in our industry, and the clearest example involves a website vendor. In an order finalized in April 2025, the FTC required accessiBe, which sells an automated web accessibility widget, to pay $1 million. Its product could not make any website conform to the accessibility guidelines, the complaint alleged, despite claims that it could, and the company also formatted third-party articles and reviews so they read as independent opinion while failing to disclose its connections to those reviewers. The second allegation is the one worth carrying into your shortlist. Paid coverage dressed as independent opinion is something a regulator has alleged and settled against a vendor in this market, not a thing we are inventing to make a point.

Two practical moves follow. Check whether the glowing third-party review of an agency is on a site the agency has any relationship with, and check whether an accessibility or compliance promise is being made about an automated product. We have written on both sides of this: what you can and cannot publish in the testimonial rules, and what actually reduces legal exposure in accessibility compliance.

The eight checks, and what a bad result actually means

Everything above condenses into one page. Nothing in this table requires the agency's cooperation, and none of it costs money.

The two questions the records cannot answer

Two things stay invisible until you read the paperwork, and both decide what you are left with when the relationship ends. The first is ownership. A website is not automatically yours because you paid for it, the domain and hosting accounts should be in your company's name before anything is built, and the transfer of copyright is a specific clause rather than a general assumption. We have taken that apart properly in what to check before you sign, and it is the piece to read next if a contract is already in front of you.

The second is what happens after launch. Hosting, platform updates, somebody being responsible when a form stops sending, and the cost of changes are four separate purchases that usually arrive as one vague monthly figure. Ask for them itemized before you sign, not after the first invoice, and see the four bills behind one number for what each of them tends to cost.

What we sell, and the readers we would turn down

Two shapes of work are set out on our offer page: a build you own outright, paid once, and a monthly partnership for businesses with continuing work. What it costs depends on what already exists, so the number gets agreed on a call rather than guessed here. A focused build typically goes from kickoff to live in about two weeks. Any partnership inquiry gets a reply within 48 hours. Out of more than 200 shipped projects the outcome we point at most often is a treatment center rebuild, after which Cornerstone Healing Center saw 20% more website conversions. We work across local service businesses and fourteen other fields, and the industry pages set out how.

Plenty of readers should not hire us, and being specific about that is more useful than a disclaimer. A site that already converts, whose only fault is looking dated, does not need us, and that money is better spent closer to revenue. If the checks above are new to you because this is a first website for a business that has not yet proven demand, a hosted builder will get you trading faster and cheaper, and you can hire properly once you know what the site has to do. If what you actually want is somebody to run paid ads and post to social every week, that is a different trade with different specialists, and we would be an expensive way to buy it.

Where we earn the money is when the site itself is the constraint. A practice whose phone rings but whose online booking never gets used. A company whose old vendor holds the domain. A business selling something considered and expensive, where the website has to do the arguing before anyone will book a call.

We ran all eight checks on ourselves

A method you will not turn on yourself is a sales tactic. Here is our row, filled in on 19 August 2026, in the same order as the table, so you have a completed example to hold the others against. Two of the eight come back badly.

One, domain registration. khanwork.com was registered on 11 November 2022 through GoDaddy, IANA registrar ID 146, and it expires on 11 November 2026. All four client-side locks are set, covering transfer, update, renew and delete. The zone is not signed with DNSSEC, which is a gap rather than a crisis and is on our list. So we are a young domain, and nothing on our site claims otherwise. If it did, that record is where you would start.

Two, portfolio, live. This one we fail. Our case study pages name the client and the outcome, but they do not currently link out to the live sites, which by the rule we set out above is the weaker format. Hold that against us; it is exactly the gap this check is designed to surface. Ask us for the URLs and you will get them, and the pages should carry them without being asked.

Three, portfolio, historic. The Internet Archive holds 47 captures of our own root URL, the first from 17 March 2023, about four months after the domain was registered, which is what you would expect from a site that was being built during that gap.

Four, our own speed. Published above, including the run we would rather not have shown you. Mobile performance came back at 96 and then 86 within 74 seconds, with Largest Contentful Paint at 1.7 seconds and then 4.1 seconds. Accessibility scored 97, best practices and SEO both scored 100. A run landing at 4.1 seconds means a real visitor on a throttled connection can wait that long for our main image, so it is a defect on our side rather than a quirk of the tool, and it is the kind of thing you should expect a candidate to name rather than explain away. The field data row says No Data on both runs, because the site does not yet get enough traffic for Google to report real-user measurements. Any agency claiming a verified Core Web Vitals pass for a site this size is describing a lab test.

Five, platform. This site is a statically generated build rather than a CMS, which is why it does not appear in the chart above. That is a deliberate trade and it costs us something real: editing content here needs a developer in a way that a Squarespace site does not.

Six, guarantees. We do not guarantee rankings, positions or scores, for the reasons set out in this article. The two-week figure above is what a focused build typically runs to, not a promise, and projects with unsettled scope run longer.

Seven, account access. We ask clients to verify their own Search Console property and add us as a user on it. Where a previous vendor already holds verified owner status, untangling that is usually the first hour of the engagement.

Eight, reviews. The testimonials on our site are from named clients on named projects rather than anonymous praise. Ask for references and we will give you names and numbers to call, which is the check Google recommends and the only one on this list that needs us to do anything.

Run the same eight rows for every firm on your shortlist, including the one you are already leaning towards. The firm that comes back with the fewest gaps between its claims and its records is not automatically the best designer in the pile, but it is the one whose proposal you can read at face value. That is worth more than a page of adjectives about process.

What follows is the same question answered only with prices that companies print on their own websites, and with rules quoted from the agencies and statutes that wrote them. Everything below was read on 18 August 2026 and linked where it appears. We sell websites, so read this as a page written by one of the options on it. The gyms we are the wrong choice for get their own heading further down, and it is not a short list. There is also a legal half to this question that the cost guides skip completely, and for a gym it is the half that decides what your join page has to look like.

The 30-second answer

A gym website costs between about $900 and $5,988 in its first twelve months if you buy from a company that publishes prices, and very little of that is design work. Most of it is subscription pricing for gym management software that happens to include a website. The floor is $75 a month from a vendor that hands every account a website at no extra charge. The ceiling is $499 a month for a plan where the website is one line among booking, billing, a branded app and a CRM. Two things move this decision more than any price. Several platforms a gym already pays for include a website, so the real question is whether the included one is costing you signups. And if you sell memberships online, your join and cancel flow is regulated.

What the published market actually charges

Six companies selling software or websites to gyms print their prices in public, which makes them the only honest basis for a comparison. Their plans are not equivalent, and the table says so: some include a website, some sell one as an upgrade, and two do not offer one at all.

Gymdesk prices by member count and nothing else: $75 a month up to 50 active members, $100 for 51 to 100, $150 for 101 to 200, $200 for 201 to 400, and a custom quote above that. Every tier lists the same three lines, unlimited staff accounts, a responsive gym website and all features included, and the page adds that visitors, frozen members, canceled clients and prospects do not count toward the total. Plans are month to month with a 30-day trial and no card required to start.

GymMaster also prices by member count, but per site: Foundation at $89 a month for up to 100 current members, Advanced at $129 for up to 400, Professional at $209 for up to 1,300, and a quote above that, with 10% off for paying yearly and a 60-day money-back guarantee. Its feature list includes website integration, online timetables and contact forms in every tier, plus a customizable member portal. That is not the same thing as a website, and the distinction matters more than the price.

PushPress publishes three plans on its pricing cards, Free at $0, Pro at $159 and Max at $229, and then a fuller ladder in its own pricing FAQ further down the same page: Train from $79 a month for workout tracking up to 20 clients, Grow at $329 a month described as a full CRM with automated workflows, email and SMS, and websites, a branded app add-on at $81 to $97 a month, and a Full Stack bundle at around $559 a month. Read that ladder twice. The website does not appear until $329.

Zipper is the clearest of the six about what changes as you climb: Starter at $139 a month gets a basic branded website and pays a 1% platform fee on payments, Studio at $299 gets a custom-branded website and no platform fee, and Growth at $499 adds unlimited members and API access. Extra locations are $129 a month each, there are no setup fees, and the page states plainly that you also pay Stripe's standard 2.9% plus 30 cents.

Two more are worth naming. TeamUp prices by customer count, showing $189 a month for 101 to 200 customers, with a custom branded app at $99 a month, and lists no website in what is included. Wodify has three tiers, Essentials, Accelerate and Ultimate, and puts a Custom Website only in Ultimate, but it does not print a price against any of them. What it prints instead is a promotion: starting at $199, struck through, $99 a month per location, with $100 a month off for life if you book a demo by 31 August. A price you can only get by booking a call, expiring on a date, is not a published price, and we have left it out of the chart for that reason.

Which platforms hand you a website, and which do not

Ask six vendors whether a website is included and you get four different answers, and the one your vendor gives is the single biggest factor in what a gym website should cost you. Gymdesk includes one on every plan, including the cheapest. Zipper includes a basic one at $139 and a custom-branded one from $299. PushPress puts websites in Grow at $329. Wodify puts a Custom Website in its top tier only. GymMaster and TeamUp do not sell you a website at all, and expect you to bring your own and wire their booking into it.

That line under the pricing cards is the sentence that should reframe this whole purchase for most gym owners. If your software already gives you a site at no extra cost, the honest question is not what a website should cost. It is what the included one cannot do, and whether that gap is worth paying to close.

Here is our answer to that, and it is a judgment rather than a sourced fact, so weigh it accordingly. A platform website is usually fine at three things: showing an accurate timetable, taking a booking, and existing. It is usually weak at three others: saying what makes your gym different from the one four minutes away, ranking for anything other than your own name, and handling an intro offer as something other than a generic sign-up button. If your problem is the first list, do not buy a website. If it is the second list, the included site is now the constraint.

One structural warning that applies to every bundled option, and we have written about it at length elsewhere: when the software company builds and hosts your site, leaving the software means leaving the site. Ask before you sign, not at renewal.

The price that is not on the pricing page

The cheapest plan in this market is not the cheapest way to run a gym, because payment processing is priced by plan, and past a certain volume the gap between the rates is bigger than the gap between the subscriptions. PushPress publishes its rates next to each tier: 4.99% plus 30 cents per card transaction on the free plan, 2.89% plus 30 cents on Pro at $159 a month, and 2.75% plus 30 cents on Max at $229. Bank transfers follow the same shape, 2.89% on the free plan against 0.79% on both paid plans.

The arithmetic is ours, done on their published numbers. Moving from the free plan to Pro costs $159 a month and saves 2.10 percentage points on every card payment you take. Those cancel out at $7,571 of monthly card volume. Below that the free plan is genuinely cheaper. Above it, the free plan is the more expensive product and the difference grows with every member you add. A gym with 150 members averaging $100 a month is at $15,000 of volume, which is roughly twice the break-even, so the free plan there hands over about $315 a month in extra card fees, a net penalty of roughly $156 once you allow for the subscription it avoids.

Zipper's version of the same trade is smaller and easier to read: 1% platform fee on Starter, nothing on Studio and Growth. The $160 a month between the two plans is repaid at $16,000 of monthly card volume. Whichever vendor you are looking at, find the processing rates before you compare the subscription prices, because on this shopping list the headline number is the smaller one.

What you actually pay once

Almost nothing in this market is a one-time cost, and the few things that are tell you something useful about where a gym's money really goes. GymMaster publishes its hardware prices openly, which is rare and to its credit.

The full list is longer than the chart. A door access controller is $550 and a door reader $300, a reception reader is $130, and a branded Bluetooth reader carries a $700 setup fee. Alongside those, GymMaster prices key fobs from $1.30 each plus $170 per 500, the tailgating camera system at $400 plus a $40 monthly module fee, and a branded member app at $150 a month on top of the $450 setup. My PT Website is the only vendor here charging a website setup fee at all, currently $499 against a struck-out $999, and it is charged once on every one of their three plans.

Put those next to each other and the point arrives on its own. The hardware on your door costs more than the setup fee on a done-for-you website, and the reader on your reception desk is the only item on the list that comes in under it. If your instinct has been that the website is the big capital item in opening or refreshing a gym, the published prices disagree.

The cancellation rule everyone still quotes was thrown out

If you have read anything about gym websites and memberships in the last two years, you have read that a federal click-to-cancel rule requires you to let members cancel online as easily as they signed up. That rule is not in force, and a good deal of what is still being written about it has not caught up.

The sequence is short and each step is on the record. In November 2024 the Federal Trade Commission published a final rule amending its Negative Option Rule, at 89 FR 90476, which among other things required sellers to obtain unambiguously affirmative consent and to provide simple cancellation mechanisms that immediately halt recurring charges. Industry groups petitioned four circuits, the petitions were consolidated in the Eighth Circuit, and in Custom Communications, Inc. v. FTC, 142 F.4th 1060, that court found the Commission's failure to issue a preliminary regulatory analysis procedurally insufficient and vacated the 2024 rule outright.

The Commission then made it official. On 12 February 2026 it published a final rule, effective the same day, recodifying the text of the Negative Option Rule as it existed before the 2024 amendments and changing the name of the rule back to Use of Prenotification Negative Option Plans. On 13 March 2026 it opened a fresh rulemaking on the same subject and took comments until 13 April. Its own regulatory agenda, published on 14 August 2026, still lists the removal of the vacated 2024 amendments from 16 CFR 425 as live work.

Two things this does not mean. It does not mean deceptive sign-up flows are now legal, because Section 5 of the FTC Act was never the thing that got vacated. And it does not mean nobody is regulating your join page, because state law moved into the space, and in one state it is more specific about your website than the federal rule ever was.

What binds you if you sell memberships online

California's Automatic Renewal Law is now the practical spec for any recurring membership sold on a website, and it is written in enough detail to be read as a build requirement. The controlling section is Business and Professions Code 17602, as amended by AB 2863, which applies to contracts entered into, amended or extended on or after 1 July 2025.

The rule that changes your build is subdivision (d). A business that lets a consumer accept an automatic renewal offer online must let them terminate it exclusively online, at will, without further steps that obstruct or delay immediate termination, and the method has to be either a prominently located direct link or button, which may sit inside a customer account or profile or in device settings, or an immediately accessible termination email that the business formats and provides and the consumer can send without adding anything. You may require the member to log in first. If they will not or cannot, you have to let them cancel by one of the offline methods instead.

Subdivision (e) then anticipates the retention flow every gym runs, and permits it on one condition. You may show a discount, a retention benefit or an explanation of what canceling means, but only while simultaneously displaying a prominently located and continuously visible link or button labeled click to cancel, or words to that effect. The phrase the federal rule became famous for is written into California's statute, and this half is still in force.

The rest of the section reads like a specification document, and every line lands somewhere in a build:

• Cancellation must be available in the same medium the member used to sign up, or the one they normally use with you.
• You must keep verification of the member's affirmative consent for at least three years, or one year after the contract ends, whichever is longer. That is a data retention decision, and it belongs in whatever your join form writes to.
• Before you confirm billing information you must clearly disclose that the membership renews until the member cancels, the length and terms of the renewal, the amount or range of charges and how often they will happen, at least one way to cancel, and your contact details. If that notice goes out electronically it has to carry a link to the cancellation process.
• If you change the fee, including a change the member previously agreed to, you must give clear notice of it and of how to cancel, no less than 7 and no more than 30 days before it takes effect.
• Annual memberships need an annual reminder, in the same medium, naming the service, the frequency and amount of charges, and how to cancel.

One honest limit on all of this. Other states run their own versions of these rules and several are stricter in places, but the primary sources for the two we most wanted to quote here were unreachable from our network on the day we wrote this, and we do not publish law we have not read. Treat the California text as the shape of what a careful build looks like, and get your own state checked by somebody qualified before you rely on it.

Three older rules that shape the join form itself

Underneath the renewal rules sits a much older body of law aimed squarely at gyms, and it constrains the form on your website in ways most designers have never heard of. In California these are the health studio services provisions of the Civil Code, and they were written for paper contracts but drafted broadly enough to cover the electronic ones.

Section 1812.82 requires every contract for health studio services to be in writing, and requires a copy to be physically given or emailed to the customer at the time they sign. Your online join flow needs to send that copy, and it needs to send it then, not in a welcome sequence three days later.

Section 1812.84 is the one that reads like a design brief. A health studio contract cannot run longer than three years, and payments cannot be required beyond its term. The contract must carry a statement of the initial or minimum length of the term, printed in at least 14-point type or presented in an equally legible electronic format, and that statement must be placed above the space reserved for the buyer's signature. There is a typographic minimum, and a required position on the page, sitting in a consumer statute. Cancellation, where the title allows it, may be done in person, by email from an address on file with the studio, or by first-class mail.

Section 1812.86 caps the total a health studio contract can require at $4,400, inclusive of initiation or initial membership fees and exclusive of interest and finance charges. If you sell a multi-year package with a joining fee on top, that ceiling is worth doing the arithmetic against before the pricing page goes live.

None of this is exotic or new. It is simply missing from the gym website cost guides we read while writing this, which treat the join page as a design problem and leave out the part where its type size and its layout are legislated.

What is left for the website to do

Once the software is handling booking, billing, waivers and check-in, and once the join flow is doing what the law asks, the honest job list for a gym website is short. That is not a criticism of websites. It is what makes a good one cheap to specify.

The site has to make the trial or intro offer obvious and one tap away, on a phone, at nine at night. It has to show a timetable that is actually current, which in practice means pulling from the same system your coaches update rather than a page somebody edits by hand. It has to state prices, because the single most common reason a gym inquiry never happens is that the visitor could not find out what it costs. It has to say where you are and where to park. It has to introduce the coaches, because for most independent gyms the coach is the product. And it has to say what kind of gym this is, clearly enough that the wrong person self-selects out and the right one recognizes themselves.

Scale is worth a moment here, because it explains why the differentiation matters more than the craft. Bureau of Labor Statistics Quarterly Census of Employment and Wages data for the first quarter of 2025 counts 45,564 private establishments in the fitness and recreational sports centers industry, the code that covers gyms, in the United States, employing 648,903 people at an average weekly wage of $435. The Occupational Employment and Wage Statistics for May 2025 put 322,930 exercise trainers and group fitness instructors in the country, with a median wage of $47,160 and a mean of $52,420. You are not competing with all 45,564. You are competing with the three or four within a few miles, and the visitor deciding between you is not comparing your typography.

One more thing belongs on that list. The reviews and transformation photos a gym leans on are, in our experience, the part of the site carrying the most legal exposure, and the rules there are federal, current and unrelated to anything above. We have written a separate piece on what you can and cannot publish.

Where the money goes first

The order you work through these matters more than which vendor you end up with.

Start by finding out what you are already paying for. Log into your platform and check whether a website is included, whether it is the basic or the custom version, and what it would cost to move up a tier. On the numbers above that answer ranges from free to $329 a month, which is a wide enough spread to be worth ten minutes.

Then fix the join flow, whichever site you are on. Same-medium cancellation, an honest pre-billing disclosure, a contract copy sent at signing, the term stated above the signature. These are small pieces of work, they are cheap to do while you are already in the code, and they are expensive to retrofit after a complaint.

Then, and only if the included site is holding back signups rather than merely looking dated, buy or build one you own, with a small honest set of pages: who this gym is for, the timetable, the prices, the intro offer, the coaches, and how to get here. A short site that answers the questions a stranger actually has will out-earn a twenty page template, every time.

Content and ads come last, if at all. A gym is a local business with a physical door and word of mouth is still the strongest channel most of them have. Before you buy a monthly content plan, ask the vendor for three pieces they wrote for other gyms and read them the way you would read a coach's programming.

Our offer, and the gyms we are wrong for

We work in two shapes, set out on our offer page: a one-time build you own outright, and a monthly partnership when there is continuing work. Scope and price are agreed on a call, because both depend on what you already have. Two weeks from kickoff to live is what a focused build usually runs to, and what decides that timeline is mostly content and approvals rather than build speed. Partnership requests get an answer inside 48 hours. Our fitness and wellness page sets out how we work in this field. The nearest outcome we can point to is a treatment center rebuild, where Cornerstone Healing Center ended up with 20% more website conversions.

We are more use to you if we are specific about who should close this tab. A gym with a full class schedule, a waiting list and a working Gymdesk site does not need us, and we would rather say so than sell you a rebuild you will not feel. If $75 or $89 a month is the shape of spending your cash flow can absorb, the included website is a reasonable purchase and we are not going to pretend it is a scandal. Want somebody posting to Instagram three times a week and filing a monthly blog? Hire a fitness marketing service. That is their product, and it is not ours.

The work we are good at starts where the website itself is the bottleneck: a gym whose classes fill but whose personal training never sells, a multi-location operator whose platform site cannot express three different timetables, a rebrand where the old vendor holds the domain, or a gym competing against a chain that outspends it and needs the site to do the arguing.

The join flow test, run it on your own site tonight

Every number in this article is decidable once you have watched your own join page behave. Do this on a phone, not a laptop, with your reading glasses off, and time it.

One. From your home page, how many taps to a trial or intro offer? More than two and the offer is decorative.

Two. Can you find the price of a standard membership without submitting a form? If not, you are asking a stranger to phone you before they know if they can afford you.

Three. Is the timetable on screen the one your coaches are teaching this week? Check one class against your software.

Four. Sign yourself up for the cheapest plan you sell. Does the page tell you, before you enter card details, that it renews until you cancel, how much and how often it will charge, and how to stop it?

Five. Now cancel that membership, using only the website. If you cannot, or the path pushes you to phone somebody, you have found the most expensive defect on your site, and it is a legal one in California and a trust one everywhere.

Six. Look at the confirmation email. Is your contract attached or linked, and is the length of the term stated where a member would see it before agreeing?

Six answers, twenty minutes, and no vendor involved. If all six pass, your website is doing its job and your money belongs somewhere else this quarter. If two or more fail, you now know exactly what you are buying, which is the only position worth being in before anybody quotes you a price.

Here is the same question answered only with prices that companies print on their own websites, and rules quoted from the agencies that wrote them. Everything below was read on 17 August 2026 and linked where it appears. We sell websites, including to practices in this field, so read this as a page written by one of the options on it. Near the end there is a section on the practices we are a bad fit for. There are also three rules that apply to a therapy website and to very few other kinds of small business site, and the cost guides we read while writing this leave all three out.

The 30-second answer

A therapy practice website costs between about $830 and $4,200 in its first twelve months if you buy it from a company that publishes its prices, and the entire spread is subscription pricing rather than build cost. The floor is a hosted plan at $69 a month with no required setup fee. The ceiling is a plan at $349 a month that bundles monthly content, secure forms and a directory membership. A one-time build you own outright, like ours, sits in the lower half of that range and then stops charging you. What almost nobody tells you before you buy is that the cheapest line in a private practice's marketing is not the website at all. It is a $29.95 directory listing, and the job left for the website is different, and smaller, than any of these plans imply.

What the published market actually charges

Two companies build websites specifically for therapists and print full price lists in public, which makes them the honest basis for a comparison. A third company sells the software most practices already run on, and its prices matter to this decision for reasons that become clear further down.

TherapySites publishes six tiers: Core at $69 a month, Advantage at $99, Premium at $158, Premium Plus at $207, Elite at $266 and Elite Plus at $336. The Core list is the useful one to read, because it tells you what the base product is: domain name management, hosting, a mobile responsive and SSL secured site, online appointment scheduling, unlimited email accounts and unlimited live support. Everything above Core is marketing services stacked on the same website. Optional add-ons are priced separately, including website design with concierge onboarding at $199 one time, secure emails and forms from $15 a month, reputation management from $40 a month, Google Ads from $600 a month plus a $99 setup fee, and Facebook Ads from $158 a month plus the same setup fee.

Brighter Vision publishes three named plans plus a custom tier. Billed monthly they are Start at $99, Grow at $179 and Flourish at $349, with a $100 setup fee on Start and Grow and none on Flourish. Billed annually the same plans work out at $78, $143 and $299 a month, with the setup fee still applying to the two cheaper ones, so a year of Flourish costs $3,588 up front against $4,188 paid monthly. Start buys a site built by a design team, hosting, security, a library of pre-written content and two email addresses. Flourish adds one optimized content page a month, HIPAA compliant forms and five secure email addresses through Hushmail, an analytics dashboard and a GoodTherapy Pro membership.

Notice two things before you pick a row. The gap between cheapest and dearest is about five times, and every plan on the chart produces a site a prospective client would call perfectly good. And the number that decides your real annual spend is not on the chart at all, because none of these totals include the channel that actually brings people to you.

The $29.95 line that changes the arithmetic

The cheapest line in this whole comparison is a directory profile, and in our view it is also the one doing the most work. Psychology Today states its price in one sentence on its own signup page: members pay a fixed monthly fee of $29.95, with no contracts and the option to cancel at any time. That membership also includes a telehealth video platform the company describes as free and HIPAA compliant.

The second half of that claim is a judgment rather than a statistic. Nobody publishes a reliable national breakdown of how therapy clients find their clinician, so anyone quoting you a percentage is guessing, and we are not going to add another guess to the pile. What is verifiable is the price, and the price is the part that should worry you.

Set that beside the plans above and the picture inverts. The listing that gets you found costs less than half the cheapest website plan and under a tenth of the dearest one.

The black bars are the two purchases a practice usually makes before it thinks about a website: the directory profile, and practice management software. SimplePractice publishes three plans at $49, $79 and $99 a month, currently with half off the first three months. Its own feature lists for those plans name paperless intake documents, online appointment requests, appointment reminders, electronic claim filing and something it calls a Client Portal website and mobile app. The booking and paperwork half of what people picture a practice website doing is therefore already bought before the website conversation starts.

So the honest framing of the website decision is not what should a therapist website cost. It is what should a therapist website do, given that something cheaper already handles being found and something else already handles booking. That question has a good answer, and it changes what you should be willing to pay.

What the site is actually for

Three jobs, and being found is not one of them.

The first is the second look. Somebody reads your directory profile, then types your name into Google before writing to you, because choosing a therapist is a high-trust decision made by an anxious person. Your site is what they find on that second look, and it is the only version of you they can read without a matching algorithm sitting in between. If it loads slowly on a phone, says less than the profile did, or looks like the template of every other practice in your city, it has cost you the referral rather than earned it.

The second is the referrals you already have. A physician, a school counselor, a former client and a colleague who is full all pass your name along, and none of those people arrive through a directory search. They arrive by typing your practice name. Everything that makes a first session easier to say yes to belongs on that page: your fee, your availability, what the first appointment is like, where to park, whether you use insurance and how.

The third is ownership. A directory profile is rented and the terms can change. A profile on an insurance platform belongs to the platform. Your website and your domain are the only pieces of your practice's presence that no company can reprice or switch off, which is worth something the day a directory raises its rate or your specialty gets crowded. We wrote the long version of that argument in who owns your website.

Judge every price on the chart against those three jobs, and most of the top of the range stops making sense for a solo practice. What the expensive plans really sell is content marketing and social posting, which is a different purchase from a website and deserves to be evaluated as one.

The one page a federal rule puts on your site

If you see any client who is uninsured or paying out of pocket, and most private practices see at least some, a federal regulation requires a specific notice to be on your website and to be findable by a search engine. It is the most concrete website requirement in this article, and not one of the cost guides we read while researching this piece mentions it.

The Good Faith Estimate rules under the No Surprises Act, at 45 CFR 149.610, require a provider to inform every uninsured or self-pay individual that a good faith estimate of expected charges is available. Paragraph (b)(1)(iii)(A) says how: the information must be written in a clear and understandable manner, prominently displayed and easily searchable from a public search engine on the provider's website, in the office, and on-site where scheduling or questions about cost occur. Paragraph (B) adds that it must also be provided orally when someone schedules or asks about cost, and paragraph (C) that it must be available in accessible formats and in the languages your clients speak.

Read the phrase easily searchable from a public search engine again, because it is doing real work. A notice buried in a PDF, hidden behind a form, or dropped into a paragraph on a page that carries a noindex tag does not satisfy it. This is a page that has to exist, be crawlable, and be written plainly. The rule also treats any discussion or inquiry about potential cost as a request for an estimate, so the page saves you administrative time as well as satisfying a regulator.

Two practical consequences for the build. You need a real page, not a footer line, and it should be linked from your main navigation or your fees page rather than orphaned. And whoever builds the site has to leave it indexable, which sounds obvious until you meet a staging site that shipped with search engines blocked.

What your fees page is allowed to say

Fee language on a therapy website is regulated advertising, and the wording most practices reach for is the wording that is specifically ruled out. California's healing arts advertising statute is the clearest example, and it covers everyone licensed under that division of the state's code, therapists and counselors included.

Business and Professions Code 651 makes it unlawful for a licensee to disseminate any public communication containing a false, fraudulent, misleading or deceptive statement likely to induce the rendering of professional services. The definition of public communication in the statute names the Internet and a directory of healing arts practitioners explicitly, so your website and your directory profile are both covered. Subdivision (c) is the one that catches everyday copy: any price advertisement shall be exact, and the statute names the phrases it will not accept, including "as low as", "and up" and "lowest prices".

Put that next to the federal estimate rule and the two point the same way. Publish the actual number. Sixty minutes, this many dollars, sliding scale places available or not. A single exact figure satisfies the state rule, gives the federal notice something to point at, and answers the question that stops most people from writing to you in the first place.

The stock photo rule almost nobody knows

Open a dozen therapy websites and most will show a purchased photograph of somebody who reads as a client. In California that carries a labeling requirement. The same statute, at subdivision (b)(3)(B), says that using a photograph or image of a model without clearly stating in a prominent location in easily readable type that the image is of a model is a violation. A model, for that purpose, is anyone other than an actual patient of the licensee advertising the service.

The rule was written with cosmetic before-and-after advertising in mind, and it sits in a statute that covers every healing arts licensee. Whether an enforcement action would ever follow from a calm stock photograph on a counseling page is a fair question to ask your own board. The design conclusion holds regardless, and it is the one we would give anyway: photographs of real rooms, real light and the actual person doing the work convert better than a stranger looking thoughtful on a sofa. A client is choosing a human being. Show them the human being.

Testimonials, and the box you cannot tick

Social proof is the standard tool for a small service website, and it is largely unavailable to you. A review from a psychotherapy client discloses that the person was in psychotherapy, so asking for one raises a confidentiality problem before it raises an advertising one, and the exact limits sit in your own board's rules and your association's ethics code rather than in anything a web designer can tell you. The same California statute reaches it from the advertising side: a statement or claim is deceptive if it is likely to mislead or deceive because of a failure to disclose material facts, which is exactly what a curated wall of the happiest outcomes does.

The federal layer applies too. Testimonials and reviews on a business website sit under the Federal Trade Commission's 2024 rule and its endorsement guides, which we covered in detail in what you can and cannot publish. The point for a practice is that the usual workaround, a page of five-star quotes with first names, is the wrong tool here.

What replaces it is more work and better material. Write about how you actually practice, including the modalities you use and the ones you do not. Say who you are a poor fit for. Publish a genuine answer to what happens in the first session. Colleagues, physicians and past clients making referrals are reading for competence and fit, not for stars.

Read the add-on list to learn what the base plan is not

Both website vendors price HIPAA features as extras, and that pricing is the most useful disclosure on either page. Brighter Vision sells HIPAA compliant email at $14.99, $19.99 and $24.99 a month, or $165, $220 and $275 a year, and its HIPAA compliant forms arrive with the $349 Flourish plan rather than the cheaper ones. TherapySites prices secure emails and forms from $15 a month on top of any plan.

Which tells you the thing to check before you sign anything: an ordinary contact form on an ordinary website is not a secure channel, and a prospective client will absolutely type clinical detail into it. Either the form goes to a compliant vendor, or the form asks for a name and a phone number and nothing else, with a line saying not to include health information. Both are defensible. A standard form with a message box and no agreement in place is not.

Note also that no website is certified compliant. TherapySites lists HIPAA compliant and ADA friendly as Core features, and both phrases describe an intention rather than a certificate that exists. Nobody certifies a website against either standard, a point we made at length about accessibility and about HIPAA. Ask instead for the specifics: who signs a business associate agreement, where form submissions are stored, and whether the site has ever been tested with a keyboard and a screen reader.

Where you are allowed to say you practice

A service area page is standard practice for a local business site, and for a licensed clinician it is a compliance question rather than a marketing one. This is also the fastest moving fact in this article.

The Counseling Compact is switching on one state at a time. Its own status notice, updated 30 July 2026 when Arkansas joined, lists it as live in Arkansas, Arizona, Georgia, Indiana, Louisiana, Minnesota and Ohio. A counselor licensed in and residing in one of those states can apply for the privilege to practice in the others. The same page says 31 additional states and the District of Columbia are working through the steps needed to issue and recognize privileges.

So the copy on your site has a shelf life. If you are licensed in a live compact state, your reach genuinely changed this summer and your website should say so. If you are not, writing that you serve clients across several states is a claim your board may read differently than your marketing did. Date that page, put the states in a list rather than a sentence, and put a reminder in your calendar to re-check it while this is still moving.

What this costs next to what you earn

The prices in this article are small numbers to an agency and large numbers to a solo clinician, so it is worth putting them beside the income of the people paying them. Federal wage data settles the comparison. The May 2025 occupational survey counts 491,930 substance abuse, behavioral disorder and mental health counselors earning a median of $59,350 a year and a mean of $64,440. Alongside them are 132,810 mental health and substance abuse social workers at a $60,280 median, 66,740 marriage and family therapists at $66,940 with a mean of $76,960, and 75,990 clinical and counseling psychologists at $100,580 with a mean of $112,750.

Against a $59,350 median for a mental health counselor, the $69 a month plan is roughly one and a half percent of a year's wage and the $349 plan is about seven percent. That is the real decision. Seven percent of gross earnings, every year, indefinitely, for a website plus content and social posting that a practice with a full caseload may not need at all.

The same numbers explain why the subscription model dominates this niche. A one-time build is a difficult purchase on a counselor's income, and $69 a month is an easy yes. That is a legitimate service and for many practices it is the right call. It is worth knowing that you are choosing to pay indefinitely for something that could be bought once.

What we would spend, in order

For a solo or small group practice, the sequence matters more than the vendor.

Start with the directory profile, because at $29.95 it is the cheapest client acquisition in this article and, in our experience, it outranks a brand new practice website for the searches that matter. Write it properly. Specific presenting problems, a real photograph, the fee, and availability that is actually current.

Then buy or build a site you own, once, with a small and honest set of pages: who you are, what you treat, what it costs, the good faith estimate notice, how to start, and where you are licensed. Six to eight pages that answer real questions beat a twenty page template every time. At published rates that is somewhere between about $830 for a year of the cheapest hosted plan and a few thousand dollars for a build.

Only then consider content. The $349 tier buys you one optimized page a month. Written well by somebody who understands clinical nuance, that can compound into real search traffic over a year or two. Written as filler, it is a subscription to nothing. Ask to read three articles the vendor produced for other clients before you agree to it, and judge them the way you would judge a colleague's writing.

Ads come last, if at all. TherapySites prices Google Ads management from $600 a month plus setup, before a dollar of ad spend. For a practice with a waiting list that is money burned, and for one with empty slots it is worth testing only after the profile and the site are doing their jobs.

Our offer, and the practices we are wrong for

We have two ways in, set out on our offer page: a flat one time build you own, and a monthly partnership when there is continuing work to do. Most focused builds take about two weeks from kickoff to launch. Email us and you will hear back within 48 hours. There is more on how we approach this field on our mental health page. Outside private practice, the outcome we can point to is a treatment center rebuild: Cornerstone Healing Center ended up with 20% more website conversions.

Now the honest part. A solo practice with a full caseload and a working directory profile does not need us, and probably does not need a new website at all. If $69 a month is the shape of spending that fits your cash flow, a hosted therapist plan is a reasonable purchase and we are not going to pretend otherwise. And if what you want is somebody writing a clinical blog every month and running your social accounts, hire a marketing service built for this field, because that is their product and it is not ours.

The work we are good at begins where the website itself is the bottleneck: a group practice whose clinicians cannot be found individually, a niche practice that needs the site to do the explaining the directory cannot, a rebuild where the previous vendor holds the domain, or a practice adding services where the site has to grow rather than be replaced again.

The arithmetic that settles it, using your own numbers

Every number in this article becomes decidable when you put it next to your session fee, which is a figure only you have. Do this on paper in five minutes.

Write down your standard fee for a fifty minute session. Multiply it by the average number of sessions a client completes with you, which you can get from your own records, and you have the value of one new client. Now take the annual cost of each option in the table, and divide.

Say a client is worth roughly a thousand dollars over their course of work. The $828 plan needs to bring you one client a year to have paid for itself. The $4,188 plan needs four and change, every year, on top of what the directory was already sending you. A one time build needs one and a half in the first year and nothing at all in the second.

Then ask the question that number cannot answer on its own: where did your last ten clients actually come from? If the answer is the directory and word of mouth, spend accordingly and keep the website small, sharp and owned. If the answer is that people found you by searching and read three of your pages before writing, the case for the bigger plan just got made with evidence rather than a sales call.

So here is the version built only from prices the companies charging them publish on their own websites, and rules quoted from the source that wrote them. Everything below was read on 15 August 2026 and linked where it appears. Websites are what we sell, home service companies included, so treat this as a page written by one of the options on it. There is a section near the end naming the situations where a different company is the better buy. There is also a change happening to Google Local Services Ads this month that almost every article on this topic still describes wrongly, and it matters more to your phone ringing than any design decision you will make.

The 30-second answer

An HVAC company website costs between roughly $3,000 and $28,000 in its first twelve months on a plan somebody publishes openly, and the spread has almost nothing to do with the website. At the bottom is a flat monthly program with the site included and no build fee. In the middle is a subscription with a one-time build charge of about $2,000 to $4,500. At the top is a custom build billed over twelve months by an agency that also sells you the marketing around it. A one-time custom build you own outright, ours included, sits below the bottom of that range and carries no monthly agency fee, only hosting. What decides your real cost is not the page count. It is whether you are also renting the channel that brings the calls, and whether the site is yours when you stop paying.

What the published market actually charges

Three companies serving this market put real numbers on a public page, which makes them the only honest basis for comparison. They are selling different things, so the table says what each price includes.

Footbridge Media sells a flat $249 a month program to HVAC contractors, with $0 setup fees, no contracts and a 90 day money back guarantee, and the price includes a custom website with unlimited pages alongside SEO, Google Business Profile management, review automation and a contractor CRM. The company says it has served more than 2,000 contractors over more than 20 years in business.

Websites for HVAC runs the subscription model: four tiers at $399, $639, $1,039 and $1,999 a month on an annual term, each with a one-time build and territory lock fee of $1,995, $2,749, $3,499 or $4,499. Month to month costs $100 to $500 more per month and $300 to $500 more on the build fee. The plans are sized by pages and cities rather than by features, from five pages and one city up to fifty pages and ten cities, and the page states that you own the content and domain.

Hook Agency, which works with contractors, publishes a standard templated website at $1,000 a month described on the page as $12,000 total split into twelve months, and a fully custom site at $2,000 a month, $24,000 total, with a note that more complex sites run up to $50,000. Its pricing FAQ answers the ownership question directly: you own your website 100%, and it adds that many companies do not operate that way.

Two things are worth noticing before you use that chart to pick a tier. First, the gap between the cheapest and the most expensive published plan is roughly nine and a half times, and every one of them delivers a website that a homeowner would call good enough. Second, none of these numbers include the money that actually buys phone calls.

At most agencies, the website is the cheapest thing they sell you

The website is not where a contractor's marketing budget goes, and one agency's own price list proves it better than any survey could. Hook Agency publishes six prices on the same page, and the two website lines are the lowest two on it.

Read the two aqua bars first. A templated site costs $1,000 a month for a year and then stops. Local SEO costs $2,800 a month and does not stop. Answer engine optimization, the newer service aimed at getting your company named inside AI answers, is listed at $4,000 a month, or $2,000 when added to an SEO package. Paid search management starts at $2,000 a month before a cent of ad spend, and the page notes the management fee rises as spend rises.

Twelve months of the custom site plus twelve months of local SEO is $57,600 at published rates, and only $24,000 of that is the website. This is the single most useful thing to understand before you ask anyone for a quote. You are not choosing a website price. You are choosing how much of your lead generation to outsource, and the site is the small line at the bottom of that decision.

The channel a lot of contractors lean on is changing this month

Google Local Services Ads, the pay per lead unit that sits above the normal search results with the badge next to it, is being folded into Google Ads right now, and home services is the first group to move. Google's own migration notice says the first phase began in August 2026 for select home and storefront service advertisers in the United States, and it names the categories: plumbing, HVAC, electrical, appliance repair, house cleaning, lawn care, roofing, pest control and moving. Broader groups follow in late 2026, and non US accounts plus everything remaining move in 2027.

The pay per lead model survives the move. You still pay for valid leads rather than clicks, the ads still appear only on Google Search and Google Maps, and the campaign stays keywordless, targeted by the services and areas you choose. Verification carries over, and Google says insurance and license reverification are not required.

Four changes are worth putting in your calendar rather than discovering later. Manual bidding and industry level target CPA are being deprecated, so whatever bid strategy you tuned by hand gets replaced. Better Business Bureau callouts are gone, and Google's instruction is to pick at least six other structured callouts instead. Your old Local Services dashboard closes on migration day and previous performance reports do not carry over, so anything you want to keep has to be downloaded first. And the account administrator gets an email 14 days before migration with a reminder seven days later, which is the entire warning you get.

Google also tells you to expect a wobble: while many ads start running right away, allow up to two weeks for the migration to complete and for campaign performance to ramp back to stable levels. If your booked jobs dip in the two weeks after your migration date, that is the reason, and it is not a reason to fire your web designer.

The badge changed too, and the guarantee behind it is gone

The Google Guarantee money back promise is being discontinued, which is news to almost every marketing page still selling it as a reason to advertise. Google's help page on the new badge says the company is simplifying advertiser badging into a single Google Verified badge, and then says plainly that it will be discontinuing the Money Back Guarantee associated with the Google Guarantee badge. The page limits reimbursement to services booked through Local Services Ads before 7 December 2025, with requests due within 30 days of the initial service completion date, which leaves almost nothing open by now.

What remains is the screening, and the screening is the part that was always worth something. Google describes it as an extensive verification process that varies by category and location and may include background, business registration, insurance and license checks, plus minimum review requirements. Existing advertisers keep their badge automatically. New ones earn it by passing those checks.

There is a ranking consequence for anyone starting out. Google offers pre badge ads that let you take leads while you finish onboarding, but says your listing will appear below providers who have completed all onboarding requirements. So the order of work for a new contractor is not website, then ads. It is licenses and insurance documents and first reviews, then ads, then the website that catches everything else.

The two free things that decide whether the phone rings

They are your Google Business Profile and your reviews, and the reason to start there is that neither one is for sale. Google's guidance on local ranking states it flatly: there is no way to request or pay for a better local ranking on Google. Local results are mainly based on relevance, distance and prominence, where prominence is partly built from how many websites link to your business and how many reviews you have.

Read those three words as a budget. Distance you cannot change. Relevance is a complete, accurate Business Profile with the right category and hours, which is an afternoon of work and no money. Prominence is reviews and links, which is a habit rather than a purchase. A website contributes to the third one and to nothing else on that list.

The service area rules matter here too, because HVAC companies are usually service area businesses rather than storefronts. Google's representation guidelines allow one profile for the central office with a designated service area, say that a rented mailing address you do not operate from is not eligible, and put a limit on reach: the boundaries of a profile's service area should not extend farther than about two hours of driving time from where the business is based. If a vendor is selling you visibility three states wide, that is not how the map works.

About those forty city pages

City landing pages are one of the most commonly sold upsells in HVAC web design, and the pattern most vendors ship is the one Google's spam policies describe by name. The policy on doorway abuse lists, as an example, having multiple domain names or pages targeted at specific regions or cities that funnel users to one page, alongside creating substantially similar pages that are closer to search results than a clearly defined, browseable hierarchy.

That is not a ban on writing about the towns you serve. It is a description of what happens when twenty pages differ only by the place name in the headline. The distinction a homeowner can see is the same one the policy draws. A page about the neighborhood with your actual response time to it, the systems common in its housing stock, a job you did on that street and the permit office you deal with is a real page. A page produced by find and replace is a template with a zip code in it.

This is worth checking before you buy a plan sized by city count, because that is exactly how several published plans are priced. Ask for three live examples of city pages the vendor has produced for another client, read them side by side, and count the sentences that differ. If the answer is one, you are buying pages that will not earn their keep, whatever the invoice calls them.

What you own when you stop paying

On a subscription site, the honest question is not the monthly fee, it is what survives cancellation. Two of the vendors above answer it in writing, which is worth more than a reassuring sentence on a sales call. Hook Agency's FAQ says you own your website 100% and observes that many companies do not work that way. Websites for HVAC states on each plan that you own the content and domain.

Notice what those sentences do and do not cover. Owning the content and the domain is not the same as owning the build, and none of these pages promise you the theme, the page templates or an export you can host elsewhere. Nor is exclusivity free of consequences: Websites for HVAC sells a territory lock, so the value you are buying is partly that the vendor will not sell the same thing to the shop across town, which also means your bargaining position at renewal is weaker than it looks.

Three questions settle it. Ask who the domain is registered to today. Ask what specifically you receive as files if you cancel, and whether that includes the design or only the text and images. Ask whether your Google Business Profile, your Google Ads account and your call tracking numbers are owned by your company or by the agency, because those are the assets that carry your history. We wrote a longer version of this argument in our piece on who owns your website, and the short version is that the contract decides, not the invoice.

The line item that is never in the quote

In most states your website is advertising in the legal sense, and your license number belongs on it. California's rule is the plainest example. Business and Professions Code 7030.5 requires every licensed contractor to include the license number in all construction contracts, in subcontracts and calls for bid, and in all forms of advertising as prescribed by the registrar of contractors. A website is a form of advertising, and so are the Facebook page, the truck wrap and the pay per click ad.

Requirements differ by state and by trade, and some states add rules about which license classification you may name and how a company name may be presented. This is a ten minute check with your own licensing board that no web designer will do for you, and it costs nothing until an inspector or a competitor notices. Put the number in the footer of every page and in the same place on every profile you control, and the whole issue disappears.

While you are there, two other pieces of text pay for themselves on an HVAC site: the service area written as the towns you actually serve, and your after hours policy stated in words rather than implied. Emergency work pays well, and a homeowner who cannot tell whether you pick up at nine at night will call the company whose site says so.

What the same money buys in labor

Set the website prices next to a payroll line and they stop being abstract. The Bureau of Labor Statistics occupational survey for May 2025 counts 409,670 heating, air conditioning and refrigeration mechanics and installers in the United States, with a median annual wage of $61,010 and a mean of $64,780.

A $12,000 website is about ten weeks of a median technician's wage before employer costs. A $2,988 flat program is about two and a half weeks. The $57,600 custom site plus SEO package is closer to a full year of one technician, and unlike the technician it does not carry a truck to a job. That comparison does not decide anything on its own, because a website works while you sleep and a technician does not. It does put the decision in a currency you already price in your head every time you consider adding a van.

What we would actually spend

For a single location shop doing residential service and replacement, the order is not the one most vendors sell. First, spend nothing and fix the Business Profile: correct category, correct hours including the ones you really answer the phone, a service area that matches your driving reality, photos of your own trucks and crew, and a standing habit of asking every completed job for a review. That is the highest return work available to you and it has no invoice.

Second, buy a site you own, once. Ten to fifteen genuinely useful pages beat forty templated ones: what you fix, what it costs, how fast you come, who you are, the towns you cover written like a human wrote them, and a phone number that is impossible to lose on a phone screen. That is a one time build, not a subscription, and at published rates it is between about $1,500 and $12,000 depending on who does it.

Third, buy leads deliberately and count them. Local Services Ads charges per valid lead, and Google says lead prices vary by location, job type, lead type and bidding mode, so nobody can quote you a national number honestly. What you can do is set the weekly budget low, run it for a season, and divide what you spent by the jobs you booked. That one division tells you more than any proposal you will be sent.

Fourth, only after those three, consider a retainer. A $2,800 a month SEO engagement is a reasonable purchase for a company with several vans and a growth target, and a poor one for a shop that has not yet claimed its own profile. The order matters more than the vendor.

Our two prices, and the shops we are wrong for

Two numbers, both public. A flat fee buys a one-time build, and a monthly partnership covers work that carries on after launch. Our offer page sets out what each includes. Focused builds are usually live about two weeks after kickoff, and we answer partnership emails inside 48 hours. Home service companies can see how we approach this market on our local services page. One result we can point to outside this trade: a treatment center we rebuilt the site for, Cornerstone Healing Center, ended up with 20% more website conversions.

Here is where a different company beats us. If your problem is that nobody in your market knows your name, a website will not fix it and a full marketing program from a contractor specialist probably will, so buy the program and skip the design conversation. If you want one company answering for the site, the ads, the reviews and the CRM on one invoice, hire a home services agency, not us, because that bundle is their product and it is not ours. And if you are shopping purely on monthly price, the $249 program undercuts anything we would build and it is a genuinely reasonable purchase. The same subscription pattern runs through the neighboring trades, and when we priced it for auto repair shops the published website rates ran from $49 to $699 a month.

Hire us when the site itself is the constraint: when it is slow on a phone in a driveway, when it does not say what you charge or how fast you come, when it was built by a vendor who still holds the keys, or when you want to own the thing outright and stop renting it. Those are the jobs we are good at, and they are a small fraction of the marketing budget this article has been describing.

Your fourteen day window, and what to do inside it

This is the only deadline in this article with a date attached to it, and it applies to every contractor running Local Services Ads. When the migration email arrives, you have 14 days before your account moves and your old dashboard closes for good.

Inside that window, do five things in an hour. Download every performance report you want to keep, because Google says previous reports do not carry over. Write down your current cost per lead and cost per booked job from that data, so you have a before number that nobody can argue with later. Check which bid strategy you are on, since manual bidding and industry level target CPA are being retired and something automatic will replace them. Pick your six structured callouts to fill the space the Better Business Bureau callout used to occupy. Confirm the phone number your leads route to, because in the new setup you can edit it yourself in real time and that is the one field a wrong entry will cost you money on immediately.

Then leave it alone for two weeks. Google says performance takes up to that long to stabilize, and the temptation to change three things at once during a migration is how contractors end up believing their website broke. It did not. The channel moved, and the only thing you needed was the fourteen days of notice you just used properly.

So here is the version built from arithmetic instead of ranges. Every price below is published by the company charging it, every rule is quoted from the rule, and every penalty is one a firm actually agreed to pay. All of it was read on 14 August 2026 and linked where it appears. We build websites for a living, including for firms in this market, so we are one of the options on this page and there is a section near the end about when to pick somebody else.

The 30-second answer

A financial advisor website costs between about $1,700 and $16,000 in its first year on a published plan, and the spread has less to do with design than with how much of the work somebody else does. Specialist platforms start near $99 a month plus a setup fee. Custom builds and full marketing retainers run into the thousands a month. What none of the cheap quotes include, and what actually decides the true cost, is the compliance work: your website is an advertisement under SEC rules, every version of it is a record you must keep for five years, and the SEC has fined firms six figures for what was on the page. Budget for the site, then budget for who is responsible for what it says.

Your website is not marketing collateral, it is a filing

Start here, because it explains every price on this page. For a registered investment adviser, the website is an advertisement in the legal sense. The SEC Marketing Rule defines an advertisement as any direct or indirect communication an adviser makes to more than one person that offers its advisory services to prospective clients. A public website is exactly that, on every page, all the time.

The rule was adopted at 86 FR 13024 and took effect on 4 May 2021, and when it adopted the rule the Commission set a compliance date 18 months after the effective date, which lands in November 2022. It has been fully in force for close to four years. That matters when a vendor tells you their template is compliant, because the template is not what the rule regulates. The words on it are.

The practical consequence for a budget is simple. In most industries a website is a design project with a marketing goal. Here it is a design project with a supervision requirement attached, and the supervision does not stop at launch. Every price below should be read as a price for a system that keeps producing compliant pages, not for a one-off delivery.

What the specialists actually charge

Four companies serving this market publish their prices in full, which makes them the only honest basis for a comparison. Their plans are not identical, so the table lists what each one says you get.

Snappy Kraken publishes website plans at $99 a month on an annual term with a $499 setup fee, and $199 a month with a $2,499 setup fee for a premium template with custom design. Worth knowing when you search: advisorwebsites.com, one of the oldest names in this category, now redirects to Snappy Kraken's pricing page.

FMG Suite prices in three tiers, described on its page as starting at $178, $418 and $1,044 a month, each with a setup fee between $994 and $3,494. Advisor Designs takes the opposite approach with two plans built around page count, $75 a month plus $995 for a five-page site and $115 a month plus $1,395 for eight pages, billed annually, with further edits charged by the hour. Paladin Digital Marketing sits at the top of the published market at $950, $2,950 and $4,950 a month with no onboarding fee and no annual term, and lists compliance and automatic archiving inside the monthly price.

Two things jump out of that chart. The first is that the cheapest and the most expensive published options differ by roughly ten times, and none of it is explained by how the site looks. The second is that the annual-term prices are annual-term prices. Advisor Designs states that all plans are billed annually, and Snappy Kraken's monthly figures are labeled as the annual-term rate, so treat the first-year number as the real number and ask what month 13 costs.

What a custom build buys that a template does not

Templates in this market are good and getting better. The honest case for paying more is narrower than agencies pretend, and it comes down to three things.

The first is the page count you are allowed to have. Advisor Designs prices five pages and eight pages as separate products. Snappy Kraken's entry plan allows up to 20 basic pages and its higher plan removes the limit. If your growth plan depends on writing, or on a page for each service and each town you serve, a page cap is not a detail. It is the ceiling on your search strategy, and you will hit it in year two.

The second is who can change what. Entry plans include a fixed allowance of help, then charge by the hour. Advisor Designs publishes $100 an hour on its lower plan and $75 an hour with 30 minutes included on the higher one. That is fair pricing, clearly stated, and it tells you something important: on a template plan, changes are a transaction. On a custom build you own the code and any competent developer can work on it, which is worth more the more often you expect to change things.

The third is whether the site is yours at all. A vendor plan is a rental, and rentals end. Before you sign anything, find out who holds the domain, who holds the content, and what leaves with you, which we covered in detail in our piece on who owns your website. In this market it matters more than most, because your compliance archive has to survive the move as well.

The rule that decides your testimonial page

Client reviews are the single most requested feature on an advisor website. The SEC described its own new rule as one that would permit the use of testimonials and endorsements, subject to certain conditions, which tells you plainly enough where the industry stood before it. Those conditions are design decisions before they are legal ones.

Read paragraph (b)(1)(i) again and notice what it does not say. It does not say disclose somewhere on the site. It says clearly and prominently, at the time the testimonial is disseminated, that the person is a current client, that they were paid if they were paid, and that there is a conflict if there is one. A modal, a footnote at the bottom of a long page, or a small asterisk linking to a disclosures page is a design pattern that argues with the rule.

There are two more conditions that quietly shape the build. If the testimonial is compensated in any way, the rule requires a written agreement with the person giving it, describing the scope of the activity and the terms of payment. And where no compensation, or only de minimis compensation, is involved, paragraph (b)(4) lifts some of those conditions, which is why an unpaid quote from a current client is the simplest thing to publish and a paid influencer arrangement is the most complicated.

None of that makes a review section impossible. It makes it a component with rules: the badge that says current client, the line that says paid or unpaid, the conflict statement, all rendered next to the quote rather than parked on another page. Build it once properly and it costs nothing extra to use it forever. This is a different regime from the one ordinary businesses live under, and if you also run a non-advisory brand it is worth reading our piece on testimonials and the FTC rules to see how differently the two are policed.

Awards, rankings and the badge from 2019

Third-party ratings are the other thing every advisor site wants above the fold, and they carry their own paragraph. Under paragraph (c) of the Marketing Rule, a rating may not appear unless the adviser has a reasonable basis for believing the questionnaire behind it was not built to produce a predetermined result, and unless the advertisement clearly and prominently discloses the date the rating was given, the period it covers, the identity of whoever created and tabulated it, and any compensation the adviser paid in connection with getting or using it.

In practice that turns a decorative logo strip into a small content problem. Each badge needs a year, a period and a source next to it. It is a five-minute build decision that most templates get wrong by default, because a logo row is designed to look clean rather than to carry four facts per item.

It is also, as the next section shows, one of the specific failures named when the SEC charged nine firms in 2024.

Numbers, calculators and anything that looks like a projection

Performance is where the website meets the part of the rule with the sharpest edges. Two provisions matter for a build.

Paragraph (d)(1) says gross performance may not appear unless net performance appears with at least equal prominence, in a format designed to facilitate comparison, over the same period and using the same methodology. Equal prominence is a layout instruction. It rules out the common pattern of a big number in the hero and the qualifying figure in the small print underneath.

Then there is hypothetical performance, which the rule defines broadly enough to catch things that do not feel like performance at all. It includes model portfolios, backtested results, and targeted or projected returns. A hero panel showing what a strategy would have returned since 2015 is hypothetical performance. So is a projection of what a portfolio might do.

The rule does carve out interactive tools, and the carve-out is the most useful paragraph on this page for anyone planning a retirement calculator. A tool where the visitor produces their own simulation is not hypothetical performance, provided the adviser describes the criteria and methodology including limitations and key assumptions, explains that results vary with each use and over time, describes the universe of investments considered and any selectivity where that applies, and discloses that the outcomes are hypothetical in nature. Four disclosures, attached to the tool. Build them in and the calculator is fine. Leave them out and you have published hypothetical performance to the general public, which is precisely what the SEC went looking for.

What firms have actually paid, and for what

This is not theoretical enforcement. Three separate sweeps have now ended in settled charges, and in every one of them the conduct was on a website.

In September 2023, nine advisers were charged for advertising hypothetical performance to the general public on their websites without the policies and procedures the rule requires, and agreed to $850,000 in combined penalties, ranging from $50,000 to $175,000 each. Two of those firms were also found to have failed to maintain required copies of their advertisements, which is the record-keeping point below showing up in a real case.

In April 2024, five more advisers settled for $200,000 combined over the same hypothetical performance problem. Four of the five received reduced penalties because they had fixed the issue before the SEC staff contacted them, which is the cheapest sentence in this article to act on.

Then in September 2024, nine firms agreed to $1,240,000 in combined civil penalties over advertisements with untrue or unsubstantiated claims, or testimonials, endorsements and third-party ratings that lacked the required disclosures.

The findings are worth reading as a build checklist, because they are so ordinary. Four firms claimed to provide conflict-free advisory services and could not substantiate it. One published an advertisement claiming membership of an organization that did not exist. One presented two testimonials that did not come from current clients, and ran endorsements without disclosing that the endorser was a paid non-client. And four included third-party ratings, some of them more than five years old, without disclosing when the rating was given or what period it covered.

Not one of those is a technology failure. Every one of them is a sentence somebody wrote, or a badge somebody left in place after it went stale. That is the actual cost center in an advisor website, and no template prevents it.

The line item that never appears in a quote

Here is the part almost every published cost guide omits entirely. Under 17 CFR 275.204-2(a)(11), an adviser must keep a copy of each advertisement it disseminates. Your website is an advertisement. Therefore your website is a record.

The retention period sits a little further down the same rule: records made under that paragraph must be kept in an easily accessible place for not less than five years, the first two in an appropriate office of the adviser. Not the current version. Each version, for five years.

Think about what that means for a site that publishes a market commentary every week. Every one of those pages, in the form it appeared, needs to be retrievable. A content management system with a revision history is not the same thing as an archive, and a vendor plan that ends when you stop paying is emphatically not the same thing.

Ask any prospective vendor two questions: does the plan capture and retain page versions, and can you export the archive if you leave? Paladin lists automatic archiving inside its monthly price. Snappy Kraken and FMG both sell compliance workflow features. One thing we could not do is price the specialist archiving vendors, because none of the ones we tried publishes a rate openly. That makes this a line item you will have to price by asking, and we would rather say so than invent a range.

If you are with a broker-dealer, add three more things

Everything above applies to registered investment advisers. If you are a registered representative of a broker-dealer, FINRA's rules sit on top, and three of them have direct consequences for a website project.

First, approval. FINRA Rule 2210 defines a retail communication as any written, including electronic, communication distributed or made available to more than 25 retail investors in any 30 calendar-day period, and requires an appropriately qualified registered principal to approve each one before the earlier of its use or its filing. Your launch date is therefore not yours alone. Build the review window into the schedule instead of discovering it in launch week.

Second, and this one surprises people, new firms have to file first. For one year from the date FINRA membership became effective, a member must file with the Advertising Regulation Department at least 10 business days prior to first use any retail communication published in electronic or other public media, and the rule names a generally accessible website among them. A new broker-dealer cannot simply launch a site on a Tuesday.

Third, the BrokerCheck link. Rule 2210(d)(8)(A) requires each of a member's websites to carry a readily apparent reference and hyperlink to BrokerCheck on the initial webpage intended for retail investors, and on any page carrying a professional profile of a registered person who does business with retail investors. That is a template decision affecting your home page and every bio page.

Filing also has a price, and it scales with the size of the site. Section 13 of Schedule A to the FINRA By-Laws sets the review charge at $300 for printed or website material, plus $10 for each webpage past the first ten, with expedited review at $600 plus $50 a page and a three business day turnaround. The current schedule dates from 1 January 2025, so check the rule rather than an older article.

What turns up on the invoice later

Four costs land after the project everyone budgeted for is finished.

Edits by the hour. Every template plan meters help. If your practice generates a few changes a month, an hourly rate at $75 to $100 with a small monthly allowance is fine. If you are actively marketing, it is a slow leak, and the monthly plan one tier up is usually cheaper than the hours you will buy.

The content treadmill. A website that is finished stops earning within a year. Somebody has to write, and in this industry somebody also has to approve. The done-for-you tiers exist because that combination is the hardest part to sustain, which is why the top of our chart is dominated by content programs rather than by design.

The archive. See above. If it is not in the plan, it is a separate purchase.

Maintenance and the platform itself. This is the same set of bills every business pays after launch, and we broke it down in our piece on website maintenance cost. Nothing about being regulated makes those cheaper.

What to spend, by where you actually are

Three honest positions, with a recommendation for each.

Newly registered, no marketing budget, few clients. Buy the cheapest published plan that does not cap you below about 20 pages, spend nothing on custom design, and put the time into your disclosures and one page that explains who you serve. At roughly $1,700 for the first year, the site is not the constraint on your growth yet.

The middle case is the one people get wrong. An established practice with referrals and a website nobody visits. Here the site is doing real work and doing it badly. The published mid-tier plans cost $2,700 to $4,900 in year one, and a custom build can land in the same range or under it, so the question stops being price and becomes control: page count, edit access, and whether you own what gets built.

A firm running actual campaigns. If you are buying ads, running events, and publishing regularly, the site is one component of a system and it should be budgeted as such. That is the $11,000 to $16,000 first-year territory on the published plans, and at that level you are mostly buying labor, not pixels. Compare it honestly against hiring, which we did the arithmetic for in our comparison of in-house against agency.

One piece of context for all three. The Bureau of Labor Statistics puts the median annual wage for personal financial advisors at $105,070 in May 2025, across 266,800 people. Set any of these numbers against a year of your own time and the cheap plans stop looking like savings and the expensive ones stop looking outrageous. The real question is which one removes work from your week.

How we price this, and when to pick somebody else

Our offer is not tiered. A one-time build is a flat fee, and there is a monthly partnership when the work carries on past launch. Our offer page sets out what each includes. Most focused builds are live within two weeks of kickoff. Anyone who writes in about a partnership hears back inside 48 hours. Across 200-plus projects the pattern we keep seeing in regulated markets is that the website is rarely the weak part; the review process around it is. Work we did for Cornerstone Healing Center, a treatment center, produced 20% more website conversions after the redesign.

Now the part that saves us both time. If your firm's compliance department mandates a specific vendor platform, and many broker-dealers do, buy that platform and stop reading comparison articles. You will not win that argument, and the approved platform is genuinely the cheaper path when the alternative is a review process that never ends. If what you need is a compliance archive above all else, buy from somebody who sells archiving as their product rather than as a feature. And if you have no marketing plan yet, a new website will not create one; spend the first $1,500 finding out which clients you want more of, then build a site that speaks to them.

We are a good fit when the site itself is the bottleneck: when it is slow, when it does not explain the practice, when nothing on it has changed in three years, and when you want to own the result. Firms in this market can read how we work on our finance page, and if you are weighing a rebuild against a rescue, our piece on redesigning without losing rankings covers the part where rankings are at stake.

Put one name against each of these six things

Finish with this rather than a checklist, because in every enforcement action above the failure was ownership, not knowledge. Somebody assumed somebody else was watching the page.

Write down six items, and put a single human name next to each one. Not a company, not a department. A name.

1. Who approves new page copy before it goes live. If the answer is the web designer, you have a problem, because they are not the one the SEC will ask.
2. Who checks that testimonials carry their disclosures. Including when a new one is added six months from now by whoever handles the inbox.
3. Who owns the age of every award badge. Ratings go stale silently, and a date on the page is the whole fix.
4. Who signs off anything containing a number. Performance, projections, calculators, and the marketing line that quietly implies a return.
5. Who holds the archive, and where. Name the person and the system that can produce last March's version of your homepage.
6. Who is responsible for the site being current in twelve months. This is the one nobody claims, and it is why so many advisor websites read like 2021.

If the same name appears six times and it is yours, that is workable but fragile. If any line is blank, that blank is the most expensive thing on your website, and it costs nothing to fill in this afternoon.

Which turns the question into a better one. Not how to lose nothing, but which losses are temporary and which are permanent, because they have different causes and only some of them belong to whoever builds the site. Everything below is drawn from documentation published by Google or by the platform being described, read on 13 August 2026 and linked where it appears. We build and rebuild websites for a living, so we are an interested party here, and there is a section near the end on who should hire someone else.

The 30-second answer

A redesign should not cost you rankings permanently, but a temporary wobble is expected rather than a sign of failure. Google's own site move guidance says to expect ranking fluctuation during a move, and that a medium-sized site can take a few weeks for most pages to move in its index. Permanent loss comes from a short list instead: changing URLs that did not need to change, redirects that are missing or of the wrong type, deleting content that was earning the traffic, and launch-day mistakes in the settings that control crawling and indexing. The single most protective decision is also the cheapest one. Keep your URLs.

The dip is normal, the loss is not

Start by separating two things that get reported as one. Reindexing lag is Google catching up with a changed site, and it resolves on its own. Permanent loss is Google concluding that a page is gone, duplicated, or no longer the best answer, and it does not resolve on its own.

Google is unusually direct about the first one. Its guide to moving a site with URL changes tells you to expect temporary fluctuation in site ranking during the move, and adds that with any significant change to a site you may see ranking fluctuations while Google recrawls and reindexes. On timing, its general rule is that a medium-sized website can take a few weeks for most pages to move in the index, and that larger sites take longer.

Two things follow from that. If your agency promises no dip at all, they are promising something the search engine does not offer. And if your traffic is down four days after launch, you do not yet have evidence of anything, which is exactly when most people start making panicked changes that genuinely do cause damage.

The honest planning window is weeks. Set the expectation before launch, agree what you will measure, and hold your nerve through the first couple of weeks unless one of the specific failures below is present. Those you fix immediately.

The safest redesign changes no URLs at all

Here is the lever that matters most, and it is usually decided by accident. A redesign is a visual project, but it becomes a search project the moment the URLs change, and they usually change because the new system has its own idea of what a URL should look like rather than because anybody chose it.

Google draws the same line. It publishes two separate guides, one for a site move with URL changes and one for a site move without them. The second guide is mostly about hosting, DNS and crawl rate. It is a smaller, calmer document, because a rebuild that keeps every address is a much smaller event in search terms. Your pages keep their history instead of having to inherit it through a redirect.

So make URL preservation a requirement rather than a preference, and put it in writing before the build starts. If a page is at /services/roof-repair today, it should be at /services/roof-repair the day after launch, even if the content on it is completely new. When a URL genuinely has to change, that is fine, but it should be a decision somebody made and can explain, not a side effect of a template.

This one requirement removes most of the risk in this article. Everything after it is about the cases where a URL really does have to move.

Not every redirect passes the signal along

When a URL does change, the redirect you use decides whether the new page inherits anything. Google splits redirect methods into permanent and temporary, and only the permanent ones act as a canonicalization signal, which is the technical way of saying only they tell Google the new URL is now the real one.

Google states the practical consequence plainly in its redirects documentation: permanent redirects signal to Google to display the new target URL in search results, while temporary redirects maintain the source page in results. A 302 on a permanent move is therefore not a small stylistic difference. It asks Google to keep the old URL, which is the opposite of what you want.

The same page ranks the methods by how likely Google is to interpret them correctly, and puts server-side redirects at the top. That ordering is the whole recommendation. Do it at the server, use 301, and treat meta refresh or JavaScript redirects as a fallback for cases where you have no server access. One detail catches people out: an instant meta refresh is read as permanent, but a delayed one is read as temporary, so a five second redirect on a moved page is a temporary redirect wearing a disguise.

Your platform decides what you are allowed to redirect

Now the part almost no redesign guide mentions, and the one that can quietly cap how well a migration goes. If you are moving onto a hosted platform, your ability to preserve URLs is a product feature with published limits, not something your developer controls.

Every number there comes from the company's own documentation, which matters because the second-hand versions disagree with each other. Search for the redirect limit on Webflow and you will find several different hard caps confidently stated by people who do not work there. Webflow says it has no hard limit and recommends 1,000 as best practice, because every rule is uploaded to the site's manifest.json file and that file has to be downloaded by visitors' browsers.

Squarespace takes a different approach again, with a single URL mappings field limited to 400 KB, which it says is usually around 2,500 redirect lines. Shopify publishes a maximum of 100,000 URL redirects, rising to 20,000,000 on the Plus plan. Wix does not publish a total, but its CSV import handles 500 at a time, and it says outright that any redirects listed beyond that limit in a file will not be imported, which is the kind of silent failure that shows up as missing pages three weeks later.

There is a subtler constraint shared across all three hosted platforms, and it changes how you test. Squarespace says the original URL can no longer exist for a 301 to work. Shopify says redirects only work for URLs that return 404 errors. Webflow requires you to delete a page, save it as a draft or change its slug before a redirect to a new location will fire. Read together, that means the redirect map on a hosted platform cannot be fully dry-run before launch, because the old URLs still exist until the moment they do not. On a self-hosted site you can test the entire map against staging in advance. If you are migrating a site with hundreds of ranking URLs, that difference is worth more than any design feature you are comparing.

The 180-day cliff nobody plans for

If your redesign also moves to a new domain, there is a clock running that most checklists describe incorrectly. Search Console has a Change of Address tool, and its job is to forward signals from the old site to the new one and tell Google to prefer the new site when picking canonical pages. It does that for a fixed period.

Google's help page for the tool puts the number at 180 days, repeatedly. Signals are forwarded for 180 days after you start the migration. You see move notifications for 180 days. You can cancel the request for 180 days. And then the sentence that should shape your redirect policy: after the 180 day period, Google does not recognize any relationship between the old and new sites, and treats the old site as an unrelated site, if still present and crawlable.

Set that against the instruction in the site move guide, which says to keep the redirects for as long as possible, generally at least 1 year. The two are not in conflict, they are describing different things. Google's assistance expires after six months. Your redirects are what carry the move after that, which is why the cheapest mistake to avoid is letting the old domain lapse a year later because nobody remembered why it was still being renewed.

Now the correction that matters most, because it applies to the majority of redesigns. The tool is only for moving between domains or subdomains, and Google lists the moves you should not use it for.

Read that second bullet closely. Moving some pages from one location to another within your site is explicitly not a Change of Address case, and Google's instruction for it is to just add redirects and update your sitemaps as appropriate. If you are redesigning on the same domain, which most businesses are, there is no tool, no forwarding and no notification. There are only your redirects, your sitemap and your own attention. Any guide telling you to submit a change of address for a same-domain redesign is telling you to do something the tool will not let you do.

The instructions that quietly go missing

Beyond URLs, every page carries machine-readable instructions that a rebuild either reproduces or does not. They are invisible in a design review, which is precisely why they get lost in one.

Two of them do the heavy lifting. The canonical tag tells search engines which URL is the real one when several could show the same content, and the robots meta tag says whether a page may be indexed and its links followed. They are common enough on the measured web that your current site very likely uses them, and a new build that omits or reassigns them is changing instructions you never knew you were giving.

The failure mode with real consequences is a single word left behind. A staging site is normally built with noindex on every page so that an unfinished site does not appear in search results, and that is correct practice. It becomes a catastrophe when it ships to production, because the new site politely asks Google to remove every page it has just launched. It is the most common serious redesign failure we see, it produces a total collapse rather than a dip, and it is invisible unless somebody looks at the source of a live page.

There is a quieter version too. The same HTTP Archive analysis found invalid HTML inside the head element on 10.1% of desktop pages, which can cause the head to be treated as ending early so that everything after it lands in the body. Tags in the body do not work. A rebuild that injects a stray element into the head can therefore disable the canonical tag sitting below it while every tag still appears correct in the source.

One small file can switch the whole site off

The other launch-day file is robots.txt, which governs crawling rather than indexing. It is a plain text file, usually tiny, and it can hide an entire website from search engines with one line.

Absence is not the danger here. A site with no robots.txt at all is simply not blocking anything. The danger is the file that exists and says the wrong thing, and the classic way that happens is a staging configuration containing a blanket disallow rule being copied to production along with everything else on launch day.

A detail that catches even careful teams: robots.txt and noindex are not interchangeable, and using both together can backfire. Blocking a page in robots.txt stops it being crawled, and a directive the crawler never sees is a directive it cannot obey. If you want a page out of the index, it needs to stay crawlable long enough for the noindex on it to be read. Blocking first and adding noindex second is a common sequence that leaves pages in the index with no way for Google to learn otherwise.

Deleting pages is the risk nobody budgets for

The technical work above protects addresses. It does nothing to protect what was on them, and a redesign is the single most likely moment for content to be quietly thrown away, because cutting pages makes a site look cleaner in a design review.

It tends to look like this. A site with dozens of pages comes back with a handful, on the reasonable-sounding argument that nobody reads the old ones. Sometimes that is true. Frequently the pages being cut are the ones answering specific questions that brought in the visitors who eventually called, and a redirect to the homepage does not preserve their value. It preserves the address while discarding the answer.

So run this in the opposite order to how most projects run it. Before the design work starts, pull the list of pages that received organic search visits in the last twelve months and mark them as keep. They can be redesigned, rewritten and improved. What they cannot be is deleted or reduced to a paragraph on a combined page without somebody deciding that on purpose, with the traffic number in front of them.

Google's line about any significant change bringing ranking fluctuations covers content as well as addresses. A page that keeps its URL but loses two thirds of what it said is a significant change, and it is one no redirect check will ever flag.

Do not take the site down while you build

One more decision gets made casually and deserves better. Businesses often want the old site replaced by a holding page while the new one is built, on the theory that an outdated site is worse than none.

Google disagrees, in writing. Its guidance on pausing a website says that completely closing a site even for just a few weeks can have negative consequences on Google's indexing of your site, and recommends limiting site functionality instead. If a genuine shutdown is unavoidable for a short period, the documented approach is a 503 response with a retry-after header, which says the site is temporarily unavailable rather than gone.

Build on staging, keep the current site serving, and switch when the new one is ready. A site that looks dated is still answering searches and still taking calls. A holding page is answering nothing while asking search engines to reconsider what your domain is for.

What a safe switchover actually looks like

Pulling the above into the order the work happens in, a competent launch looks like this.

1. Staging is unreachable to search engines, and production is not. Password protection is the safest method, because it cannot be forgotten in the way a noindex tag can. Whatever is used, the last task before launch is confirming production carries no leftover blocking.
2. Every old URL has a decision. Kept as is, redirected to the closest equivalent page, or deliberately retired. The output is a map, not a mental note, and the homepage is not an acceptable destination for a page about one specific service.
3. Redirects are permanent, server side, and tested. On a self-hosted site test the whole map on staging first. On a hosted platform, where redirects only fire once the old URL is gone, test the first batch within minutes of going live.
4. The new sitemap goes to Search Console on launch day. Google says to submit the new sitemap so it learns the new URLs, and that you can then remove the old one.
5. The head of a live page gets read by a human. The canonical tag and the robots meta tag in the head, then robots.txt at the root, all on the real site after launch. Three checks, five minutes, and they catch the failures that cost entire quarters.
6. One move at a time. Google's advice on site moves is not to chain them, so a redesign, a domain change and a platform migration on the same weekend is three overlapping recoveries with no way to tell which one hurt.

What to require from whoever builds it

You do not need to run any of this yourself, but you do need it to be somebody's named job. When a redesign loses traffic permanently, it is almost never because the work was impossible. It is because nobody was responsible for it and everybody assumed it was included.

Put four things in the agreement. First, that existing URLs are preserved by default and any change is flagged for approval. Second, that a redirect map covering every old URL is a deliverable you receive a copy of, in a format you can read. Third, that staging is blocked from search engines and production is verified clear on launch day. Fourth, that somebody checks Search Console for coverage errors in the weeks after launch, with an agreed date rather than good intentions.

Ask one question during the pitch and listen closely to the answer: what happens to our current URLs? A team that starts talking about the redirect map, and asks whether you have a list of your top organic pages, has done this before. A team that says search engines figure it out has told you who will be handling it, which is nobody.

What we do, and who should hire someone else

We treat URL preservation as the default and the redirect map as a deliverable, because a rebuild that loses the traffic it inherited is not a successful project regardless of how it looks. Our own redesigns are not exempt from any of the above. The measured outcome we publish is Cornerstone Healing Center, which saw 20% more website conversions after our redesign. We have two ways in, a flat fee for a one-time build and a monthly partnership when the work keeps running, both set out on our offer page. Two weeks is the usual gap between starting and going live on a focused build, though on a rebuild the redirect work is what decides whether that launch is safe. Any partnership request gets a reply within 48 hours.

Three cases argue for hiring someone else. If your site has thousands of URLs, years of accumulated redirects and an international structure, hire a technical SEO specialist to own the migration and a design team to work under that plan, because at that size the migration is the project and the design is a workstream inside it. If your current site gets no organic search traffic at all, none of this applies to you and you should spend the money on the design and the offer instead. And if the real goal is more traffic rather than a better website, a redesign is an odd way to buy it, so our piece on traffic that produces no leads is a more honest place to start.

Service businesses weighing a rebuild can read how we work with that market on our local services page, and if you are still choosing between building it yourself and hiring somebody, our comparison of platforms against designers covers that decision.

Freeze these five numbers before anyone touches the site

Finish with this, because it costs an hour and it is the only thing that will tell you afterwards whether the redesign helped, hurt, or did nothing. You cannot measure a change you did not record, and after launch the old version of your site stops being available to measure.

Record these while the current site is still live, and save them somewhere that is not the website.

1. The full list of URLs. Export your sitemap, or crawl the site, and keep the file. This becomes the checklist that every redirect is measured against, and no substitute for it exists after launch.
2. Your Search Console performance export. Queries, pages, clicks and impressions for the last twelve months, exported to a file. This is the before picture, and it is the evidence that settles any later argument about what changed.
3. Your top twenty landing pages from organic search. From analytics, with their visit numbers written down. These are the pages that are not allowed to quietly disappear in a content cull.
4. Whatever redirects already exist. Most sites that have been around a while are already redirecting old addresses. Those rules have to survive the rebuild, or you break moves that were made years ago and have been working ever since.
5. A copy of the current pages. The actual text, saved as files. When a page loses rankings after launch, the first useful question is what it used to say, and the second is what it says now.

Then put two dates in the calendar: one three days after launch to check indexing, redirects and the head of a few live pages, and one thirty days after launch to compare against the numbers you froze. If the first check is clean and the thirty day comparison is close, the dip you saw in week one was exactly what Google said it would be.

So this is built the other way around. Every figure below is a price a named company prints on its own page, or a count taken from a directory anyone can open. We sell websites, which makes us an interested party, and the section near the end says plainly who should buy from someone else. All prices were read on 12 August 2026 and are linked where they appear.

The 30-second answer

A working real estate website in 2026 costs between roughly $700 and $9,200 in its first twelve months of software, and where you land inside that has almost nothing to do with design. The listings feed itself is cheap: roughly two thirds of United States multiple listing services charge nothing to send data to your site, and the highest fee published in the directory we read is $30 a month. What moves the total is the platform subscription, which runs from $59 to $724.95 a month at published rates, plus a one-time setup fee of $0 or $500 depending on how you pay, plus advertising packages that cost more than everything else combined. Design and build sit on top of that as a separate purchase.

Four separate purchases hiding behind one question

The confusion in this market comes from bundling. Four different things get sold as "a real estate website", and their prices have nothing to do with each other.

• The listing data. Permission from your MLS to publish its listings, delivered as an IDX feed. Priced by the MLS, usually per website per month.
• The software that displays it. A plugin bolted onto a site you already have, or a hosted platform that is the site. This is the recurring bill people remember.
• The site itself. Design, copy, structure, speed, the pages that are not search results. Sold as a project, a template, or thrown in free with the subscription.
• Traffic. Advertising packages, lead guarantees, ad management fees. Not a website expense at all, but it usually arrives on the same invoice.

Price them separately and the decision gets easy. Price them as one number and you will overpay for the two that are cheap in order to get the two that are not.

The MLS data fee is the smallest line, not the biggest

Start with the number everybody gets wrong, because getting it right removes most of the fear from this decision. Cost guides quote a single blended "IDX service fee" and let you assume the multiple listing service is behind it. Luxury Presence, which sells real estate websites and says so, puts that recurring fee at between $50 and $200 per month in 2026, with a one-time IDX setup running from $200 to more than $2,000. Its overall verdict is that an IDX website in 2026 typically costs from $250 to $1,500 or more per month.

That blend hides three separately priced things, and the one people fear most is the cheapest of them. The MLS's own share is published. One IDX vendor, Showcase IDX, runs an open coverage directory with a page for every MLS it connects to, and each page prints the subscriber's share of that MLS's data fee. Reading all 343 active United States entries gives a clean picture.

Two hundred and twenty-four of the 343 charge nothing at all. Seventy charge $5 a month, thirty-seven charge $10, and only twelve charge more than that. The single highest figure in the set is the $30 a month that Northwest MLS in Washington charges, and it bills per team rather than per website. The median MLS in this directory charges zero.

Size does not predict the fee either, which is worth knowing before anyone tells you a large market is expensive. Bright MLS, listed with 98,910 agents across the Mid-Atlantic, charges nothing. Stellar MLS in Florida, listed with 64,893, charges nothing. California Regional MLS, the largest in the set at 120,142 agents, charges $10.

Two caveats keep this honest. This is one vendor's directory, and the fee it lists is what that vendor passes through to you, so a different provider may present the same MLS differently. Showcase's own pricing page puts the range at $0 to $33 a month and says that although rates vary slightly, 99% of IDX companies charge these fees. And some MLSs charge the agent directly for feed access, separately from anything a vendor collects, which is a question for your board rather than your web developer.

The practical instruction is short. Look up your own MLS before you budget anything. If you were working from a $50 to $200 a month figure, the MLS's part of it is very likely zero, and over three years that gap is larger than the cost of a decent website.

What the platforms charge to run the site

Here is where the money actually goes. The software that turns a data feed into a searchable website is sold by subscription, and the published range is wide enough that the same job can cost twelve times more depending on what else is bundled in.

At the bottom, Placester sells a hosted site from $59 a month for Agent Essential, $79 for Agent Plus and $129 for Agent Premier, with 20% off for annual billing, no contracts and a fourteen day trial. Extra IDX feeds are $25 a month each.

In the middle sits the plugin approach. Showcase IDX charges $94.95 a month for Essentials and $124.95 for Premium, with no setup fee, or $949.50 and $1,249.50 paid yearly, which is two months free. It is a WordPress plugin, so it assumes you have a site already. That assumption is the entire cost difference between this row and the next one.

At the top are the all-in-one platforms. Real Geeks charges $399 a month plus a one-time $500 setup fee, includes two users, and adds $10 per MLS board per month for feeds. Sierra Interactive publishes two columns: on an annual subscription its packages start at $299.95, $399.95 and $599.95 a month, and month to month the same three are $359.95, $474.95 and $724.95. One MLS feed is included and additional feeds are $25 per site each month.

Compare what you get, not the number. The $59 plan and the $724.95 plan both put listings on a website. The gap is a customer relationship manager, lead routing, automated messaging, seats for a team and an onboarding program. If you have a team of five and no CRM, the top of that chart is defensible. If you are one agent with a phone and a spreadsheet, $724.95 a month is $8,699.40 a year for features you first heard named during a demo.

Setup fees, contract terms and the price of leaving

The monthly figure is not the whole subscription. Three other numbers sit around it, and only one of them appears in most comparisons.

Setup fees depend on how you pay rather than what you get. Sierra charges $0 to set up an annual subscription and $500 to set up the same thing month to month. Real Geeks charges $500 once, on top of the $399. Showcase IDX and Placester both charge nothing, and Showcase says so in as many words: there are no setup fees or long term contracts.

Commitment length is the second. Real Geeks offers six month, twelve month and annual terms, and its lead guarantee is written against the full length of the term rather than any single month. Showcase IDX sells a monthly plan you can leave with thirty days notice, or a yearly plan at two months off, and if you leave the yearly plan early you get a prorated refund minus a $200 early cancellation fee. Placester says it has no contracts and includes a fourteen day trial.

The third number is the one nobody looks for. Real Geeks states that if it bought your domain for you, moving that domain to a website it does not host costs $50. Fifty dollars is nothing. Discovering the clause during a migration, when your listings are down and your email is routing through somewhere you no longer control, is not nothing. We wrote a whole piece on who owns your website because that clause has a hundred cousins.

Most of a big real estate bill is advertising

The subscriptions above are the small half of a big real estate bill. The large half is advertising, sold beside the software and priced per month so that it reads like part of the same product.

Real Geeks sells three lead packages: a social package running Facebook and Instagram listing ads at $299 a month, a search package running Google and Bing property search ads at $599, and a seller package running Google valuation ads at $599. Sierra takes a different shape, offering managed Google Ads to customers with a minimum ad spend of $500 a month and charging a 10% monthly management fee on top.

What makes Real Geeks unusually worth studying is that it publishes the guarantee behind those prices, which means you can do arithmetic instead of guessing.

Divide one by the other and the advertising line becomes legible. The social package at $299 a month guarantees twenty leads a month in a Tier 1 metro, which is $14.95 per guaranteed lead, and forty in a Tier 3 metro, which is $7.48. The search package at $599 guarantees the same counts, so $29.95 and $14.98. The seller package, also $599, guarantees ten a month in Tier 1 and fifteen in Tier 3, which works out at $59.90 and $39.93.

None of those numbers are bad. Seller leads are worth more than buyer leads and cost more to buy, and a company willing to put a floor under the count in writing is being more transparent than most of this market. The point is the comparison they make possible. If you would not pay $59.90 for a seller contact from any other source, the package is not cheaper because a website came with it. And if you would happily pay it, then what you are buying is advertising, and the website is the delivery mechanism, not the purchase.

One thing the pages do not settle: whether the media spend sits inside the package price or beside it. Sierra states its ad spend minimum separately and prices management on top, which is unambiguous. Ask any vendor to put the answer in writing before you sign, because the difference between the two readings is several thousand dollars a year.

The clause that charges you for editing your own site

Read the footnote under any lead guarantee before you sign the front of it. The Real Geeks guarantee, in the company's own words, is invalidated if the site owner makes changes to the code or content on the templated website it provides, and it names the home search pages, the home valuation pages and the lead forms specifically. It is also invalidated if you change your target market more than thirty days after launch.

There is nothing sinister in that. A company promising a lead count cannot promise it on a funnel somebody else has been editing. But look at what you have actually bought. You are renting a site you may not touch, in a market you may not change, for the length of a term, and if the leads fall short the remedy is a one month credit on a twelve month term or a half month credit on a six month one. That is a media buy with a website attached, and it should be evaluated as a media buy.

The alternative is not exotic. Own the site, run the ads separately, and keep the freedom to rewrite the page that is not converting on the afternoon you notice it is not converting.

What the build itself should cost

Nothing above pays for design. Platform templates are included with the subscription, which is exactly why they look like each other, and an agent site that reads like every other agent site in the county is not competing on anything except the ad budget behind it.

A build is a fixed-scope project: a home page that says who you serve, area or neighborhood pages that can rank, a valuation or inquiry path that works on a phone, the search experience wired into whichever feed you chose, and the technical work underneath. On a plugin route it is a real project. On a hosted platform it is mostly configuration, because the platform will not let you do much else, and paying an agency four figures to configure a template is a poor trade.

Ours is a flat fee for a one-time build, or a monthly partnership when a business wants design and development running continuously rather than once. Those live on our pricing page so nobody has to sit through a call to hear them. Whoever you hire, get the build quoted as a separate line from the software, because otherwise you cannot tell which of the two you are unhappy with a year later.

The same twelve months, priced four ways

Four routes, one year, using only the published figures above. Advertising is excluded from every row on purpose, because it is a separate decision and it would swamp the comparison. The build is excluded too, for the same reason.

The arithmetic is deliberately visible: $94.95 times twelve is $1,139.40, $399 times twelve plus the $500 setup is $5,288, and $599.95 times twelve is $7,199.40. Take the last row month to month instead of annually and it becomes $724.95 times twelve plus a $500 setup, or $9,199.40, which is the top of the range in the answer at the start. The distance between that and the $708 row is $8,491.40 a year.

Notice what the fourth column does not measure. The cheapest row assumes you already have, or will pay for, a website. The two most expensive rows include one. That is the honest tension in this table, and the next section prices the missing piece.

What the site has to earn before any of it makes sense

Now put those totals next to what the average buyer of them earns. The National Association of REALTORS, in its 2025 Member Profile released on 6 August 2025, reported a median gross income of $58,100 for 2024, up from $55,800 the year before. The typical member had twelve years of experience, closed ten transactions and did $2.5 million in sales volume. Eighty-seven percent are independent contractors, which means the technology bill is theirs, not their brokerage's.

The association's REALTOR Technology Survey, published 18 September 2025, found that 34% of respondents spent an average of $50 to $250 a month on technology for their individual business over the previous twelve months. That is a third of the profession, and the figure covers the whole technology budget, not the website line. A $724.95 platform sits about three times above the top of it before a dollar of advertising.

The same survey is worth reading for what it says about where business comes from. Social media was the top lead-generating technology at 39%, followed by customer relationship management at 23% and the local MLS at 17%. If your website is meant to be your lead engine, that survey is a reason to check the assumption rather than fund it.

The website you may already be paying for

Before buying anything, find out what your brokerage already provides. In the same technology survey, 38% of members agreed and a further 29% strongly agreed that their brokerage gives them all the technology tools they need to do their jobs. Two thirds is not a rounding error, and brokerage-provided sites usually arrive with the MLS relationship already handled.

The catch is the one this whole article circles. A brokerage site is the brokerage's, and it goes away when you do, along with the content, the rankings and often the leads. That trade may be worth making for a new agent with no book of business and no budget. It is a bad trade for someone twelve years in with a farm area they have spent a decade building.

Ask three things: does the brokerage site let you publish your own pages, does the domain belong to you, and can you take the content with you. If the answers are no, no and no, then what you have is a profile, not a website, and the honest way to think about a paid site is as the thing that survives a move.

The two lines that decide whether it is actually yours

Everything above changes value depending on where two things sit. The first is the domain. It should be registered in your name, in an account you can log into today, before anyone builds anything. The second is the content and the pages you write. A hosted platform holds both by default unless you have arranged otherwise, and platforms are bought and sold and repriced.

This is not a reason to avoid platforms. It is a reason to know which route you picked. A plugin on your own site means you keep the site if the plugin disappears. A hosted platform means the site is the subscription, and cancelling ends both. Neither is wrong. Choosing one without noticing is.

Buyers, meanwhile, are not choosing you off a search box. NAR's 2025 Profile of Home Buyers and Sellers found that 88% of buyers purchased through an agent or broker, and that agents were the most trusted and most used information source, ahead of online listings. Ninety-one percent of sellers used an agent, matching the record, and only 5% of homes sold without one. The website's job is to be found, to be credible on a phone, and to make contact easy. It is not the transaction.

What we do, and who should not hire us

We build the site and leave the feed to whichever vendor your MLS approves. That is the split we recommend to most agents: a fast, owned site with pages that can rank, a search experience running on a plugin costing under $100 a month, and no long term contract on the design. The one measured result we can point to is a 20% rise in website conversions at Cornerstone Healing Center following its redesign, and that came out of structure and speed rather than a bigger advertising budget.

Three situations point elsewhere. If you want a customer relationship manager, a dialer, drip campaigns and a website in one login with one bill, buy a platform, because stitching those together yourself is worse value than the subscription. If you are brand new, have no listings and no budget, use whatever your brokerage gives you for a year and spend the money on getting in front of people instead. And if what you actually want is leads next month, that is an advertising decision, and our comparison of in-house against agency costs will not help you with it.

Agents and teams who want the longer version of how we work with this market can read the real estate page. We answer partnership requests within 48 hours, and a focused build is usually live about two weeks after we start.

Choose your route in three questions

Three questions settle this faster than any comparison table, because each answer removes an entire branch. Answer them in order.

1. Do you need a customer relationship manager you will actually use? Not one you intend to use. One you already use, or one you are replacing because the current one is failing. If yes, you are buying a platform, and the website is included in that decision rather than driving it. If no, you have just eliminated the $299 to $724.95 tier and saved yourself several thousand dollars a year.
2. Will you publish anything other than listings? Neighborhood guides, market notes, seller resources, the pages that earn search traffic over years. If yes, you need a site you can edit and own, which means a plugin route, and the guarantee clauses attached to templated platform sites are a direct conflict with that plan. If no, a hosted template at $59 to $129 a month is honestly enough, and you should not let anyone talk you past it.
3. Is your advertising budget already working? If you are running ads that produce business today, adding a platform will not improve them, and the money is better spent widening what works. If you have no advertising at all, be clear that a website does not create demand on its own, and decide whether you are buying a place to send people or a machine for finding them. They are different purchases with different prices.

Whatever the three answers point at, look up your own MLS's data fee before you budget for it. Yours is probably zero, and now you know what to do with the difference.

We build websites for a living, so read this as an interested party setting out the rules it works under. Every legal point below links to the statute, the policy or the government page it came from, all read on 11 August 2026. None of it is legal advice. For a contract you will sign more than once, an hour with a lawyer in your state is money well spent.

The 30-second answer

Paying for a website does not transfer its copyright. In the United States, copyright in a design or in code belongs first to the person who created it under 17 U.S.C. 201(a), and it moves to you only through a written transfer signed by that owner under 17 U.S.C. 204(a). A "work made for hire" clause often fails to do it, because a website is not one of the nine kinds of commissioned work the statute allows. Ownership of the files is also a separate question from control of the domain, the hosting and the analytics, none of which copyright law touches at all.

What the law hands you by default, which is not much

By default, the person who did the work owns the work. Section 201(a) vests copyright in the author the moment something is fixed in a file, and the author is the human who made it, not the business that paid the bill. Section 201(b) creates the one big exception: for a work made for hire, the employer counts as the author from the start. That exception is doing enormous quiet work in the web industry, because almost every website is built by someone who is not your employee.

The Supreme Court settled the employee question in Community for Creative Non-Violence v. Reid, 490 U.S. 730 (1989). A charity commissioned a sculpture, directed the work closely, paid for it, and still lost the copyright fight, because the sculptor was an independent contractor rather than an employee under common law agency principles. Courts weigh a list of factors: the skill required, who supplies the tools, where the work happens, how the person is paid, whether taxes and benefits are handled like employment. A design studio working from its own machines, on its own hours, invoicing you as a vendor, lands on the contractor side of that line in almost every case.

So the starting position for a normal agency or freelance engagement is uncomfortable: they own it, you licensed it by implication, and the terms of that implied license are whatever a court later decides they were. Nobody wants to find that out during a dispute.

Why "work made for hire" in a web contract often does nothing

The phrase is in almost every template, and for websites it usually has no legal effect. 17 U.S.C. 101 allows only two routes to a work made for hire. The first is work by an employee inside the scope of employment. The second is a work "specially ordered or commissioned for use as a contribution to a collective work", plus eight other named categories, and only if both parties sign a written agreement saying so.

Here is that full list, from the statute: a contribution to a collective work, a part of a motion picture or other audiovisual work, a translation, a supplementary work, a compilation, an instructional text, a test, answer material for a test, and an atlas. Read it twice. A website is not there. Neither is a logo, a design system, an application, or a set of page templates. Some lawyers argue a site could be squeezed into "collective work" or "compilation" on particular facts, which tells you how thin the ground is.

When the category does not fit, the clause simply fails, and the copyright stays where section 201(a) put it. That is why competent agreements never rely on the phrase alone. They pair it with an assignment: this is a work made for hire, and to the extent it is not, the contractor assigns the copyright to the client. Two sentences instead of one, and the second one is the one that works.

California adds a twist worth knowing before you insist on the phrase. Under Labor Code 3351.5(c), a person engaged by a signed work made for hire agreement for a commissioned work of authorship is an employee for workers' compensation purposes, and Unemployment Insurance Code 686 makes the commissioning party an employer for unemployment insurance. California buyers who demand work made for hire language can therefore create a payroll relationship they never wanted. An assignment gets you the same ownership without that side effect.

The words that actually move ownership

The operative language is an assignment, in writing, signed. Section 204(a) requires an instrument of conveyance, or a note or memorandum of the transfer, in writing and signed by the owner of the rights being conveyed. A verbal promise, a line in a proposal deck, and a friendly email thread are all weaker than one signed sentence.

Four details separate a real assignment from a decorative one.

• Present tense. "Hereby assigns" transfers now. "Agrees to assign" is a promise to sign something later, which is only as good as the relationship on the day you ask.
• Named deliverables. Final files, layered source files, code, and content produced for you. If source files and working files are not named, expect to receive flattened exports.
• A payment trigger you can meet. Assignment on receipt of final payment is normal and fair. Assignment on some vaguer condition, or no trigger at all, is where disputes start.
• Honest carve-outs. Most studios reuse internal frameworks, component libraries and scripts across clients, and will not assign those. That is reasonable. What you need instead is a perpetual, irrevocable, transferable license to use and modify them as part of your site, so a future developer can work on it without calling anyone.

Two more clauses are worth reading closely. Portfolio rights, which let the studio show the work, are standard and harmless. A right for the studio to keep and reuse "all designs" without limit is not, because that is the clause that lets your competitor down the road get a site that looks like yours.

What that clause is worth when things go wrong

Ownership decides who holds the enforcement tools, and those tools have numbers attached. Under 17 U.S.C. 504(c), a copyright owner can elect statutory damages instead of proving lost profits: not less than $750 and not more than $30,000 for each work infringed, rising to as much as $150,000 where the infringement is proved willful, and dropping to as little as $200 where the infringer proves it had no reason to know.

Most disagreements never get near a courtroom. That is the point. The party who can credibly send a letter about statutory damages is the party who gets a fast, quiet resolution, and that party is whoever the paperwork says owns the work. If you are not the owner, none of the numbers in that chart are available to you, and the pressure runs the other way.

Registration costs less than one hour of anyone's time

Owning a copyright and being able to sue over it are two different things. 17 U.S.C. 411(a) blocks a civil action for infringement of a United States work until registration of the claim has been made. Section 412 goes further on money: statutory damages and attorney's fees are off the table for infringement that begins after publication and before registration, unless registration happens within three months of first publication. Three months from launch is a real deadline with real consequences, and it is the one nobody in the industry mentions.

The fees come from the Copyright Office fee schedule. Note the wording on the cheapest option: single author, same claimant, one work, not for hire. A genuine work made for hire does not qualify for it, which is a small, satisfying reminder that the phrase has consequences beyond the sentence it sits in.

Whether registering a whole website is worth it depends on how much of your brand lives in it and how distinctive the design is. For a bespoke site that took months, at $65 it is difficult to argue against. For a template build with stock components, there is much less to protect, and the honest answer is to skip it.

The domain is a separate argument, and it runs on a clock

Copyright law has nothing to say about your domain name. Domains are governed by contract and by ICANN policy, and the only thing that matters is whose name sits in the registrant field. The ICANN Transfer Policy, in the version published on 21 February 2024, is explicit that the registered name holder and the administrative contact are the only parties who can approve or deny a transfer to another registrar, and that the registered name holder wins any dispute between the two.

The same policy sets out when a registrar may refuse to let a domain move at all. A transfer can be denied within 60 days of the creation date, within 60 days of a previous transfer between registrars, and during the 60-day lock that follows a change of registrant. Those windows are the practical reason to sort domain ownership at the start of a project rather than during a divorce. If you change the registrant on the day you fire your agency, you can find the name frozen for two months, right when you need to point it somewhere new.

The fix is unglamorous. Register the domain yourself, in the business name, on a card the business controls, before anyone starts designing. If it is already registered to someone else, ask for the registrar account or a change of registrant now, while everyone is friendly, and put the renewal on a calendar you own. In real estate, some platforms register the domain for you and charge to move it afterwards, which is one of the line items in our breakdown of what a real estate website costs.

The accounts that never make it into the contract

A site is not one asset. It is a stack of accounts, and each one has a different owner by default. This is the part clients discover in the worst possible week, so here it is in advance.

The pattern behind that table is simple. Anything with a login and a credit card attached is controlled by whoever set it up, and no clause about copyright changes that. Ownership of the files and control of the accounts have to be handled as two separate lists.

The parts of your site you will never own outright

Some of what makes up a website is licensed, not owned, and no contract can change that. Being clear about which parts keeps everyone honest.

If your site runs on WordPress, the software is under the GPLv2 or later, and WordPress.org states that derivative works such as plugins and themes inherit that license. You are not buying WordPress. You are receiving it under a license that also gives you the freedom to modify it and move it, which is exactly the freedom you want when a relationship ends.

Commercial themes come with their own terms, and they are usually better than people expect. The ThemeForest Regular License permits one end product for yourself or for one client, says that the developer can transfer that single end product to the client for any fee and that the license transfers with it, and forbids selling the end product to anyone other than that one client. In plain terms: one purchase per site, and the license is meant to land with you. If a studio built five client sites off one purchase, that is a licensing problem in your website, so ask which theme was used and ask to see the purchase code.

Fonts and photography are where the real exposure sits, because both are usually licensed to whoever bought them, for a defined use, and the receipt lives in the vendor's account. Rather than trust a summary of anyone's terms, ask for the licenses themselves: which typefaces, bought where, licensed to which entity, for what page-view tier, and the same for every photograph that is not yours. If nobody can produce them, that is your answer, and replacing a font is far cheaper than the alternative.

What the law now asks of you, the buyer

If you hire a solo freelancer rather than a firm, the paperwork is no longer optional in a growing number of places, and the duty falls on you. California's Freelance Worker Protection Act, enacted as SB 988 and now Business and Professions Code 18100 and following, applies to contracts entered into or renewed on or after 1 January 2025 worth $250 or more, counted alone or aggregated across the previous 120 days.

The obligations are concrete. The hiring party has to put the contract in writing, furnish a signed copy to the freelancer, keep it for at least four years, and include the parties' names and addresses, an itemized list of services with their value and the rate and method of pay, the date payment is due or how it will be determined, and any deadline for submitting an invoice. Payment is due on the contract date, or within 30 days of completion if the contract is silent. A worker who asked for a written contract and was refused is awarded an extra $1,000, and late payment can cost up to twice the unpaid amount.

Section 18102(b)(2) is the clause that ties this section to the rest of the article: once work has started, a hiring party cannot require, as a condition of paying on time, that the freelancer grant more intellectual property rights than the contract agreed. Ownership has to be negotiated up front, not extracted at invoice time.

Other places have their own floors, all lower than most project budgets.

New York's Freelance Isn't Free Act added Article 44-A to the General Business Law on 28 August 2024, covering freelance workers hired for $800 or more alone or across the preceding 120 days, with payment due within 30 days when the contract does not say. The Illinois Freelance Worker Protection Act took effect on 1 July 2024 at $500 across 120 days, with the same 30-day default, as summarized by Jackson Lewis. The city of Los Angeles has run its own ordinance since 1 July 2023 at $600, and its Office of Wage Standards now says it will only take complaints on contracts that state law does not already cover.

One limit matters for anyone about to panic. California and New York both define a freelance worker as a person, or an organization of no more than one person. Hiring a twelve-person studio does not trigger these laws. Hiring the designer who works alone does, and that is a very common way to buy a website.

Eight questions to ask before you sign

Print these. Any studio worth hiring answers all eight without checking with anyone, and the answers tell you more about how a company operates than its portfolio does.

1. Does the contract assign the copyright to me, in present tense, and when does that happen? Look for "hereby assigns" and a clear trigger such as final payment. A promise to assign later is not the same thing.
2. What exactly is assigned, and what is only licensed? Expect internal frameworks and libraries to stay theirs. Make sure the license to keep using them is perpetual and survives the relationship.
3. Do I get source files and the repository, including history? Layered design files and the code as it is actually built, not exports.
4. Whose name goes in the domain registrant field? The answer should be yours, from day one.
5. Which accounts will be in my name and billed to me directly? Hosting, analytics, Search Console, business profile. Ask for the list before the project starts.
6. Which fonts, images, plugins and themes are licensed, to whom, and what do they cost each year? Ask for a written inventory at handover.
7. What happens on the day we part ways? A wind-down clause with a fixed handover window and a defined list of what gets transferred beats goodwill.
8. Who can I call in six months, and what does that cost? Ownership is worthless if nobody can use it. Prices for ongoing help should be written down before you need them.

What a complete handover contains

A handover is a delivery, not an email saying thanks. When we close out a build, the client receives a single document listing the domain registrar and account, the hosting account and where it is billed, admin credentials for the CMS at owner level, the repository with its history, the design source files, an inventory of every third-party license with renewal dates and costs, the analytics and Search Console properties under their own Google account, and the signed assignment. It takes an hour to assemble if the project was set up correctly and two miserable days if it was not.

Ask for that list in writing at the proposal stage. A studio that hesitates is telling you something useful about how the last engagement ended.

What we do, and when we are the wrong choice

Our position is simple: you paid for it, so you take it with you. Our agreements assign the copyright in the deliverables on final payment, our clients hold their own domains and hosting from day one, and the handover document above ships with every project. Our offer page sets out what each includes: a focused one-time build, and a monthly partnership for continuing design and development work.

We are the wrong choice in three situations. If you want a site you will never touch again and never move, a template on a builder platform will do the job for less, and our comparison of builders against hiring a designer lays out that tradeoff honestly. If you need a lawyer to draft or review a bespoke agreement, hire the lawyer, not us. And if the person you actually need is a full-time employee sitting in your standups, our breakdown of in-house against agency costs will save you a discovery call. Businesses in local service industries often get more from a small, well-owned site than from a large one they cannot edit.

A 20-minute audit of the site you already have

You do not need a contract review to find out where you stand. Six checks, done in one sitting, will tell you which assets you actually control. Do them in this order, because the early ones are the expensive ones.

1. Look up your own domain. Use ICANN's public lookup service. Registrant details are usually redacted, so what you are really testing is the next step.
2. Log in to the registrar yourself. Not a shared screen, not a screenshot from your agency. If you cannot get in with your own credentials, you do not control your domain today.
3. Log in to the hosting or platform account and find a billing receipt with your business on it. No receipt usually means the account is somebody else's.
4. Open your analytics property and check your own access level. Administrator or nothing. The same goes for Search Console and your business profile, where you are looking for owner rather than manager.
5. Ask your developer for read access to the repository. The speed of the reply is data.
6. Search your email for the contract and read one sentence. The one containing the words assigns, assignment or ownership. If it does not exist, or it only says work made for hire, you have found the gap.

Failing a check is not a crisis and rarely means bad faith. Most of it is drift: an agency registered the domain in 2019 to get a project moving, and nobody revisited it. The fix is an email asking for the assignment to be signed and the accounts to be transferred, sent while the relationship is still good. That email is much harder to write after the relationship is not.

We build websites, so read this as an interested party showing its arithmetic. Every figure below comes from a published study, a payment company's own fee schedule or a platform's own pricing page, all read on 10 August 2026 and linked where you can check them. Where a source says its own numbers cannot be used a certain way, this piece says so instead of quietly using them anyway.

The 30-second answer

Mobile is where nonprofit websites lose money. Across 180 nonprofits in the 2026 M+R Benchmarks study, phones and tablets brought 52% of website visits but produced 43% of donation transactions and only 28% of online revenue. The average gift was $168 on desktop and $88 on mobile. On the main donation page, 11% of desktop visitors completed a gift against 8% of mobile visitors, and just 4% of mobile visitors at small nonprofits. Across all traffic, 1.6% of visitors donated, worth $1.33 per visitor. Fix the phone path before you rewrite the copy.

The study everyone quotes, and the sentence everyone skips

One study anchors most of what gets written about online giving, and it is worth knowing what it is. M+R Benchmarks 2026 collected data from 180 participating nonprofits, who self-identified their sector and were sorted by size using 2025 online revenue: Small below $1,000,000, Medium from $1,000,000 to $5,000,000, Large from $5,000,000 to $10,000,000, and Extra Large above $10,000,000. The figures published as averages are medians, chosen so a couple of spectacular programs cannot drag the number around.

Then there is the instruction almost nobody repeats. M+R writes: "Do not compare this year's M+R Benchmarks findings to previous editions!" The participant pool changes every year, the size bands were redrawn, and some metrics are now calculated differently, including page speed, which moved to the speed index reported by Google PageSpeed Insights. So when an article tells you that donation page conversion fell from last year's published figure to this year's, it is doing the one thing the people who gathered the data asked readers not to do. Note the difference between that and the year-over-year changes inside the report, which are calculated from the historical data of this year's participants and are safe to quote.

Used properly, the study answers a narrower and more useful question: what does normal look like right now, for organizations roughly your size, in roughly your sector. That is enough to tell you whether your own numbers are a problem or a preference.

Half the traffic, a bit over a quarter of the money

The device gap is the largest single distortion in the data. M+R reports that mobile users, meaning phones and tablets together, made up 52% of all visits to nonprofit websites in 2025, with desktop at 48%. Desktop then produced 57% of donation transactions and 72% of revenue. The average gift followed the same shape: $168 on desktop, $88 on mobile.

Be careful about what this does and does not prove. It does not show that mobile visitors are worth less as people. Two very different mechanisms produce the same chart. Some donors browse on a phone, think about it, and give later at a desk, which loads the desktop column with gifts the phone actually earned. And some donors genuinely give up on a small screen because a form is hostile on a small screen. The study measures the gap. It does not attribute it, and neither should you until you have looked at your own assisted conversions.

What the gap does settle is a question of priority. If the phone version of your site is the one that only has to avoid looking broken, you are treating the majority of your audience as an afterthought, and the revenue column is what that costs.

What the donation page itself converts at

Zoom in to the page where the decision happens and the pattern holds. Of visitors who reached the main donation page, 11% of desktop users completed a gift, against 8% of mobile users. At small nonprofits, those with online revenue under $1,000,000, mobile conversion was 4%.

The definition matters here more than the number. M+R calculates this as donations to the main donation page divided by pageviews of that page, using the one-time page where a separate monthly page exists. So it is not the share of your website visitors who give, and it is not comparable to whatever your ads platform calls a conversion rate. It is a measure of how well one page finishes what the rest of the site started.

That narrowness is what makes it usable. Ten minutes in your analytics will give you the same two figures for your own page, and unlike a revenue total they are not dominated by whether you had a good campaign in November.

The whole-site number: 1.6% of visitors, $1.33 each

Across all traffic sources, 1.6% of nonprofit website visitors made a donation, and nonprofits received an average of $1.33 per visitor. Sector matters enormously: visitors to Hunger and Poverty organizations donated at 3.3% and were worth $5.58 each.

Revenue per visitor is the most underused number on this list, because it is the only one that puts a price on a website project. Suppose your site takes 300,000 visits a year and earns the median $1.33 per visitor. Moving that to $1.60 is 27 cents a visitor, which is $81,000 a year. That arithmetic is ours, applied to somebody else's median, and it is an illustration rather than a forecast. Run it on your own two numbers and the argument for or against spending money on the site usually settles itself in about a minute.

What to change on the phone path, in order

Start with the thing that removes typing. Card numbers are miserable to enter on a phone, and digital wallets skip that step entirely. M+R reports how widely nonprofits offer them on the main donation page: PayPal at 79%, Google Pay at 58%, Apple Pay at 57% and Venmo at 44%. Note what that measures. It is adoption, not lift. The study does not claim wallets raise conversion, and this article will not either. The useful signal is scarcity: a donation page with no wallet button now sits in a shrinking minority, on precisely the device where its conversion rate is worst.

Second, measure speed properly. M+R now uses the speed index from Google PageSpeed Insights, which blends field data from real visitors with lab testing rather than relying on a simulation. You can run the same tool on your own donation page in under a minute, and the field data section is the part to read, because it reflects the phones and connections your donors actually have.

Third, count the work you are asking for. Fields on the form, taps to reach a $50 gift, whether an account is required, whether the amount buttons are big enough for a thumb, and whether the donation page shares a template with the rest of the site or arrives from a different system looking like a different organization. None of these is measured as a lift in the study, so treat them as hypotheses to test rather than promises. The point of the benchmark is that at 8% and 4% there is plenty of room to test into.

The road to the page got narrower in 2025

The donation page cannot convert people who never arrive, and the largest free road to it shrank. Organic search still accounted for 39% of all visits to nonprofit websites, but the share fell month by month across 2025. M+R attributes the decline to zero-click results and to people asking chatbots instead of searching, and argues that nonprofits now need an answer engine strategy alongside the search strategy they already have. We wrote about how to get recommended by ChatGPT and Google AI from the same starting point, and the practical work turns out to be mostly the same work: publish the facts about your organization in plain, quotable form on pages a machine can read.

There is also a paid road that costs nothing, and it is specific to this sector. Google Ad Grants gives each qualifying nonprofit up to $10,000 a month in search ads shown on Google.com. It has eligibility rules and account requirements, it buys clicks rather than affection, and plenty of grants sit half spent because nobody owns them. It is still the cheapest traffic any organization in this article can buy, and it lands on the same donation page, which is a reason to fix the page first.

Email is the other road, and it had a better year

While search narrowed, email grew. Email revenue rose 16% on average in 2025, and 11% of all online revenue was sourced directly to email. Nonprofits raised $54 for every 1,000 fundraising messages sent, up 4% on 2024, and $2.40 per subscriber across the year against $1.87 the year before. The average subscriber received 50 messages, 31 of them fundraising.

The website's role in that is unglamorous and easy to skip: it is the place list growth happens. A sign-up path that only exists in the footer is a decision to grow the channel that is working more slowly than you could.

The page has to survive the second gift

One donation is not the unit of value, and the retention numbers explain why the monthly option deserves the best position on your form. Among donors whose first online gift came in 2024, 24% gave again in 2025. Among prior donors, those who gave in 2024 and in at least one of the five previous years, 66% came back. Overall one-time retention was 48%. Monthly donors behave differently again: 10% stop within two months of setting up a gift, 81% are still giving after seven months, and 71% are still active a full year later.

Monthly giving supplied 27% of all online revenue in 2025, and the share scales with size: 22% at small nonprofits and 37% at the largest. Revenue from one-time gifts grew faster last year, 17% against 12%, which M+R reads as donors responding to a run of emergencies. That is a break from the long-run pattern, not a reason to stop building the sustainer program.

Two design consequences follow. The monthly option should be a real choice on the page rather than a checkbox under the amount buttons. And the place where a sustainer updates a card should be findable in ten seconds, because a stolen card is a common way for monthly gifts to end and it has nothing to do with how the donor feels about you.

Thirty-seven percent of the year arrives in December

Timing decides when a website project is safe. Nonprofits received 37% of all 2025 online revenue in December alone. The last week of the year accounted for 10% of annual revenue, and the last day accounted for 4%.

Read that as a project calendar. A donation page that goes live in November is a large bet placed on the biggest month of the year with no time left to test, fix or roll back. If a rebuild cannot be finished, measured and adjusted before the end of September, the honest scheduling answer is usually January, when a mistake costs a fraction as much and you have eleven months to learn from it.

The wider context is worth naming, because it is why so many of these organizations are looking at their websites at all. Online revenue for the average nonprofit rose 15% in 2025, and among study participants who received government funding, two-thirds reported receiving less of it than the year before. A lot of the growth in these charts is individual donors covering for something that disappeared.

What the page costs to run

Processing is the smallest decision on this page, and it gets the most attention. Here is what the companies publish about themselves.

Sources for the table: PayPal's merchant fees page, Stripe's pricing page and Donorbox's pricing page, all read on 10 August 2026. PayPal adds 1.50% for international charity transactions on top of the domestic rate. Donorbox presents its numbers as bands on the plan cards without splitting platform fee from processing fee, so use your own invoice as the authority rather than this table.

Now the arithmetic that decides the plan question. Moving from the Donorbox free plan to the $150 a month Pro plan saves roughly 1.5 percentage points at the middle of each band, so the plan pays for itself at about $10,000 a month in online donations, which is $150 divided by 0.015. Below that, the free plan wins. That calculation is ours, it uses the middle of two published bands, and your own contract may differ.

The wider point is the spread. Across the routes above, the fee on a $100 gift moves by roughly two dollars, before any monthly plan fee. Between an 8% mobile conversion rate and an 11% one, the same page earns nearly 40% more from the same visitors. The money is in the page, not in the processing fee, and the processing fee is where most of the debate goes.

Who actually does this work

The staffing data is the part of the study that made us wince in recognition. The average digital team at a participating nonprofit is six people, with medians running from five at small organizations to thirteen at the largest. Most teams either stayed the same size or shrank in 2025, in a year when revenue grew and the number of channels did not.

Outside help is already normal rather than exotic: 75% of participating nonprofits worked with partner agencies on digital work, including 62% of small organizations and every single one of the largest. So for most readers this is not a decision about whether to bring in help. It is a decision about what to point it at, and the data above suggests the answer is the phone version of one page rather than another channel.

What we charge, and who should not call us

We publish prices, in the same spirit as everything above. One flat fee covers a focused build. For teams that want the work to keep going, with someone accountable for the numbers every month, there is a monthly partnership. Both figures sit on our pricing page. Two hundred projects have gone out the door since we started. Eden Digital is a site we built that went on to grow revenue tenfold, and Cornerstone Healing Center came to us for a redesign, after which website conversions rose by 20%. Two weeks is the usual runway on a focused build, and nothing waits more than 48 hours for a reply.

Plenty of organizations should not call us. If your giving already runs through a platform your team likes and your mobile conversion is at or above the benchmark, the page is not your constraint and the money belongs in getting more of the right people to it. If your traffic is healthy and your form is fine and gifts still are not arriving, the problem is further upstream and the reasons traffic does not become inquiries is the more useful read. We are a fit when the site itself is the weak link, nobody currently owns the giving path end to end, and there is enough traffic that a few points of conversion adds up to real money. More on how we work with organizations in this sector sits on our nonprofits page.

A one-hour test you can run this week

None of this needs a consultant to diagnose. Take your own phone off the office Wi-Fi and go through your giving path as a stranger would.

Search for your cause, not your name. Type the phrase somebody would use if they did not know you existed. Note whether an AI answer sits above every real result, and whether it mentions you.

Open the donation page on mobile data and time it. Then run the same URL through Google PageSpeed Insights and read the field data, which is what real visitors experienced, rather than the lab score.

Count the work. How many fields, how many taps to give $50, and whether you are forced to create an account before you can give money away.

Look for a wallet button. If there is no Apple Pay, Google Pay or PayPal option on the page, you have found something to fix without needing a single further measurement.

Give yourself $5 and read the receipt. It is the most-read page in your entire program and usually the least designed. Ask whether it gives a first-time donor any reason to come back.

Then write two numbers on the wall. Revenue per visitor for the last twelve months, and mobile conversion on the donation page. Everything in this article is a way of moving one of those two, and neither of them moves quietly.

We build websites and ordering pages, so read this as an interested party showing its arithmetic. Every rate below came off the seller's own pricing page, a regulator's own page or a company's own filing, read on 9 August 2026. Where a figure could only be verified from an older announcement, the date is in the sentence. Where the honest answer is that the apps are worth paying, the article says so.

The 30-second answer

Marketplace delivery costs 15% to 30% of the order. Uber Eats publishes 20% on Lite, 25% on Plus and 30% on Premium, with 7% on pickup when your in-app prices match your in-store prices and 10% when they do not. DoorDash introduced 15%, 25% and 30% tiers with 6% pickup in April 2021. Taking the same order on your own site costs payment processing: 2.5% plus $0.29 through Uber Eats Webshop, or 2.9% plus $0.30 through Stripe. On a $40 order that is $12.00 against $1.29. The marketplace fee is the price of a customer you did not have. Paying it again on their tenth order is the part worth fixing.

What the apps charge in 2026

Start with the numbers each company puts on its own page. Uber Eats publishes three marketplace packages: Lite at a 20% Marketplace Fee, Plus at 25% and Premium at 30%, with Plus and Premium opening at a 0% intro rate for 30 days. Every package carries a 7% Pickup Fee, and Uber is specific about the condition: that rate depends on proof that pricing for in-app pickup matches in-store pricing, and without it the pickup fee is 10%. Running your own drivers through the app costs a 15% self-delivery fee. Borrowing Uber's couriers for orders that came from your own channels is Uber Direct, which starts at $7.99 per delivery. Selling through your own site with Uber's Webshop costs a 2.5% order processing fee plus $0.29 per order.

DoorDash sets up the same way. When it moved to tiered pricing, Restaurant Dive reported on 27 April 2021 that Basic charges 15%, Plus 25% and Premier 30%, that pickup is 6% across all three, and that Premier carries a growth guarantee reimbursing commissions for restaurants receiving fewer than 20 orders a month. The same report noted that the Storefront product for ordering on a restaurant's own site was free apart from payment processing fees. Those are the rates the industry has quoted ever since, and they are five years old, so treat your own contract as the authority.

Two details hide inside the percentages. The fee applies to the order value, not to your profit on it, and the consumer pays their own service and delivery fees on top, which raises the total your customer sees while leaving your side of the transaction unchanged. So the menu price a delivery customer pays is rarely the menu price you set, and the app is the one setting the difference.

The increase that landed in March 2026

Rates on a channel you do not own can move by email. Uber told merchants that new US marketplace fees took effect on 11 March 2026: Lite rose to 20%, Plus stayed at 25% except on Uber One member orders, which now carry 30%, Premium stayed at 30%, and pickup went from 6% to 7%. Custom negotiated delivery rates rose by 3 percentage points, capped at 30%. Restaurant Dive covered the change on 10 March 2026, the day before it started.

Read the Lite line again, because a move from 15% to 20% is a third more fee on every order that package touches. Restaurants that wanted out had until 10 March to say so, and a submission for the alternative had to be in by 11:59pm on 6 March. That window is the real lesson here. A pricing change on a channel you rent gets announced, not negotiated, and the reply-by date is measured in days.

What a 15% cap means in New York now

New York City is where owners think they are protected, and the protection changed in 2025. The Department of Consumer and Worker Protection now lists four separate caps on what an app may charge a restaurant per online order: delivery fees capped at 15%, transaction fees at 3%, basic service fees at 5%, and enhanced service fees at 20%. Add them up and the ceiling in the city famous for its 15% cap is 43%.

The history is short. The City Council voted on 26 August 2021 to make the caps permanent, holding apps to 15% for delivery and 5% for other fees. Then, as Restaurant Business reported on 1 May 2025, the Council passed Int 762-B, which kept 15% for delivery, 3% for credit card processing and 5% for other services, and allowed an additional 20% for enhanced services on top. The bill also lets restaurants charge higher prices on the apps than in the dining room, and put marketing materials in delivery bags.

Two things follow for anyone operating under a cap. The basic package still has to exist, so the 23% version of the deal is always available to you, and buying enhanced visibility is a choice you make rather than a rate you are handed. And the permission to insert your own marketing into a delivery bag is worth more than it sounds, because it is the one moment when a customer the app owns is holding something you printed.

What the platforms actually keep

The commission is not the platform's profit, and knowing that changes how you negotiate. In its quarterly report filed on 5 August 2026, DoorDash reported Net Revenue Margin, its own measure of revenue as a percentage of marketplace order value, at 13.5% for the quarter ended 30 June 2026, on $33.1 billion of Marketplace GOV and 970 million orders. Uber, filing the same day, reported Delivery revenue of $5,245 million on Delivery Gross Bookings of $27,463 million, which is 19.1% of order value.

Those bars are not strictly comparable, and the caption says why: both platforms count consumer fees, advertising and subscriptions, both are net of courier pay, and both cover grocery and retail as well as restaurants. Read them as scale, not as a like-for-like margin. The useful conclusion is that a 30% fee funds a demand machine with couriers, discounts and marketing inside it. You are not being overcharged by a company pocketing 30 cents on the dollar. You are buying access to something expensive to run, at a price set by whoever runs it.

What the fee actually buys

Demand, and it is real. DoorDash's annual report for 2025 counts more than 56 million monthly active users in December 2025 and more than 35 million members across DashPass, Wolt+ and Deliveroo Plus at the end of the year. Marketplace order value grew from $66.8 billion in 2023 to $80.2 billion in 2024 to $102.0 billion in 2025. No independent restaurant is going to reproduce that on its own, and pretending otherwise is how bad advice gets written.

What the fee does not buy is the customer. The app holds the account, the payment method, the notification permission and the reason to open something at 7pm on a Friday. That is the asset, and it is being rented back to you at 15% to 30% every time the same person orders again. The platforms are candid about it in their own filings. DoorDash's annual report lists among its competitors "merchants that have their own online ordering platforms".

What a direct ordering page costs

Payment processing, and not much else. Uber's own Webshop, which puts ordering on your site while your listing stays in the marketplace, is 2.5% plus $0.29 an order. Stripe publishes 2.9% plus $0.30 per successful domestic card transaction with no setup fee and no monthly fee. DoorDash's Storefront was free apart from processing when it launched. On a $40 order, all three land between $1.29 and $1.46, against $8.00 on a 20% marketplace fee.

Then be honest about the rest of the bill, because a free ordering page is not a free channel. Somebody has to build the site and keep it working, which is a real recurring cost we broke into four separate bills in a previous piece. If you want delivery on your own orders you still need couriers, and Uber Direct starts at $7.99 a delivery, which on a $40 order is close to the 20% marketplace fee anyway. Direct ordering wins clearly on pickup and on delivery you drive yourself. On courier-fulfilled delivery it wins by a smaller margin than the chart suggests.

The other real cost is demand. Traffic that used to belong to the app has to come from somewhere, which usually means your search listing, your signage, your receipts and your regulars. That is slower than switching on a marketplace tier, and it compounds instead of repricing itself in March.

Why a percentage hurts more than it reads

A commission is charged on revenue and paid out of margin, and margin is thin right now. The National Restaurant Association's 2026 State of the Restaurant Industry, published 12 February 2026, projects $1.55 trillion in sales and 15.8 million jobs, and reports that 42 percent of operators said their restaurant was not profitable last year. Sixty percent reported softer customer traffic, and more than nine in ten named food, labor, insurance, energy and swipe fees as significant challenges.

We are not going to invent a margin figure for your restaurant, because the real one depends on your rent, your labor model and your menu. Do the arithmetic on your own numbers instead. Take last month's marketplace sales, apply your tier, and put the result next to your rent. Owners who have never done this are routinely surprised to find the apps are one of their largest single suppliers, sitting somewhere between the meat order and the lease.

Why leaving the apps is usually the wrong move

The advice to delete your listing is usually written by somebody selling ordering software. Marketplaces are a customer acquisition channel with the largest audience in the category, and turning one off removes discovery, not just fees. The sane version of this decision splits the two jobs: let the app do acquisition, and stop paying acquisition rates for retention.

In practice that means the first order from a new customer is worth 20% to 30%. The tenth from the same person is not, and every one of those you move to your own page is close to pure margin recovered. The measure worth tracking is not marketplace sales, and not direct sales, but the share of repeat customers who order direct. If that number is flat year over year, nothing you have built is working, no matter how good the site looks.

A sensible order to build in

Do these in sequence, because each one makes the next cheaper. This is roughly how we sequence the work for restaurants and hospitality clients.

One, put real ordering on your own site. Not a PDF menu and a phone number. An ordering page that works on a phone in under three taps, with a card field and a pickup time. This is a build, not a plugin decision, and it is the only piece that has to be right.

Two, decide your pricing position. Menu parity is simplest and honest. Marking up delivery prices is legal in New York under the 2025 bill and is common practice, but note that Uber's 7% pickup rate is conditioned on in-app pickup prices matching in-store prices, so a blanket markup can cost you the cheaper pickup tier.

Three, work the bag. Every delivery order leaves your kitchen with a paper bag stapled shut, and in New York the 2025 bill explicitly allows your marketing inside it. A card with a direct-order code is the cheapest acquisition you will ever run, because the app already paid for the introduction.

Four, win pickup first. Pickup is where direct ordering wins outright: no courier, no marketplace fee, and 7% to 10% saved on every order you convert. It is also the easiest habit to shift, because the customer is already driving to you.

Five, rent couriers instead of a channel. On-demand delivery for orders that came through your own site is a per-delivery cost, which behaves very differently from a percentage as your average order rises.

Six, look at the split every month. Marketplace orders, direct orders, and repeat customers on each. Fifteen minutes with those three numbers will tell you more than any dashboard the apps give you.

What we charge, and who this is not for

Our prices are on the pricing page for the same reason the rates above are public. A focused build is paid once. Continuing work, where the site keeps changing and somebody owns the results, runs monthly. The count is past 200 projects shipped, among them Eden Digital, a site we built that went on to grow revenue tenfold, and a redesign for Cornerstone Healing Center that lifted website conversions by 20%. A focused build usually takes about two weeks, and requests get an answer inside 48 hours.

This is not for every restaurant. If you do 15 delivery orders a week, the fee is not your problem and a build will not pay for itself soon enough to matter. If your ordering already runs through a point of sale you like, the honest advice is to switch on its direct channel and spend the money on the reasons your existing traffic is not converting instead. We are worth calling when delivery is a real share of revenue, the site is the weak link, and nobody currently owns the job of moving repeat orders off the marketplace.

Before you sign anything, ask about these

Five short questions, useful against an app, an ordering vendor or an agency including this one.

What is my all-in rate per order, including processing? Get one number for a typical order, not a tier name. Percentages and per-order fixed fees behave differently at $18 and at $80.

How and when can that rate change? Ask for the notice period in writing. March 2026 was a few days of warning for a change of a third on one Uber tier.

Who owns the customer data? Names, emails, order history and the right to market to them. If the answer is complicated, the answer is that you do not own it.

What does the fee include that I would otherwise buy? Couriers, insurance, support, placement, promotions. Some of it is genuinely expensive to replace.

What happens to my orders if I stop paying? On a marketplace, everything stops. On your own domain and your own processor, you keep the page, the list and the history. That difference is the reason to build one, and it does not show up in either price.

We sell one of these, so read it as an interested party showing its arithmetic. Every price below was read off the seller's own pricing page on 8 August 2026. The vulnerability figures come from Patchstack's published 2026 security report. The platform and PHP numbers come from the WordPress.org statistics API, which counts live installs rather than surveying opinions. Where the honest answer is that you should spend nothing, the article says so.

The 30-second answer

Website maintenance costs roughly $30 to $400 a month at published 2026 rates, and the range is that wide because one phrase covers four separate purchases. Managed WordPress hosting is $25 a month at Flywheel and $30 at WP Engine on annual billing, and already includes platform updates and daily backups. The monitoring software on top is close to free, $1 to $2 per site per month at ManageWP. A care plan with people behind it is $39 a month per site at GoWP or $99 at WP Maintainer, which then bills hands-on work at $99 an hour. Design and conversion work is not maintenance at all. As for why any treadmill exists, 11,334 new vulnerabilities were reported in the WordPress ecosystem during 2025, 91% of them in plugins, and for the heavily attacked ones the median time to first exploitation is five hours.

The four bills inside one invoice

Split the quote before you compare it to anything else. Maintenance is four purchases wearing one label, and they are not equally worth paying for.

The server and the platform. Hosting, certificates, PHP and database versions, the core updates a host applies whether or not anyone is watching. Cheapest of the four, and the easiest to buy well.

The update treadmill. Watching for plugin and theme vulnerabilities, applying fixes without breaking anything, and having a plan for the days no fix exists yet. This is the part people picture when they hear the word maintenance.

Somebody being responsible. A named human who notices the site is down, restores the backup and replies on a Sunday. No software supplies this, which is why the price jumps the moment it appears in a plan. Trades where the phone is the business feel this hardest: our breakdown of what an HVAC company website costs prices the same responsibility inside a contractor marketing plan.

The changes you will ask for. New pages, a price update, a staff photo, one more form field. Most businesses need more of this than they budget for, and most quotes treat all of it as extra.

Priced separately, those four are simple to shop. Rolled into one figure they cannot be compared at all, which is a large part of why they usually arrive rolled into one figure.

Bill one, the server and the platform

Start here, because good managed hosting already covers more than most owners realize, and this is where paying twice is easiest. WP Engine publishes Startup at $30 a month, Professional at $55, Growth at $109 and Scale at $276, each published as a "starting at" figure on annual billing, with a Core plan from $400. Startup covers one site, 25,000 visits a month and 10 GB of storage. Scale covers 30 sites and 400,000 visits. Flywheel publishes a similar shape at $25 a month for a single site billed at $300 a year, $96 for its Freelance plan and $242 for Agency, with extra sites at $20 a month each.

The tiers matter less than the feature list every plan carries: "Daily & on-demand backups", "Managed WP, PHP & MySQL updates", "Security patching & plugin risk scans" and round-the-clock WordPress support. Hold a care plan proposal next to that list and a good share of the proposal is often the same work, sold a second time.

So the first question to ask about any maintenance quote is dull and worth asking out loud: which of these lines is my host already doing? If the answer is most of them, what you are being offered is supervision rather than labor. Supervision has real value on a busy site. It is not worth $200 a month on five pages that change twice a year.

Bill two, the update treadmill

This is the bill you cannot responsibly skip, and the underlying numbers got worse last year. Patchstack's State of WordPress Security in 2026, with data updated 25 February 2026, counted 11,334 new vulnerabilities in the WordPress ecosystem during 2025, a 42% increase on 2024. Of those, 4,124 (36%) were serious enough to require a protection rule and 1,966 (17%) carried a high severity score, meaning they were likely to be picked up by automated mass-scale attacks. Highly exploitable issues rose 113% year over year.

Where the flaws live is the useful part. 91% of the new vulnerabilities were found in plugins and 9% in themes. WordPress core accounted for six, all low priority. The platform is not what puts your site at risk. The stack of plugins bolted onto it is.

Two further findings should change what a sensible plan looks like. On speed, roughly half of high impact vulnerabilities are exploited within 24 hours, and once the figures are weighted by how intensely each one was attacked, the median time to first exploitation is five hours. On coverage, 46% of vulnerabilities had received no fix from the vendor by the time they were publicly disclosed, so for almost half of them, keeping everything updated is not an option that exists on day one.

The network layer does not rescue you either. Patchstack tested common defenses against known exploited vulnerabilities and found they blocked 12% of WordPress-specific attacks. A broader test across popular hosting companies still blocked only 26% of vulnerability attacks. On relying on updates alone, the report's verdict is short: "not a viable defence".

None of that means a five-page brochure site is under constant siege. It means the honest description of the security half of a care plan is fast reaction and clean restores, not prevention. Anyone selling prevention as a guarantee is selling something that nobody currently has.

What skipping it looks like across the whole web

Most sites are behind, and you can check that yourself rather than take our word for it. The WordPress.org statistics feed, read on 8 August 2026, puts 65.1% of installs on the current major version, 7.0. About 35% report something older, and the tail runs all the way back to releases from the early 2010s.

The PHP figures are starker. 17.7% of WordPress sites still run PHP 7.4, which reached end of life on 28 November 2022. Another 11.9% run 8.1, which stopped receiving security fixes on 31 December 2025, and 4.3% run 8.0, finished since 26 November 2023. Add everything on 7.3 and below, and 40.0% of WordPress sites are on a PHP version that receives no security fixes at all. Only 8.2, 8.3, 8.4 and 8.5 are still supported, and 8.2 loses its security support on 31 December 2026, which will move another 24.9% of WordPress sites into the dark bars on that date unless they are upgraded first.

This is the strongest argument for the boring half of maintenance. Nobody skips a PHP upgrade deliberately. It happens because the upgrade might break a plugin, because nobody owns that risk, and because the site keeps loading either way, so the decision slides to next quarter until it is four years late and the afternoon job has quietly become a project.

Bill three, somebody actually doing the work

Here is the number that reframes the whole category: the software costs almost nothing, and the people cost everything. ManageWP, one of the dashboards agencies run this work through, is free forever for update management, monthly cloud backups, security checks and performance checks across unlimited sites. Its premium add-ons are billed per website per month at $1 for uptime monitoring, automated security checks, automated performance checks and client reports, and $2 for backups and vulnerability protection. Agencies can bundle each add-on for up to 100 sites from $25 a month, or take everything for $150.

Attach a person and the price multiplies by a factor of twenty or more. GoWP publishes a maintenance plan at $39 a month per site covering WordPress updates, 90 days of off-site backups, daily security scans with malware cleanup and a visual check on every update. WP Maintainer publishes a $99 a month subscription covering core, theme and plugin updates, scheduled backups and security scanning with cleanup, then bills on-demand developer work at $99 an hour, only with your approval first.

Compare the first aqua bar with the three below it. Monitoring a site costs about two dollars a month. Having somebody accountable for it costs $39 to $99, and the moment that person has to make a judgment call it becomes an hourly rate. That is the right shape for the pricing, because automation is cheap to run and judgment is not. Nobody is charging you $99 to click update. They are charging for the decision at eleven at night when the update took the checkout down.

Two things follow from that. Cheap plans are being honest about being thin rather than cutting corners, and even the priciest line on the chart costs less than one incident handled badly. If your site takes bookings or payments, price an hour of downtime before you price any plan. That arithmetic is what our work with online stores usually turns on.

Bill four, the changes you will ask for anyway

Budget for monthly changes, because you will want them, and check whether they are included, because usually they are not. GoWP prices content edits as a separate product at $99 a month per site, sitting on top of the $39 maintenance plan rather than inside it. WP Maintainer bills the same kind of work hourly at $99. If you would rather have a named person than a ticket queue, GoWP publishes dedicated staff from $1,299 a month for a developer, copywriter or designer, and $1,099 for a virtual assistant.

That $1,299 line is the most instructive number on this page, because it marks the edge of the category. Below it you are buying upkeep. At it you are buying capacity. A business that sends three small requests a week is usually buying upkeep when it needs capacity, and paying for both.

What each option covers, side by side

That grid settles nothing on its own. It exists because these four options are almost never quoted in the same units, and once they are, most owners spot their own answer inside a minute.

When you can safely spend nothing

Plenty of sites need no maintenance plan, and we would rather say so than sell one. If your site sits on Squarespace, Wix or Shopify, updates and patching are the platform's job and already inside the subscription. If it is a handful of static pages with no forms, no logins and no payments, hosted somewhere sensible, there is very little that can rot. If your traffic is a few hundred visits a month and the business runs on referrals, the risk of an outage is measured in mild embarrassment.

The honest floor in those cases is a backup you have tested restoring at least once, a renewal reminder for the domain, and someone's phone number for the day something breaks. That is a few dollars a month, not a few hundred, and a plan sold against those conditions is selling comfort rather than protection. Our comparison of builders and hiring a designer walks through where that line usually falls.

The moment the calculation changes is the moment the site starts carrying revenue. Paid traffic landing on it, bookings running through it, patient or client information passing through a form. At that point an outage costs money on a schedule, and the maintenance question stops being about tidiness.

Does a site built without WordPress escape all this

Partly, and it is worth being precise about which part. A static site has no plugin layer, so the 11,334 vulnerabilities counted above are almost entirely somebody else's problem, and there is no database or PHP process on the public internet to keep patched. That is the single biggest reason we build the way we do, which our comparison of WordPress and modern static builds covers in detail.

What does not go away: dependencies still age. Node.js 20 reached end of life on 30 April 2026 and Node.js 18 on 30 April 2025, so a codebase left untouched for two years ends up on a runtime with no security fixes, exactly like the PHP 7.4 sites above. Frameworks ship major versions. Forms, analytics, booking widgets and payment scripts are third-party code with their own risk, which is the thread running through our piece on what a healthcare practice site has to do. And content goes stale on every platform ever built.

So a static build changes the shape of the maintenance bill rather than deleting it. Less urgent patching, no weekly plugin roulette, more attention on content and conversion. Anyone telling you their stack needs no upkeep at all has either not run one for three years or is not counting.

What we charge, and what sits inside it

We quote on a call, once the work is scoped. A focused build is flat and one time. For continuing work we run a monthly partnership, which is not a care plan with a nicer name: hosting, updates, backups, monitoring, content changes and design work all sit inside it, which is why it is priced against the capacity line rather than the upkeep line. The full breakdown is on the pricing page.

More than 200 projects have shipped under that model. One named result: 20% more website conversions at Cornerstone Healing Center once the redesign was live. Another: Eden Digital, a site we built, grew revenue 10X. A focused build usually goes live in about two weeks, and partnership requests get an answer inside 48 hours.

Who it suits: businesses whose website has to keep earning after launch, in fields where trust is doing most of the selling, such as healthcare practices and the service businesses competing for the same searches every day. Who it does not suit: a five-page site that changes twice a year. Buy a $39 plan for that one and spend the difference on something that grows.

Five questions before you sign anything

Ask these of any provider, including us, and the answers will separate a real plan from a line item in about ten minutes.

What does my host already do? Get the overlap in writing. If the plan repeats what managed hosting includes, the price should reflect that it is oversight rather than work.

Who applies updates, and when? Weekly, monthly or automatically, on staging or straight to live, and what happens when a plugin has no patch available. Given a five-hour median to exploitation, monthly is a posture, not a policy.

Where are the backups, and has a restore been tested? Retention, storage location and the date somebody last restored one successfully. An untested backup is a hope.

What counts as an edit? Get examples, a turnaround time and a monthly cap. This is where most maintenance relationships actually break down.

Who picks up the phone at 11pm, and what does that cost? A named person and a response time, or an admission that there is not one. Both answers are fine. Only the vague version is a problem.

Anything that survives those five questions is worth the money. Anything that does not is a subscription you will forget you are paying for, which is a fairly common way to spend $1,200 a year on nothing.

We build sites for a living, so read this knowing we are one of the two options on the table. What we can offer instead of neutrality is arithmetic you can check. Every plan price below was read off each company's own pricing page on 7 August 2026. Every performance number comes from HTTP Archive's 2025 measurement of millions of live sites. Every claim about what you can take with you when you leave comes from the platform's own help documentation. Where a builder wins, we say so.

The 30-second answer

Use a website builder when the site is essentially a brochure, you have more time than budget, and nothing about how it works will need to change. Published plans run from $12 a month at Squarespace to $49 at Shopify's Grow tier, billed annually, plus your own hours. Hire a designer or a studio when the website carries revenue: when it has to convert paid traffic, rank against funded competitors, connect to booking, intake or CRM systems, or represent a brand people trust with money or their health. Speed is no longer the tiebreaker. In 2025, 74% of Wix sites and 69% of Squarespace sites passed Core Web Vitals on mobile, against 45% of WordPress sites. The real tiebreakers are conversion, ownership and how many of your own weeks you are willing to spend inside an editor.

What the builders actually charge

Less than most owners expect, and more than the headline. Every platform advertises a cheap entry plan and sells most businesses something above it, so the useful comparison is two numbers per platform: the cheapest plan that lets you use your own domain, and the tier a growing business usually ends up on.

Squarespace publishes Basic at $12 a month, Core at $17 and Advanced at $25, billed annually, with a 14-day free trial and one year of domain registration included on annual plans. Wix publishes a free tier, then Light at $17, Core at $29, Business at $39 and Business Elite at $159, all yearly in US dollars. Webflow publishes a free Starter plan limited to a webflow.io address, 2 static pages and 1 GB of bandwidth, then Basic at $15 and Premium at $25 billed yearly, priced per site, with its Team plan at $2,500 a month on an annual contract. Shopify publishes Basic at US$19, Grow at US$49, Advanced at US$299 and Plus from US$2,300 a month when paying yearly.

Read those bars as the floor rather than the bill. They cover software and hosting. They do not cover the work. In some industries the software a business already runs folds a website into that floor at no extra charge, which is what we found when we priced the market for gyms and fitness studios.

The costs that arrive after you subscribe

Four line items reliably show up later, and all four are published rather than hidden. Knowing them in advance is the difference between a $200 year and a $900 one.

Transaction fees on what you sell. Squarespace charges a 2% online store transaction fee on its Basic plan and 0% on Core and Advanced, and takes 7% on digital content and memberships at Basic, 5% at Core and 0% at Advanced. Those percentages sit on top of whatever the payment processor charges. On memberships alone, every $1,000 of sales hands back $70 at Basic and nothing at Advanced, so the tier you sit on decides how much of your own revenue you keep.

The domain after year one. Free-domain offers on both Squarespace and Wix cover the first year of a new registration. Renewal is a separate line on next year's card statement, and Wix notes that the voucher cannot be used to extend a domain you already own.

Add-ons priced like separate products. Webflow sells Optimize, its A/B testing and personalization layer, from $299 a month based on page views, Analyze from $9 a month based on sessions, and Localize from $9. A $25 plan with two add-ons is not a $25 plan.

The feature you assumed was included. On Squarespace, custom CSS and JavaScript and professional email through Google Workspace both start at the Core tier, not Basic. On Webflow, the CMS starts at Premium. Most upgrade decisions are triggered by one missing feature rather than growth.

What the performance data actually says

The builders now beat the average WordPress site, and it is not close. This is the finding most agencies quietly skip, so here is the source in full: the 2025 Web Almanac CMS chapter from HTTP Archive measures real visits to millions of live origins, not lab tests of demo pages.

On mobile in 2025, the share of sites passing all three Core Web Vitals was 85% for Duda, 77% for Shopify, 74% for Wix, 69% for Squarespace, 65% for Webflow and 45% for WordPress. The WordPress figure is drawn from 4,840,897 origins, so it is not a sampling artifact. It is what happens when a platform runs on every tier of hosting and every combination of theme and plugin that anyone has ever installed.

The trend matters more than the ranking. Wix went from 40% of sites passing in 2023 to 57% in 2024 to 74% in 2025. Squarespace went from 33% to 60% to 69% across the same three years. WordPress moved from 28% to 40% to 45%, real progress, still last of that group. When one company owns the whole stack it can ship a rendering improvement on a Tuesday and every customer inherits it, whether they read the release notes or not.

Two caveats before anyone screenshots the chart. First, these are platform averages, and a carefully built WordPress or custom site sits far above its platform average, which is exactly why the ceiling on a bespoke build is higher than any hosted plan can reach. Second, speed is a floor and not a strategy. Google's own search documentation says Core Web Vitals feed its ranking systems while also making clear that the most relevant content can win with weaker page experience. Passing the metrics stops you losing. It does not make you win.

What everyone else is actually running

WordPress still runs most of the web, and the builders are smaller than their marketing suggests. W3Techs surveyed the web on 7 August 2026 and found WordPress on 41.2% of all websites, a 59.1% share of sites using a content management system at all. Shopify sat at 5.3%, Wix at 4.3%, Squarespace at 2.5% and Webflow at 0.8%, while 30.4% of sites ran none of the systems the survey monitors.

That last number is the interesting one. Nearly a third of the web is custom code, static output, frameworks or something bespoke enough that a scanner cannot label it. Popularity is a weak argument in either direction, but it does kill the idea that hiring a developer is an exotic choice.

When a builder is genuinely the right call

There is a version of this business that takes every project that walks in. It is not the one we want to run, so here is the case against hiring us. Stay on a builder, or start on one, in these situations.

Your site is five or six pages of information and a contact form, and you can write the words yourself. Your budget for the whole year is under a thousand dollars, in which case a builder is not the compromise, it is the only responsible option. You enjoy the tinkering, which sounds trivial and is not, since the person who likes editing their own site keeps it current and the person who dreads it lets it rot. You are testing whether the business works at all, where the correct move is to spend nothing on design and everything on finding out. Or you are a solo operator whose customers arrive through referrals and a Google Business Profile, and the website is a credibility check rather than a sales channel.

Recognize yourself in two of those and the answer is already settled. Open a free trial this afternoon. Spending five figures on a site that a $17 plan would have handled is one of the more common ways to hurt a young business.

Where a builder starts costing you money

The bill stops being the plan price and starts being what the site fails to earn. Ongoing upkeep is its own line too, and our breakdown of what website maintenance costs prices that half out. Four thresholds tend to mark the change.

You are buying traffic. Paid clicks make conversion rate the whole game, and template layouts are built to be safe across ten thousand businesses rather than persuasive for yours. Once ad spend passes the cost of a build, tuning the page beats buying more clicks. Our piece on traffic that never becomes leads covers where that leak usually sits.

The site has to do something. Multi-step intake, insurance verification, gated resources, a booking flow that writes to your CRM, a pricing calculator. Builders bolt these on through apps and embeds, each with its own subscription, its own look and its own failure mode. This is where a $29 plan quietly becomes a stack of subscriptions nobody owns.

You need to look like nobody else. Template families are recognizable to the people you most want to impress, and in fields where trust drives the decision, treatment centers, clinics, law firms, finance, looking like a competitor is a real cost even though nobody will invoice you for it.

Your own hours became the expensive part. This is the one owners underestimate. A few evenings a week inside an editor for a couple of months is not free, it is simply unbilled, and it is time not spent on the work only you can do.

What you can actually take with you

Less than you think, and this is the part to read before you commit years of content to a platform. Both leading builders publish exactly what happens if you leave, and both are more restrictive than the marketing implies.

Squarespace exports an .xml file aimed at WordPress. It carries layout pages, text blocks, image blocks, gallery pages and one blog page with its posts. It does not carry your product blocks, video blocks, audio blocks, drafts, style settings, custom CSS, dropdowns, or album, cover, index, portfolio and store pages. Content in page-specific headers, footers and sidebars does not travel either. You also cannot export from one Squarespace site and import into another.

Wix is blunter. Its help center states plainly: "Your site must run on Wix's servers." The company explains that its architecture relies on proprietary technology and its own services, so external hosting is not supported, and that blog posts cannot be exported to other platforms. Your domain is yours and can be transferred away. The site itself stays.

None of that is a scandal. It is the deal you accept in exchange for never patching a server. But it changes what a subscription means: you are renting the storefront, and after four years of publishing, the cost of leaving is the rebuild, not the plan fee. A site built on open technology and hosted where you choose is a different asset, which is the honest reason we build the way we do and part of what our comparison of WordPress and modern static builds exists to explain.

What hiring actually buys in 2026

Not speed, and not hosting. Both are commodities now. What you are buying is judgment and time back.

Judgement means someone decides what the homepage argues, what the first screen asks for, which four objections the page has to answer before anyone calls, and what to cut. Builders hand you infinite freedom and no opinion, which is a comfortable trap: the site ends up describing the business instead of selling it. That is a research and positioning problem wearing a design costume, and no template solves it.

Time back is the simpler half. Somebody else runs the build, the launch, the technical work and the changes that follow, while you keep serving customers. If you are weighing a studio against putting a designer on payroll instead, the wage arithmetic sits in our in-house versus agency breakdown, and it is a different decision from this one.

What hiring does not buy is a guarantee. Any firm that promises you a ranking or a conversion rate in a first meeting is selling a number it cannot control.

The two paths across twelve months

None of that decides anything by itself. Its only job is to force both options into the same units, since almost nobody quotes them that way.

What we charge, and who it is for

Our offer is set out on our offer page, and we quote it once the work is scoped. A focused build is one time and flat. The ongoing partnership runs monthly for teams that keep shipping after launch. Builds typically go live in about two weeks, and partnership requests get a reply inside 48 hours.

We have shipped more than 200 projects. Two results we can put names to: Cornerstone Healing Center saw 20% more website conversions after we redesigned their site, and Eden Digital, which we built, grew revenue 10X. Yours would be its own number, and anyone who quotes you one in a first meeting is guessing.

The fit is strongest where the website does real work rather than sitting there looking presentable. Local service businesses fighting over the same searches every day, and online stores where a percentage point of conversion changes the year. If that is not you yet, a builder is a perfectly reasonable place to spend the next twelve months.

How to decide this week

Four questions settle it in an afternoon.

What does the site have to do, stated in verbs? Explain, book, sell, qualify, collect documents. A list of one or two verbs is builder territory.

What does a full year cost on each side? Twelve months of the plan you will realistically be on, plus add-ons, plus transaction fees on anything you sell, set against a build and a partnership. Write both totals down before either option becomes emotional. If what you are selling is products rather than services, the transaction line is the one that moves, and we have taken it apart in the percentages nobody adds up.

What are your own hours worth? Estimate the evenings the DIY route needs, then double the estimate, because nobody invoices themselves and everybody guesses low.

Where will this site live in three years? If moving it is even faintly plausible, read the export documentation now rather than discovering the limits later, and know what a move does to your search rankings before you plan one: our guide to a redesign that keeps its rankings covers the redirect work that decides it.

Then repeat the exercise next year. Businesses cross this line as they grow, and plenty of good ones belong on a $17 plan today, with our invoice showing up somewhere around 2028.

It is a fair question with a bad answer online. Search it and you get confident ranges, six-figure totals and phrases like fully loaded cost, almost never with a source attached to either end of the range. The inputs you need are public, though. The Bureau of Labor Statistics publishes what these roles are paid and what employers spend on top of the paycheck. Design firms publish their prices. What follows is that comparison assembled from those two sets of numbers and nothing else, including our own prices, because we are one of the options on the table and it would be odd to pretend otherwise.

The 30-second answer

Hiring costs more than the salary and less than the internet claims. In May 2025 the median US web developer earned $92,650 and the median web and digital interface designer earned $104,000. Benefits add roughly half again on top: for private-industry professional employees, wages are 68.5% of what the employer actually pays, so a $92,650 salary sits closer to $135,000 a year in real cost, before tools or management time. On the other side, design subscriptions publish monthly prices starting at $699, and one large firm publishes a $15,000 monthly minimum. We quote our own partnership on a call. The deciding factor is not the rate. It is whether you have close to full-time work for one skill, or part-time work for three.

What these people actually earn

Start with the wage data, because every other number in this comparison is built on it. The BLS Occupational Employment and Wage Statistics program surveys employers rather than asking workers, and its May 2025 estimates are the most recent published.

Three occupations cover most of what a business means when it says it wants to hire a website person. Web developers (70,190 employed) had a median annual wage of $92,650 and a mean of $98,770. Web and digital interface designers (113,330 employed) came in higher, at a $104,000 median and a $117,490 mean. Graphic designers (197,830 employed) had a $62,960 median and a $70,560 mean.

The spread inside each occupation matters more than the midpoint. Web developers at the 10th percentile earned $48,100 and at the 90th percentile $162,290. For interface designers the same two points are $53,750 and $201,550. That is not noise. It is the difference between someone who can update pages and someone who can own an entire digital product, and you do not get to pay the first number for the second person.

Two more figures are worth having, because the job that gets invented after launch is usually marketing rather than code. Marketing managers averaged $177,770 a year and market research analysts and marketing specialists averaged $89,490.

The number salary comparisons leave out

A salary is not what an employee costs. The BLS measures the gap every quarter in Employer Costs for Employee Compensation, and for our purposes the useful cut is private industry, management, professional and related occupations, which is where designers and developers sit.

In the first quarter of 2026, employers in that group paid $78.10 per hour worked in total compensation. Of that, $53.50 was wages and salaries and $24.61 was benefits. So wages are 68.5% of the bill and benefits are 31.5%. Restricting it to full-time workers only pushes the total to $81.37 an hour.

Turn that ratio around and it becomes a multiplier: divide a salary by 0.685 and you get roughly what the employer pays, which is about 46% more than the salary itself. Applied to the medians above, a web developer lands near $135,300, an interface designer near $151,800 and a graphic designer near $91,900.

Three honest caveats, since this is the number people misuse. These are national figures and your metro will differ. The BLS wages-and-salaries category is slightly broader than base salary, so treat the multiplier as an estimate rather than a quote. And it still excludes hardware, software seats, desk space and the hours a manager spends directing the work, none of which we are going to invent a figure for.

Where that extra third goes

It goes mostly to things you cannot switch off. For the same group of professional workers in the first quarter of 2026, the $24.61 an hour breaks down into paid leave at $7.34, insurance plans at $5.73 (of which health insurance is $5.36), legally required benefits such as Social Security, Medicare, unemployment insurance and workers compensation at $5.03, supplemental pay at $3.60 and retirement and savings at $2.89.

Look at the first bar again. Paid leave is the biggest benefit an employer buys, and the thing it buys is time when the person is not working. That is completely reasonable as employment policy and completely relevant to a comparison, because the week your only designer is on holiday, your website work stops. Firms carry that risk across a team. A one-person hire hands it to you.

Salaries also do not hold still. The Employment Cost Index for private industry rose 3.3% over the twelve months to June 2026, so the offer you make this year is a floor rather than a ceiling.

One salary buys one skill

Notice that the BLS treats developers, interface designers and graphic designers as three separate occupations with three separate wage levels. That classification is not bureaucratic tidiness. It reflects that the market pays for three different skills.

Most business websites need all three at some point, plus somebody who can write. Hire one person and you get one of them, at a good level, and whatever they can improvise for the rest. Plenty of strong people genuinely cover two. Almost nobody covers three at a level you would pay a premium for, and if they claim they do, the wage percentiles suggest what that person actually costs.

If you did hire all three at the medians, the salaries alone come to $259,610, and the loaded cost lands near $379,000. That arithmetic is also, we suspect, the origin of the alarming ranges quoted in the comparison articles. They are describing a department, not a hire.

What it costs to go and find them

Recruiting is a real line item and it is smaller than most people fear. SHRM's 2026 recruiting benchmarking brief, based on 4,657 member responses collected between November 24, 2025 and January 23, 2026, puts the median cost-per-hire for nonexecutive positions at $1,300, up from $1,200 in 2025. Executive hires are a different animal at a $15,000 median.

Time is the bigger cost. The same brief reports a median time-to-fill of 39 calendar days for nonexecutive roles, down from 44 days in 2025. Add notice periods and the ramp needed to learn your business, and a hire made today starts producing useful work in the following quarter.

The market is also not slack. In June 2026 the job openings rate in professional and business services was 5.5%, against 4.4% across all nonfarm jobs, and 2.2% of people in the sector quit during that single month. You are hiring into competition, and once hired, the person can leave.

What the other side of the comparison publishes

Agency pricing is famously opaque, so use the firms that publish numbers. ManyPixels lists its design subscriptions at $699 a month for one daily output, $1,199 for two, $1,399 for a part-time assigned designer and $2,599 for two, with lower rates on quarterly and yearly billing. At the other end, Superside states that its subscriptions start at a $15,000 monthly minimum on an annual term plus a $1,000 monthly software fee, and that its dedicated team option starts at $30,000 a month on a twelve-month term.

Our own numbers sit between those poles. A focused build is charged once, and the ongoing partnership runs monthly.

The published spread runs more than twenty to one, and the variable is not quality claims. It is volume, seniority and how much of the thinking is included. A $699 subscription executes briefs you write. A $30,000 team decides what the briefs should be. Most owners reading this need something in the middle, which is exactly the band where nobody publishes anything, so ask for the number in the first call and treat reluctance as information.

The break-even is hours, not rates

Here is the comparison that actually decides it, and the BLS has already done the hard part. That $81.37 an hour worked for a full-time professional employee is the true unit price of an in-house person, and you pay it for all 2,080 hours of the year whether or not the work exists.

So count the work. If your website, campaigns and collateral genuinely generate thirty or more hours a week, every week, in-house is cheaper per unit of output and better in ways that do not show up in a spreadsheet: they sit in your meetings, they know why the pricing page says what it says, and their context compounds. Hire.

If the honest number is six hours some weeks and none the next, you are not buying a designer. You are buying availability, and availability priced at a salary is the most expensive way to get it. That is the pattern behind most of the businesses that come to us after a hire did not work out. The person was good. The work was not there, so they drifted into being a generalist marketing assistant, then left.

Somewhere between those two poles is a genuine grey zone, roughly fifteen to twenty-five hours a week of specialized work. In that band, price both and pick on management appetite rather than cost.

When hiring is clearly the right call

We would rather say this plainly than lose the trust of a reader we could never have served well.

Hire in-house if your interface is your product, because a software company outsourcing its core screens is outsourcing its roadmap. Hire if you ship creative daily, as retailers, publishers and high-volume advertisers do, since the volume alone justifies the salary. Hire if your work cannot leave the building for security or regulatory reasons. And hire if you already employ someone who can direct design work, because that is the quiet prerequisite: a designer without a decision-maker waits for briefs, and waiting is the most expensive thing they can do.

If two or more of those describe you, hire, and do not let anyone with an agency website talk you out of it.

The arrangement most of our clients end up with

One internal owner, one external team. The internal person owns priorities, content, approvals and the relationship with the rest of the business. Often it is part of an existing marketing role rather than a new headcount, which is worth noting given that marketing specialists average $89,490 and marketing managers $177,770. The external team owns design, build and the parts that need three skills at once.

It works because it splits the two things that get bundled by accident. Context belongs inside. Capacity does not have to.

Year one, side by side

Nothing in that table settles the argument on its own. It exists so the comparison happens in the same units, which is the part most quotes quietly skip.

What we charge, and what sits inside it

Two numbers, both public on our pricing page. Our build fee is flat for a focused scope, not a from-price that grows once scope arrives. The monthly partnership covers businesses that keep shipping after launch. Launch on that build is usually around two weeks, and partnership requests get an answer inside 48 hours.

What that buys, in the terms of this article, is three skills without three salaries, plus the coverage a single hire cannot give you. The evidence we point at is specific rather than promised: Cornerstone Healing Center saw 20% more website conversions after we redesigned their site, and Eden Digital, which we built, grew revenue 10X. Those are results for named clients, not a forecast for yours, and any firm that hands you a forecast in the first meeting is selling you something it cannot deliver.

Businesses in local services and early-stage startups tend to sit furthest from the hiring threshold, since the work arrives in bursts. That is the case we are built for.

Whichever way you go, settle ownership before the first invoice: our guide to who owns your website covers the assignment clause, the domain registrant field and the accounts to get in your own name.

How to decide inside a week

Four steps, none of which require a consultant.

One, log the actual hours. For two weeks, write down every website, design and marketing-asset task and how long it took. Most owners are shocked by how small the real number is, and it is the only input that matters.

Two, name the skills those hours needed. If they span all three occupations above, one hire will not cover them.

Three, ask who directs the work. If the answer is nobody, fix that before you spend anything.

Four, price both over the same twelve months, salary plus 46% against twelve monthly fees plus any build cost, and let the two totals sit next to each other before anyone falls in love with an option.

Then run the same test again a year later. Businesses cross the hiring threshold as they grow, and the honest version of this article ends by admitting that the right answer for a client today may be to hire someone in three years and stop paying us. If you want the neighboring numbers, our breakdown of what a professional-services website costs covers the build side in detail, and our piece on traffic that never turns into leads covers what tends to be wrong when the spending has already happened.

It also does far less than the panicked coverage suggests. A lot of what agencies now tell clients is banned is not banned by this rule at all. Knowing which is which is worth an afternoon, because the parts that do apply are cheap to get right at build time and awkward to fix after someone complains.

The 30-second answer

You can put testimonials on your website. What you cannot do is invent them, buy them, or hide the fact that the person praising you works for you. The FTC's Rule on the Use of Consumer Reviews and Testimonials took effect on October 21, 2024, and it bans fake or false reviews and testimonials, paying for a particular sentiment, undisclosed insider reviews, company-controlled sites posing as independent, bullying reviewers into silence, and buying fake followers. Courts can impose civil penalties of up to $53,088 per violation for knowing violations. The rule's much-discussed ban on hiding negative reviews is narrower than people think: it applies to a reviews section you host, not to a curated wall of happy quotes on a landing page. Curating your marketing is legal. Fabricating it is not.

What actually changed in 2024

The FTC announced the final rule on August 14, 2024, and it was published in the Federal Register on August 22 as 89 FR 68034. The text now lives at 16 CFR Part 465, and it took effect on October 21, 2024.

Deceptive reviews were already illegal before this, under Section 5 of the FTC Act, and the agency brought cases and recovered money under it. What a trade regulation rule adds is a more direct lever: civil penalties against knowing violators, rather than having to build the case from scratch as unfair or deceptive conduct each time. The current maximum is $53,088 per violation, set out in the penalty table at 16 CFR 1.98.

Two words in that sentence matter. Knowing, because the rule is aimed at businesses that understood what they were doing rather than at honest mistakes. And per violation, because the count is not one penalty per company.

Lina M. Khan, then chair of the FTC, described the harm this way when the rule was announced: "Fake reviews not only waste people's time and money, but also pollute the marketplace."

A review is not a testimonial

Almost every mistake we see starts here. The rule defines these as two different things and applies different provisions to each.

A consumer review is an evaluation submitted by a customer and published on a site or section dedicated to collecting and showing such evaluations. Your Google profile is full of reviews. So is the star-rating block on a product page, if customers submit to it.

A consumer testimonial is an advertising message that reads as the opinion or experience of a customer. The quote card on your homepage is a testimonial. Nobody submitted it to a review system. You chose it, designed it, and published it as marketing.

The FTC's own staff guidance puts it bluntly in its questions and answers on the rule: most consumer reviews are not consumer testimonials, and most consumer testimonials are not consumer reviews. If you take one thing from this article, take that sentence, because the rules that got the most attention apply to reviews, while the thing on your website is usually a testimonial.

The six things the rule bans

Stripped of the legal scaffolding, Part 465 prohibits six categories of conduct.

Fake or false reviews and testimonials. Writing, creating or selling a review or testimonial that misrepresents that the person exists, that they used your product or service, or what their experience was. Buying them or spreading them counts too, when you knew or should have known.

Buying sentiment. Providing compensation or incentives conditioned, openly or by implication, on the review being positive or negative.

Undisclosed insider reviews and testimonials. Quotes from your officers, managers, employees, agents or their immediate relatives, without a clear disclosure of the relationship.

Company-controlled review sites posing as independent. The FTC's own example is a company that invents a seal program and then awards the seal to its own products.

Review suppression. Two flavors: bullying a reviewer with groundless legal threats, intimidation or knowingly false public accusations, and misrepresenting that the reviews shown in your review section are all of them when you filter by sentiment.

Fake social media indicators. Selling or buying followers, likes and views generated by bots or hijacked accounts, to misrepresent influence.

Notice what is absent. Nothing here requires you to publish every kind word you receive, run a review section, or show a star rating.

Real people, real quotes, real faces

The most common violation we find on a site audit is not a fabricated quote. It is a real quote attached to a face that belongs to a stock photography model.

The Endorsement Guides at 16 CFR Part 255 deal with this directly. An advertisement presenting endorsements by people represented as actual consumers should either use actual consumers, or clearly and conspicuously disclose that the people shown are not. A smiling headshot next to a client quote represents that person as the client. If the headshot came from a stock library, the page is making a claim that is not true, and the disclosure that fixes it defeats the purpose of using the photo.

The same logic applies to generated faces. A model that produces a photorealistic person who has never existed is a faster way to arrive at the same problem.

The FTC's July 2025 case against telemedicine firm NextMed shows what this looks like when it goes wrong. Among the charges, the agency alleged the company used testimonials and before-and-after photos from people who were not its clients. The company and its principals agreed to pay $150,000 and to stop misrepresenting reviews.

The fix is unglamorous and it is what we do on every build: initials or no image at all, unless the client has given you a real photograph. An honest text card converts better than a face nobody believes anyway.

When your own team is the testimonial

Section 465.5 is the provision most likely to catch a small business acting in good faith. If an officer or manager writes a review of the business, it needs a clear and conspicuous disclosure of the relationship. If you publish a testimonial from an employee or an agent without disclosing it, and you knew or should have known, that is on you.

Clear and conspicuous has a specific meaning in the rule, and it is stricter than most disclosure design. The disclosure has to be unavoidable. The definition says outright that a disclosure is not clear and conspicuous if a consumer has to click a hyperlink or hover over an icon to see it. An asterisk leading to a footnote at the bottom of the page does not satisfy it. Text in the first line of the review does.

There are two sensible escape hatches. The first is that no disclosure is needed when the relationship is already obvious to the audience, for example when the person is filmed in company uniform in the company's office. The second is the generalized solicitation carve-out, and it is broader than people expect.

The FTC's staff guidance works through the exact scenario: you email all recent customers asking for a review, a few of them happen to be employees, and they post without disclosing. That is not a violation of this section. The answer does not change if you offered an incentive to everyone who received the email, as long as the incentive was not conditioned on saying something nice.

One caution the guidance adds is easy to miss. Insider reviews with proper disclosures can still create a problem under the FTC Act if they materially lift your average star rating, because plenty of people read only the number.

How to ask for reviews without buying them

Asking customers for reviews is normal, effective and legal. Paying for a verdict is not. The line sits exactly at whether what you offer depends on the sentiment.

A discount for anyone who leaves a review, positive or negative, is on the safe side of the rule. A discount for a five-star review is not. Neither is a prize draw for customers who "share what you loved", because the condition is implied by the wording. The rule explicitly covers incentives conveyed by implication.

BrightLocal's Local Consumer Review Survey 2026, which surveyed 1,002 US adults, gives a sense of how common each behavior is.

Seventy-eight percent were asked for feedback in the previous year and fifty-nine percent were offered some reward for it, which are both fine. Eleven percent were offered an incentive specifically to write a positive review, which is the conduct the rule bans. That is a meaningful share of the market doing something that now carries a penalty exposure.

If you want the volume without the risk, ask everyone, ask soon after the work is delivered, and make the ask identical regardless of how you expect the customer to feel. We build that request into the delivery step of client projects rather than leaving it to a marketing afterthought.

Can you show only your best testimonials

Yes, and this is where most published guidance is wrong.

The suppression provision is narrow by its own terms. It applies to reviews displayed in a portion of your website or platform dedicated in whole or in part to receiving and displaying consumer reviews. The FTC's staff guidance is explicit that the provision does not prohibit the selective use of particularly positive consumer reviews in marketing materials.

So the wall of glowing quotes on your homepage is not a violation of the suppression rule. Our own site has one.

The caveat is real, though, and the same guidance states it. Using reviews that are not representative of what customers generally experience can still be deceptive under Section 5 of the FTC Act, and the Endorsement Guides say an ad relating one customer's experience on a key attribute will be read as a claim that the experience is typical. If your average client sees nothing like the outcome in your featured quote, the quote needs context, not just a disclaimer.

If you do run a review section that collects customer submissions, the equal-treatment principle governs. You may withhold reviews for reasons applied to every submission regardless of sentiment, such as reviews containing personal information, abusive or discriminatory content, confidential commercial information, content you reasonably believe is fake, or content unrelated to what you sell. What you may not do is filter by rating and then imply the survivors are the full set.

The review widget that cost $4.2 million

The clearest illustration of that distinction predates the rule. In January 2022 Fashion Nova agreed to pay $4.2 million to settle allegations that it blocked negative reviews of its products from its website. It was the FTC's first case about concealing negative reviews.

The mechanism is what web teams should sit with. According to the complaint, the retailer used a third-party review management interface configured to post four and five star reviews automatically while holding lower-rated ones for approval. From late 2015 until November 2019, those hundreds of thousands of held reviews were never approved.

Nobody had to write a fake review. Somebody chose a default in an admin panel. That default is still shipped by review plugins today, usually described as moderation, and it is the single setting we check first when we inherit a site with a reviews section.

The other two cases in that chart show how varied the exposure is. Leader Automotive Group and its parent AutoCanada agreed in December 2024 to turn over $20 million over a set of allegations in which bogus reviews sat alongside junk fees and bait-and-switch pricing. NextMed's total was $150,000. Reviews were one charge among several in both, so neither figure is a price list for a fake testimonial.

Enforcement has not stopped. In May 2026 the FTC and the state of Illinois sued a Chicago company over thousands of fabricated home-repair business listings padded with invented five-star reviews to dilute genuine one-star ones. The company has not been found liable, and a complaint is only the government's side of the story.

Answering a bad review without making it worse

You are allowed to reply to critics, and you should. The rule restricts how, not whether.

What it prohibits is using an unfounded or groundless legal threat, a physical threat, intimidation, or a public accusation you know to be false, in order to stop a review being written or to get one taken down. Groundless has a definition: contentions unwarranted by existing law, or factual claims with no evidentiary support.

Threatening to sue over a genuinely defamatory review is still available to you, because that threat has a legitimate basis. Sending a takedown letter to every unhappy customer is the thing that turns a bad week into an enforcement matter. The FTC's guidance also notes that intimidation is not limited to physical threats and can include abusive communications and character assassination.

Contacting a customer privately to fix the underlying problem is fine, and so is asking a satisfied customer to update an old review. What draws attention is offering someone money to delete a negative review, which the guidance flags as a possible unfair practice even though the rule does not ban it outright. NextMed was alleged to have offered gift cards to consumers to remove or change negative reviews, and to have conditioned refunds on it.

The results claim hiding inside your testimonial

A testimonial that contains a number is two claims at once: that the customer said it, and that the number is achievable.

The Endorsement Guides require substantiation for the second one. If a quote describes an outcome on a central attribute of what you sell, it will be read as representative of what customers generally get. Where you cannot substantiate that, the guides call for a clear disclosure of the performance customers should generally expect, and that disclosure has to change the overall impression rather than sit in small print.

This is why we are careful with our own numbers. We say that Cornerstone Healing Center saw 20% more website conversions after we redesigned their site, and that we built Eden Digital and it grew revenue 10X, because those are specific results for named clients that we can point to. We do not present either as what a new client should expect, because we have no basis for that claim, and a reader who signs up expecting 10X is a refund request in slow motion.

Regulated industries carry more weight here. If you market treatment, therapy or medical services, a patient testimonial describing outcomes needs care that a plumbing testimonial does not, and several state boards have their own restrictions on top of federal law.

Who is actually on the hook

The business, and often the people it hires.

The FTC's guidance confirms that advertising agencies, public relations firms, review brokers and reputation management companies are not immune. An agency that writes a fake testimonial, runs an incentive program conditioned on positive sentiment, or handles review suppression for a client can be liable in its own right. We include ourselves in that. It is one of the reasons we will not implement a review filter that hides one-star submissions, whoever asks.

Two limits are worth knowing. Ordinary consumers cannot be liable under the rule for what they write, and there is no private right of action, so a competitor cannot sue you under this rule directly. That does not stop them complaining to the FTC, or suing under other law.

We are a design and development studio, not a law firm, and none of this is legal advice. For anything with real money attached, ask a lawyer who practices advertising law.

What we put on client sites

Our standing rules are short. Real names or initials, never invented ones. A real photograph or no photograph. Verbatim quotes, trimmed for length but never for meaning, with the client's written permission to publish. Any employee, investor or founder quote labelled as such in the visible text. Numbers only where we can point at the source. Review widgets configured to publish everything that passes the same content policy, or not installed at all.

None of that is expensive. It is a set of decisions made during the design phase, which is why our one time build and our monthly partnership both include it rather than pricing it as an extra. For businesses in local services, where reviews are most of the buying decision, we treat the proof section as a first-class part of the layout rather than a strip at the bottom.

The honest counterpoint: if what you actually want is a site that looks more established than the business currently is, we are the wrong studio. Proof is the one part of a website that cannot be designed into existence.

A 20 minute audit of your own site

Open your site and check six things.

One, does every testimonial correspond to a real person who really bought from you, and can you produce the permission. Two, is every photograph next to a quote a photograph of that person. Three, is any quote from staff, family or an investor labelled in the visible text rather than a tooltip. Four, does any incentive you offer for reviews depend on the review being positive. Five, if you host a review section, does anything filter submissions by rating or sentiment. Six, does any published number in a quote have a source you could show a regulator.

Most sites we audit fail two and three. Those are also the two cheapest to fix, usually inside an hour, which makes them a strange thing to still be carrying.

If the audit turns up more than a couple of problems, the underlying issue is usually that the site was assembled from a template with placeholder content that nobody replaced, in which case the testimonials are the symptom. Our piece on why traffic does not turn into leads covers what else tends to be wrong on those pages, and our guide to ADA website compliance covers the other rulebook that reaches the same markup.

The useful question is not whether you are compliant, because that word has no official meaning here. It is which specific things on your site a plaintiff can point at, and which of them are cheap to prevent. Nearly all of them are build decisions, made in the first week of a project, by people who are not lawyers.

The 30-second answer

If your business serves the public, the ADA almost certainly reaches what you offer online, and there is still no federal regulation telling private businesses which technical standard to hit. The Justice Department says you have flexibility in how you comply, but you must comply. In practice the benchmark is WCAG 2.1 Level AA, because that is the standard the government itself adopted when it wrote a binding web rule for state and local agencies. Automated accessibility widgets do not deliver that standard: the FTC required one overlay vendor to pay $1 million in 2025 over claims that its AI could. Federal web accessibility lawsuits reached 3,117 in 2025, up 27 percent in a year.

Is your business actually covered

If you appear anywhere in the statute's list of public accommodations, yes. That list, at 42 U.S.C. 12181(7), is unusually concrete for a federal law. It names "a restaurant, bar, or other establishment serving food or drink", "a bakery, grocery store, clothing store, hardware store, shopping center, or other sales or rental establishment", "the office of an accountant or lawyer", the "professional office of a health care provider, hospital", and "a gymnasium, health spa, bowling alley, golf course, or other place of exercise or recreation".

Read that list next to the industries we work in and the answer is not close. Law firms, medical and behavioral health practices, gyms, restaurants, retailers and most local service businesses are named in the text of the statute itself.

What the statute does not say is anything about websites, because it was written in 1990. The Department of Justice filled that gap with guidance rather than regulation. Its web accessibility guidance of March 18, 2022 states that "The ADA's requirements apply to all the goods, services, privileges, or activities offered by public accommodations, including those offered on the web." The same page says businesses "can currently choose how they will ensure that the programs, services, and goods they provide online are accessible", and then adds the part people skip: "But they must comply with the ADA's requirements."

That is the whole tension. An obligation exists, and the government has declined to tell private businesses exactly what satisfies it.

Why there is no rulebook, and what filled the gap

There is no technical standard for private business websites because the Justice Department dropped the rulemaking that would have created one. On December 26, 2017 it published a notice at 82 FR 60932 withdrawing four advance notices of proposed rulemaking under Titles II and III of the ADA "for further review". Nine years later, nothing has replaced them for Title III.

What filled the vacuum was litigation. When there is no regulation defining the finish line, the finish line gets argued case by case, and the reference both sides reach for is the Web Content Accessibility Guidelines published by the World Wide Web Consortium. Build to those and you are arguing about degree. Ignore them and you are arguing about whether you tried at all.

How many lawsuits are we actually talking about

Enough that this is a normal business risk rather than an exotic one. Level Access, in a litigation analysis published May 27, 2026, counts 3,117 federal web accessibility lawsuits filed in 2025, a 27 percent increase over the 2,452 filed in 2024. Adding state court filings takes the 2025 total past 5,000. New York, Florida and California still dominate the volume, with newer activity in Illinois, Missouri and Minnesota, where the analysis says more than 300 businesses were sued in 2025.

Two honest caveats before anyone panics. The company publishing that count sells accessibility software, which is a reason to read the framing carefully even when the underlying court filings are real. And a filing is not a judgment. A large share of these cases resolve quickly and quietly, which is precisely what makes them attractive to file in volume.

The practical read is not that you will be sued. It is that a demand letter is a cheap thing for someone to send, and the cost of receiving one is much higher than the cost of not being an obvious target.

So what do you actually build to

Build to WCAG 2.1 Level AA. Not because a statute names it for private business, but because the federal government picked exactly that standard the one time it did write a binding web rule, which makes it the least arguable benchmark available.

That rule covers state and local government, not you. It was published at 89 FR 31320 on April 24, 2024 and took effect that June. Its operative text now sits at 28 CFR 35.200, and it requires public entities to comply with "Level A and Level AA success criteria and conformance requirements specified in WCAG 2.1".

WCAG 2.2 is the newer version, currently published by the W3C as a Recommendation dated 12 December 2024, and it adds success criteria on top of 2.1. If your team is building to 2.2, good. The version referenced in United States regulation today is 2.1, so 2.1 Level AA is the floor to specify in a contract and 2.2 is the direction to move.

The government deadline that quietly moved this April

It moved by a year, and it does not bind private businesses, but it is worth knowing about for two reasons. On April 20, 2026 the Justice Department published an interim final rule at 91 FR 20902 extending the Title II web compliance dates. Public entities with a population of 50,000 or more went from April 24, 2026 to April 26, 2027. Smaller entities and special district governments went from April 26, 2027 to April 26, 2028.

The first reason to care is commercial. If you sell to a city, a county, a school district or a public hospital system, or you build sites for anyone who does, that obligation lands in your scope of work and now lands a year later than the schedule most vendors were quoting.

The second is that a lot of published advice is now wrong. Guides written before this April still list April 24, 2026 as a live date. If you are reading anything on this topic that has not been updated since spring, check its dates against the rule itself.

Healthcare organizations have a separate obligation on a separate clock, under section 504 rather than the ADA, and those dates also moved in 2026. We covered that in our piece on what a HIPAA compliant website actually requires, and the two should not be confused with each other.

The overlay widget problem

An accessibility overlay will not make your website conform to WCAG, and the Federal Trade Commission has now put a price on saying otherwise. On January 3, 2025 the FTC announced that accessiBe would pay $1 million to settle allegations it misrepresented the ability of its AI-powered tool to make any website WCAG compliant. The complaint alleged the accessWidget plug-in "did not make all user websites WCAG-compliant" and that the claims were therefore false, misleading, or unsubstantiated.

Samuel Levine, then director of the FTC's Bureau of Consumer Protection, put the principle plainly: "Overstating a product's AI or other capabilities without adequate evidence is deceptive."

The complaint carried a second allegation that is arguably more useful to you as a buyer. The FTC said the company "deceptively formatted third-party articles and reviews to appear as if they were independent opinions by impartial authors" while failing to disclose its connections to those reviewers. If you researched overlays in the years before this, some of what you read was advertising. The Commission approved the final order on April 22, 2025 by a 3-0 vote, barring the company from claiming its automated products can make any website WCAG compliant, or keep it compliant over time, without evidence.

Be precise about what this does and does not mean. The FTC acted on advertising claims, not on whether overlays are lawful, and it is not illegal to run one. The point is narrower and more practical: a script that loads after your page cannot supply what your page never had. If a button contains an icon and no text, the widget is guessing what that button does. If your color palette fails contrast, the widget offers a visitor a menu to change it, which is a different thing from the page being readable when it loads.

What actually breaks, in order

Six failure types account for almost all of the detectable problems, and all six are decided during the build. WebAIM, a non-profit service center at Utah State University that has worked on accessibility since 1999, tests the top one million home pages every year. In its February 2026 analysis, "95.9% of home pages had detected WCAG 2 failures", up from 94.8% in 2025 and reversing six straight years of small improvements.

The scale is worth sitting with. WebAIM detected 56,114,377 distinct errors across the million pages, an average of 56.1 per page, 10.1 percent more than the 51 per page found in 2025. And because only automatically detectable failures were counted, the report notes that the real rate of full WCAG 2 A and AA conformance "was certainly lower than 4.1%".

Translated out of jargon, here is what those six things are. Low contrast text is a palette decision, usually grey type on white that looked elegant in a design file. Missing alternative text is an image with nothing describing it, so a screen reader announces a filename or nothing at all. Missing form input labels means a field whose label is only visual, so nobody using assistive technology knows what to type. Empty links and empty buttons are almost always icons without text, the search magnifier and the hamburger menu being the classic pair. Missing document language is one attribute on one line of HTML that tells software which language to pronounce.

None of that is expensive to get right the first time. The trend data makes the point better than we can: missing document language has fallen steadily from 33.1 percent of home pages in 2019 to 13.5 percent in 2026, because it became a default in build tooling. Low contrast text went the other way, improving to 79.1 percent in 2025 and then jumping back to 83.9 percent this year, because contrast is a taste decision that gets re-made by every new designer.

Your platform matters less than your builder

It matters, but not in the way vendors imply, and the same WebAIM data lets us check rather than guess. Detected errors per home page, broken out by content management system, ranged from 29.9 on Adobe Experience Manager to 52.8 on WordPress across 252,302 WordPress home pages.

Two readings of that chart are wrong and one is right. It is not evidence that Squarespace or Wix make a site accessible: they score well partly because they constrain what an average user can build, which is also the reason businesses outgrow them. It is not evidence that WordPress is inaccessible: WordPress carries an enormous unmanaged long tail of themes and plugins, and the median WordPress site is not built by a professional.

The right reading is that no platform on that list gets you close. The best score on the chart still means roughly thirty detected barriers on a single home page. Whatever you build on, someone has to care about this, and that person is your builder. We have written elsewhere about choosing between WordPress and a modern framework, and accessibility is one of the few dimensions where the honest answer is that the decision barely moves the needle.

The same report shows every industry failing. Business home pages averaged 52.6 errors, health and fitness 54.4, law, government and politics 46.6. There is no sector quietly doing this well.

Prevention is nearly free, retrofitting is not

The cost difference between building accessibly and fixing it later is the single most useful number in this article, and it is not a number, because we will not invent one. Nobody publishes a credible average cost of remediation, and the range depends entirely on how big your site is and how bad the templates are.

What we can describe is the shape of the work. During a build, contrast is one decision about a palette, made once, applied everywhere. Labels and button names are written into components as they are created. Alternative text becomes a field in the content process. The marginal cost is close to zero because the work is happening anyway.

Retrofitting inverts all of that. Somebody has to audit every template, find every instance, change markup that other things depend on, and test the result without breaking the design that was signed off. Then the content team has to go back through the images. On a fifty page site that is a project, not a task, and it competes for budget with everything else you wanted to do this quarter.

This is the argument for doing it at the moment you are rebuilding anyway, which most businesses do every three or four years regardless.

If you sell into Europe

Different law, different standard, and it is worth knowing which one applies before you assume the ADA travels. The EU Web Accessibility Directive obliges public sector bodies to make their websites and apps accessible, referencing the harmonised standard EN 301 549 v3.2.1. It requires three things American rules do not: a published accessibility statement naming the non-accessible content and alternatives, a feedback mechanism so users can report problems, and regular monitoring by member states with reports to the Commission every three years.

That directive entered into force on 22 December 2016 with a transposition deadline of 23 September 2018. A second instrument, the European Accessibility Act, extends accessibility requirements to a wide range of products and services in the private sector, so businesses selling into the EU should get specific national advice rather than relying on a US checklist.

The practical consequence for anyone selling both ways is small: build to WCAG Level AA and you are aligned with the substance of both, then handle the statement and feedback obligations separately as documentation.

Five questions for whoever builds your site

The answers separate a team that has done this from a team that will learn on your budget, and they take about ten minutes to ask.

Which WCAG version and level will you build to, and will you write it into the contract? How will you test it, and does that testing include using the site with a keyboard only and with a screen reader, rather than only running an automated scanner? Are you installing an accessibility widget, and if so, what specifically do you expect it to fix? Which of the six most common failures does your component library prevent by default? And after launch, what stops the next person who adds a page from undoing it?

That last one catches more teams than the others combined. Accessibility is not a state you reach, it is a property that decays every time somebody uploads an image without a description.

What we do about this, and when we are the wrong choice

We build websites, and accessibility is part of how we build rather than a line item you buy separately. That means WCAG 2.1 AA as the working target, contrast checked in the design phase instead of after sign-off, real text on every control, labelled form fields, and keyboard testing before launch. We do not sell an overlay and we would talk you out of one.

There are two ways in, both set out on our offer page: a flat fee for a focused build, paid once, or a monthly design and development partnership. The record behind those figures is 200+ projects shipped, a focused build that typically launches in about two weeks, and partnership requests answered inside 48 hours. Cornerstone Healing Center saw website conversions rise 20 percent after the rebuild we did for them.

Now the part most agencies leave out. We are not accessibility auditors and we do not issue conformance reports, so if you need a formal audit, a VPAT, or an expert witness, hire a specialist firm, and if a demand letter is already sitting on your desk, call a lawyer before you call us. If your site is a large enterprise application carrying a decade of legacy templates, remediation at that scale is a different discipline than ours. Where we fit is narrower and more useful than either: you are rebuilding anyway, and you want the replacement built properly the first time. That describes most of the healthcare practices and law firms who reach us.

A ten minute check you can run right now

You can find most of your problems yourself, without buying anything. Open your home page and put your mouse away. Press Tab repeatedly and watch where the focus indicator goes. If it disappears, if it never reaches the menu, or if you cannot tell what is selected, that is failure one, and it affects everyone who cannot use a mouse.

Next, zoom the browser to 200 percent and see whether text overlaps or controls fall off the screen. WCAG 2.1 sets that one at Level AA: text must resize "up to 200 percent without loss of content or functionality". Then right click, view page source, and look at the very first line for a lang attribute on the html element. Run your two main text colors through any contrast checker and compare against the 4.5 to 1 ratio that WCAG 2.1 requires for normal text. Finally, look at your search icon, your menu icon and your social icons, and ask whether any of them carries text a screen reader could actually announce.

Five minutes of that tells you more about where you stand than any scanner score, and it costs nothing. If the results bother you, the fix belongs in your next build, not in a widget.

That distinction is the entire article. Most of a medical practice website is ordinary marketing content with no patient data anywhere in it: the service pages, the clinician bios, the directions, the blog, the careers page. The compliance work concentrates in four narrow places. Knowing which four keeps you from paying for security theater on pages that never needed it, and from missing the pages that did.

The 30-second answer

If your practice is a HIPAA covered entity, your website has to protect patient information it collects, stores or transmits, and in practice that means four things: forms that collect patient details, third-party scripts that watch what visitors do, the vendors holding that data, and any logged-in patient area. Every vendor that touches that data needs a signed business associate agreement. Public pages carrying no patient data are not the problem. Separately from HIPAA, healthcare organizations receiving federal financial assistance from HHS must meet WCAG 2.1 Level AA, and that deadline moved in May 2026 to May 11, 2027 for recipients with fifteen or more employees. HIPAA penalties start at $145 per violation and are capped at $2,190,294 a year for identical violations.

First question: is your practice even covered

Not every health business is bound by HIPAA. The rules reach three groups, defined at 45 CFR 160.103: health plans, health care clearinghouses, and health care providers who transmit health information electronically in connection with a covered transaction. If you bill insurance electronically, you are in. A cash-only practice that never files a claim electronically may not be, and a wellness app with no provider relationship usually is not. Aesthetics practices sit right on this line, which is one reason a med spa build is priced differently from a salon build; we broke the numbers down in what a med spa website should cost in 2026.

Falling outside HIPAA is not the same as falling outside everything, and this is where health startups get caught. The FTC's Health Breach Notification Rule was written for exactly the businesses HIPAA misses. The Commission finalized amendments to it in April 2024, published at 89 FR 47028 on May 30, 2024 and effective July 29, 2024. Among the changes: for a breach affecting 500 or more people, the FTC has to be notified at the same time individuals are, and in no case later than 60 calendar days after discovery.

What counts as patient information on a website

Protected health information is individually identifiable health information, and the test is not whether a page mentions a medical condition. It is whether something identifying a person is tied to their health, their care, or payment for their care. A page describing your knee replacement service is not PHI. A form where a visitor types their name, phone number and "knee pain, worse at night" is.

Demographic details count once they arrive in that context, which is why an email address submitted through an appointment request sits in a different category from the same address dropped into a general newsletter box with nothing attached to it. Two consequences follow. Most of your website is genuinely out of scope. And the moment you add a form field asking why someone is getting in touch, the page holding that field changes character.

The four places compliance actually lives

Four things on a practice website carry real obligations: forms, third-party scripts, the vendors behind them, and the logged-in area if you have one. Everything else is a normal website.

Forms that collect patient details. Appointment requests, symptom questions, insurance details, file uploads. Ask what the form genuinely needs. A field you do not collect is a field nobody can leak, and most practices are collecting a clinical paragraph on a public form when a callback request would do.

Third-party scripts. Analytics, advertising pixels, chat widgets, heatmaps, embedded schedulers, review widgets. Each one is code from another company running on your page and reading what happens there. This is the category that produces almost all the enforcement, and it gets its own section below.

The vendors holding the data. Hosting, the form processor, the email inbox those submissions land in, the scheduler, the CRM. Form data emailed to a Gmail address is form data sitting in Gmail. In some specialties the website vendors price the compliant versions of those pieces as paid add-ons, which is worth knowing before you compare plans: we broke those numbers down for therapy practices.

Any logged-in patient area. Portals, results, messaging. This is the highest-risk part of the site and usually the part that is not built by your web agency at all, which is a reason to be precise about where one system ends and the other begins.

Who has to sign a business associate agreement

Any outside company that creates, receives, maintains or transmits PHI on your behalf is a business associate, and so is a data transmission service that needs routine access to it, and so are their subcontractors. That definition also sits in 45 CFR 160.103. The agreement itself is required by 45 CFR 164.502(e), which requires satisfactory assurances documented in a written contract.

The practical test is short. List every vendor that can see form submissions or portal data. For each, either a signed agreement exists or that vendor must not receive PHI. There is no third option, and "they told us they are HIPAA compliant" is not the same as a signature.

Google is the cleanest illustration because it publishes its own list. The HIPAA Included Functionality page, in the version effective May 14, 2026, names the services its business associate agreement covers: Gmail, Google Calendar, Google Chat, Google Drive with Docs, Forms, Sheets, Slides and Vids, Google Meet, Google Sites, Google Groups, Google Keep, Google Tasks, Google Vault, Google Voice for managed users, Cloud Identity Management, Apps Script, AppSheet and the Gemini apps. Google Analytics is not on that list. Neither are third-party add-ons. Google also says outright that customers without a signed agreement must not put PHI into these services at all.

The tracking script problem

Analytics and advertising pixels are where healthcare websites actually get into trouble, and most of the enforcement has come from the FTC rather than from HIPAA regulators. On July 20, 2023 the FTC and the HHS Office for Civil Rights sent a joint letter to roughly 130 hospital systems and telehealth providers, naming the Meta or Facebook pixel and Google Analytics specifically, and warning that these tools can reveal conditions, diagnoses, medications, treatments and where somebody sought care.

The money has followed the same pattern. GoodRx agreed to a $1.5 million civil penalty on February 1, 2023 in the first case ever brought under the Health Breach Notification Rule, over unauthorized disclosures to Facebook, Google and others. BetterHelp agreed on March 2, 2023 to pay $7.8 million to consumers, over revealing sensitive data to third parties such as Facebook and Snapchat after promising to keep it private, which the FTC called its first action returning money to consumers whose health data was compromised. Cerebral agreed to an order requiring it to pay more than $7 million, announced on April 15, 2024.

This is not history. On July 29, 2026, the FTC together with Utah and California sued Hims & Hers, alleging among other things that the telehealth company shared health information with Meta, Snap and other third parties while promising privacy. Those are allegations in a complaint, not findings, and the case is unresolved. It is also five days old as we publish this.

The HIPAA side of the same question has an odd twist worth knowing, because a lot of published advice is out of date on it. OCR's bulletin on online tracking technologies took the position that an IP address combined with a visit to a public page about a health condition was protected information. Hospital groups sued, and on June 20, 2024 a federal judge in the Northern District of Texas vacated that part of the revised bulletin as beyond the agency's authority. HHS withdrew its appeal on August 29, 2024.

Read that carefully before you relax. What the court removed was a specific theory about unauthenticated public pages. Everything behind a login is untouched by the ruling and squarely covered. And nothing in that case binds the FTC, which is the agency actually collecting money. The workable rule for a practice site: keep advertising pixels off pages tied to specific conditions and off the appointment flow, never put anything identifying into a URL or query string, and treat any vendor receiving that traffic as one that needs a signed agreement or needs removing.

What it costs when this goes wrong

HIPAA penalties are tiered by what you knew and what you did about it. The statutory tiers live at 45 CFR 160.404, and the amounts are inflation-adjusted every year in a table at 45 CFR 102.3. The current published figures run from $145 per violation where you genuinely did not know, to $73,011 per violation for willful neglect left uncorrected, with a cap of $2,190,294 for identical violations in a calendar year.

Per-violation is the phrase that should hold your attention. The tiers are a range, not a forecast, but a single misconfigured script does not misfire once. It misfires for every visitor, every day, until somebody notices.

The Security Rule bits a website actually touches

The technical safeguards are shorter than most checklists suggest, and one detail surprises nearly everyone. Under 45 CFR 164.312, unique user identification and an emergency access procedure are required. Automatic logoff and encryption are listed as addressable. So is encryption in transit, under the transmission security standard. The rest of the section covers audit controls, integrity of the data, and verifying that whoever is asking for access is who they claim to be.

Addressable does not mean optional. It means you either implement it or document why an alternative is reasonable for your setup, and the documenting is the part that gets skipped. In 2026 there is no defensible reason to skip encryption on a website anyway, so treat this as a paperwork question rather than an engineering one.

HHS proposed rewriting the whole Security Rule in a notice published at 90 FR 898 on January 6, 2025, which would tighten many of those addressable items into requirements. As of August 2026 it is still a proposal. Build as though it will land, because the things it proposes are things a competent build does anyway.

The healthcare accessibility deadline that just moved

This one is a different law with a different deadline, and it moved three months ago. On May 9, 2024, HHS published a final rule under section 504 of the Rehabilitation Act, at 89 FR 40066, requiring recipients of federal financial assistance from HHS to make their web content and mobile apps conform to WCAG 2.1 Level A and AA. Then, on May 11, 2026, the department published an interim final rule at 91 FR 25496 extending both compliance dates by a year.

Two things to check before you act on that. Whether your organization is a recipient of HHS financial assistance is a genuine question with a real answer, and it is not the same question as whether you are a HIPAA covered entity. The two sets overlap without matching. Second, the rule at 45 CFR 84.85 carries exceptions, including archived web content, conventional electronic documents that predate your compliance date and are not used to access your services, content posted by unrelated third parties, and password-protected documents about a specific individual.

Our advice is unchanged by the extension. If you are rebuilding anyway, build to WCAG 2.1 AA now, because doing it during a build costs close to nothing and retrofitting a finished site costs real money. The wider picture, including what the ADA asks of every business open to the public, is in our guide to ADA website compliance. We wrote the design side of this in our piece on healthcare website design and trust.

What to ask the agency building your site

Six questions separate an agency that has done this before from one that will learn on your project.

Which vendors will touch form submissions, and which of them will sign a business associate agreement? Which third-party scripts load on condition pages and on the appointment flow, and can you show me the list? Where do form submissions actually go, and are they emailed anywhere in plain text? Who holds admin access after launch, and what is the process for removing someone? Will you build to WCAG 2.1 AA, and how will you show me you did? And if we part ways, who keeps the site, the domain, the content and the analytics history?

An agency that answers "we use HIPAA compliant hosting" and stops has answered roughly one sixth of that. Hosting is the easiest part.

What we do here, and who we are wrong for

We build healthcare websites, and we are not a compliance firm. That line matters because the two get sold as one thing. Our work is the site itself: the pages, the forms, the speed, the structure, the script inventory, and picking vendors that will sign when they need to. We do not write your policies, run your risk analysis, or give you a legal opinion, and any agency offering all of that with a website is selling something outside its competence.

You can see our numbers without booking anything. A focused build is charged once. Continuing work runs monthly. Both are on the pricing page. Behind those numbers sit 200+ shipped projects, a typical two week turnaround on a focused build, and a 48 hour answer to any partnership request. On results, Cornerstone Healing Center's website conversions rose 20 percent after the redesign we did for them, which is the closest proof we have of a clinical audience making a hard decision quickly.

Where we are the wrong call: if you need a compliance program rather than a website, hire a firm that does that. If you are a health system with enterprise procurement and a security questionnaire, we are too small. If your project is really custom portal software integrated with an EHR, that is a different discipline. Our healthcare page covers what we do build, and the practices we do it best for are ones whose site is quietly losing people who were ready to call.

An audit you can run this afternoon

You do not need anyone's help to find out where you stand. Open your site in a private browser window with the developer tools network tab showing, click through to a condition page and start an appointment request, and write down every outside domain that loads. That list is your third-party script inventory, and for most practices it is longer than expected.

Then list every form on the site and where each submission lands. List every vendor from those two exercises and mark which ones have a signed agreement on file. Tab through the site with the keyboard only, and run one page through a contrast checker. Finally, write down who currently has admin access, including former staff and former agencies.

Five lists, one afternoon, no invoice. If the finished lists make you uncomfortable, that discomfort is the finding, and it is a far cheaper way to learn it than the alternatives described further up this page.

Most guides answer this by inventing a range. We are going to do something narrower. Every figure below is one that a named company publishes about itself, on its own pricing page or its Clutch profile, read in August 2026. Where nobody publishes anything, the table says so, because a blank is information too.

The 30-second answer

A US law firm should expect a website build to land somewhere between roughly $3,000 and $25,000 in 2026, with enterprise work running past $100,000. What moves you between those bands is the volume of original content and custom design the site needs, not how good the finished site is. The seven legal web agencies below publish minimum project sizes of $1,000, $5,000 or $10,000 before anyone scopes your work, and published hourly rates that all sit between $100 and $199. After launch, budget separately for hosting, content, accessibility work and, in some states, bar advertising filings. Our own offer is set out for the same reason this article is: a focused build, flat and one time, and an ongoing monthly partnership.

What agencies actually publish

Seven agencies that build or market law firm websites publish enough about themselves to compare. Here they are, oldest first, which is the only ordering that does not smuggle in an opinion. Everything in the table comes from the firm's Clutch profile, read in August 2026.

Two things jump out. The legal-only shops are not the expensive ones, and DD.NYC, which puts only 10 percent of its work in legal, sits at the top of the published range alongside MeanPug at $10,000. Specialization and price are not the same axis, which is worth remembering when a pitch implies they are.

What a published minimum actually tells you

A minimum project size tells you who an agency wants on the phone, and almost nothing about what your project will cost. It is a lead filter, published by the agency itself and audited by nobody.

The clearest evidence is MeanPug Digital. Its published minimum is $10,000, and its own profile reports that its most common project size falls between $50,000 and $199,999. The floor and the typical engagement are a factor of five apart at the low end. On the other side, BluShark Digital and LawRank both publish $1,000 floors while listing headcounts of forty to two hundred plus people. Nobody staffs fifty people on $1,000 projects.

Use the minimum for one thing only: to work out whether you are the size of client this agency is built to serve. If your whole budget equals their floor, you will be the smallest account in the building.

Reading an hourly rate

Published hourly rates in this market are remarkably tight. Six of the seven publish one, and every band sits between $100 and $199 an hour.

Because the rates barely vary, the rate is not what makes one quote three times another. The hour count is. A quote of $6,000 at $150 an hour is forty hours of work: roughly a template configured, your content poured in, and a launch. A quote of $30,000 at the same rate is two hundred hours, which buys original design, photography direction, a page written for each matter type and a real migration plan.

So ask for the hours, not just the total. An agency that cannot tell you roughly how many hours its own proposal contains has not scoped the work, and the number you were given is a guess wearing a suit.

The one firm that publishes real build prices

PaperStreet publishes four price bands for law firm websites on its own site, which is rare enough to be worth studying even if you never contact them.

Their Essentials tier, $3,000 to $5,000, is described as pre-designed templates that get modified to fit, with $3,000 quoted as a total first-year cost. Plus runs $9,500 to $15,000 and can go higher when there is an extensive existing website to account for. Custom is $18,000 to $25,000, and they name the variables that push it up: the number of attorneys, functionality requirements, search features, branding needs and blog transfers. Enterprise starts at $40,000.

Notice that the drivers they publish are all quantities. More attorneys, more pages, more content to move. Price in this market is mostly a headcount and a page count, dressed up as a design decision.

What moves your number

Five things account for most of the gap between the cheapest sensible quote and the most expensive one.

The number of matter types you take. A firm doing personal injury, family law and estate planning needs three genuinely different pages, because someone searching after a car accident and someone searching about a custody schedule want nothing in common. Each real page is content, not a template swap.

Attorney bios and photography. Bios are the most read pages on most firm sites and the most often left as placeholder text. Real photography of real people costs money, and pound for pound it changes more minds than anything else on the invoice.

How much writing is included. The single largest hidden variable in every proposal. Ask flatly: who writes the practice area pages, and how many are included at this price? "Client provides content" is the phrase that turns a six week project into a nine month one.

Whether an existing site has to move. If you have years of blog posts and rankings, a migration with a URL map is real work. Skipping it is how firms lose traffic they had already paid for, which we broke down in our piece on traffic that never becomes leads.

Intake plumbing. Forms that route to whoever answers new matters, call tracking, and a connection to your case management system. This is where a website stops being a brochure, and it is usually quoted separately.

The line items only law firms pay

Two cost centers show up on a law firm website that never appear on a restaurant's, and most web proposals are silent on both.

The first is bar advertising regulation. Rules differ by state and change, so your own jurisdiction is the one that matters, but Texas shows the shape of the problem. Under Texas Disciplinary Rule 7.04, unless an exemption applies a lawyer must file with the Advertising Review Committee of the State Bar of Texas no later than ten days after an advertisement is disseminated, including a copy of the advertisement, a completed application and a fee set by the Board of Directors. The same rule lets you submit at least thirty days ahead for a pre-approval that protects you if the final version matches what was reviewed. Neither the filing time nor the review cycle is in your web quote, and both sit on your calendar, not the agency's.

The second is accessibility. The ADA lists "office of an accountant or lawyer" among public accommodations in 42 U.S.C. 12181(7)(F), and the Department of Justice, in its web accessibility guidance published on March 18, 2022, states its consistent position that the ADA's requirements apply to all the goods, services, privileges or activities offered by public accommodations, including those offered on the web, and points to the Web Content Accessibility Guidelines as the technical reference. Building to WCAG from the start costs very little. Retrofitting a finished site costs a lot, which is the actual argument for raising it in the first meeting.

You are the lawyer here and we are not, so treat both paragraphs as a prompt to check your own state's rules rather than as advice about them. Other regulated professions carry their own version of this problem, and we worked through what a financial advisor website costs under the SEC and FINRA rules that shape one.

What you keep paying after launch

The build price is the smaller half of a five year number. Three lines continue whatever you signed.

Hosting is the cheapest and the easiest to compare, because it is genuinely published. WP Engine lists managed WordPress hosting from $30 a month on its Startup plan up to $276 a month on Scale, on annual pricing. Anyone quoting you many multiples of that for "hosting" is bundling something else, and you are entitled to ask what.

Content is the line that decides whether the site keeps earning. A site with eight pages written once will do what eight pages do, forever. Publishing is the difference between a brochure and an asset.

Maintenance means plugin and platform updates, security patching and the small fixes nobody plans. Ask what is covered and what is billed, in writing, and ask the ownership question early: if you stop paying, do you keep the site, the domain, the content and the analytics history? Ownership at the point of a breakup is the most expensive detail people skip.

The technical floor to demand at any price

Price does not guarantee speed, and slow sites lose people who were ready to call. Google's Core Web Vitals give you a floor you can hold any agency to: the page's main content should paint inside 2.5 seconds, responses to taps should land within 200 milliseconds, and layout shift should stay at or under 0.1. Google says to judge all three at the 75th percentile of real page loads, measured separately for mobile and desktop.

Two ways to use that before you sign. Run a page speed test on a site the agency actually built, on mobile, and ask them to explain any failure. Then put the thresholds in the contract as a launch condition. Agencies that build fast sites will agree in the meeting. The rest will explain why the numbers do not matter, which is your answer.

When the client asks an AI before they ask you

Search behavior in legal has moved, and the site has to answer for it. Clio's 2025 Legal Trends Report, published on October 16, 2025, reports that a majority of consumers say they would look for their next lawyer online, that more than half have used or would consider using AI to answer a legal question, and that among those who used AI, 28 percent were directed to contact a lawyer.

That last number is the one to build for. People are arriving at your site already briefed, often by a machine that read someone's page and not yours. Pages that answer real questions in plain language get quoted by those systems; pages of adjectives about dedication and results do not. We wrote the practical version of this in how to get recommended by ChatGPT and Google AI.

What we offer, and who should not hire us

Our offer is on the offer page: a flat fee for a focused build, one time, or an ongoing monthly partnership, each quoted once we have scoped the work. We have shipped 200+ projects, a focused build typically goes live in about two weeks, and we answer partnership requests within 48 hours. The closest proof we can show a law firm is from healthcare rather than legal: after we redesigned Cornerstone Healing Center's site, they saw 20 percent more website conversions. The comparison holds because the decision rhymes: a stranger in a bad month, deciding in a few minutes whether to trust you with the worst thing happening to them.

Who should not hire us. If your bottleneck is paid media, intake staffing and case management operations rather than the website, the legal-only agencies named on this page do that work and we do not. If you want a hundred city landing pages spun up for a personal injury market, that is a different machine and a different budget. And if the cheapest possible template is the goal, someone will beat our number.

Hire us when the site is what is failing. It loads slowly on a phone, it reads like the firm two doors down, nothing on it speaks to the exact case a person is searching at eleven at night, and the visitors you already pay for leave without calling. Fixing that is the whole job, and our law firm page lays out how it gets built.

A two-week way to decide

Week one, do the homework nobody does. Write down every matter type you actually want more of, in priority order, because that list is the site map and the quote. Get three proposals from agencies in different categories, one legal-only, one design-led. Ask each for an hour estimate alongside the price.

Week two, test rather than deliberate. Run the speed check on their client sites. Ask the ownership questions. Ask who writes the practice area pages and how many are included. Then buy the smallest real scope any of them will sell you, a homepage and one matter page, instead of signing a twelve month agreement with someone you have never worked with.

The firms that get this wrong are rarely the ones who picked the wrong agency. They are the ones who bought a site, never published again, and replaced the whole thing four years later at full price.

This page is published by an agency too. We build websites for treatment centers, so we are one of the options here and we have an obvious interest in how you read it. What we can do is take the opinion out. Everything below comes from what each firm publishes about itself, on its own website or on its Clutch profile, checked in August 2026, plus a set of tests you can run on any agency in an afternoon without asking anyone's permission.

The 30-second answer

There is no single best web design agency for addiction treatment centers, and any list that crowns one is selling the crown. The market splits three ways. Behavioral health specialists such as Webserv, Lead to Recovery and Behavioral Health Partners work only in this vertical. Healthcare generalists such as Dreamscape Marketing, Stodzy Internet Marketing, Active Marketing and Onspire Health Marketing bring larger teams and a wider healthcare record. Design-led partners, which is what KhanWork is, build the site and hand it over. Pick the category that matches the gap you actually have, then test the finalists. Star ratings and years in business are the weakest signals on the table.

How this list was built

Three rules, all of them boring on purpose.

First, a firm had to have a public services page covering website design or development for addiction treatment or behavioral health providers. Second, it had to publish enough about itself to compare, meaning at minimum a location or a founding date or a public directory profile. Third, nothing on this page is our characterization of anyone's quality. We have not seen their contracts, their client rosters or their internal results, and neither have you.

Two things that follow from that are worth saying out loud. Clutch profile data is entered by the agency, not audited by Clutch, so a published minimum project size is a marketing choice rather than a fact about what your project will cost. And a firm being absent from this page means nothing at all. There are dozens of competent shops in this vertical, including several that appear on the aggregated directories but publish too little to compare fairly.

The seven firms, and what each one publishes

Listed oldest first by the date each firm publishes for itself. That order is not a ranking, it is just the only neutral order available.

Onspire Health Marketing

Onspire publishes a Kansas City, Missouri headquarters and the line "trusted healthcare marketing since 1992". Behavioral health sits inside a much broader practice that covers hospitals and health systems, academic medical centers, rural and critical access hospitals, health plans, medical device companies and a long list of clinical specialties. Services run from brand strategy and positioning through website design, content, SEO and reputation management.

Best fit when your center is part of a larger health system, or when the real problem is positioning and brand architecture rather than the website itself. Look elsewhere if you want a partner who thinks about admissions calls all day. Behavioral health is one of many verticals here.

Active Marketing

Active Marketing publishes the sentence "Active Marketing has specialized in growing admissions at addiction treatment centers since 2004", which is the longest addiction-specific tenure claim we found. The published service mix leans hard toward content, digital PR, link building and technical SEO, with PPC, CRO, email and reputation work alongside. Healthcare B2B and SaaS appear in the navigation as separate practices. No office address is published anywhere on the site.

Best fit when your gap is organic visibility and content depth over a long horizon. Look elsewhere if you need a website rebuilt as the primary deliverable, or if a published office address is part of your vendor diligence.

Dreamscape Marketing

Dreamscape publishes an address in Columbia, Maryland and a team of "70+ Digital Experts on 8 Highly Specialized Teams", extended through a partner network it describes as "350+ Industry Experts in Managed Care, Marketing, and Technology". Its Clutch profile says founded in 2005, 50 to 249 employees, $100 to $149 per hour, minimum project size $1,000, and a 5.0 rating from five reviews. Addiction treatment is one of nine published verticals that also include autism services, dental, elective medical, hospitals, mental health, non-profits, senior living and small business.

Best fit when you are a multi-location operator who needs media buying, call tracking, CRM integration and a website from one vendor with the staff to cover all of it. Look elsewhere if you want the people who built your site to still be reachable in month nine, which is harder inside a large agency.

Stodzy Internet Marketing

Stodzy publishes an address at 1700 North Dixie Highway in Boca Raton, Florida and the claim "14 Years Of Healthcare Marketing Experience". The published service list is unusually long and includes one item nobody else on this page lists: LegitScript certification support, which matters because that certification is the gate on Google advertising for addiction services. Published results claims include "1000+ Calls Per Day Across All Clients". Addiction treatment sits alongside dentistry, primary care, telehealth, ABA therapy and medispas.

Best fit when you want search and paid media handled by a team that has already walked facilities through certification. Look elsewhere if design is the thing you are unhappy about. This is a search-first shop by its own description.

Webserv

Webserv publishes an Irvine, California address and says it has supported addiction and mental health facilities since 2017 and does not serve other industries. Published numbers include "15,000+ Admissions to Facilities", "200+ Treatment centers nationwide" and "$100M+ Ad Spend Managed". Its Clutch profile lists 10 to 49 employees, $150 to $199 per hour, a $5,000 minimum project size and a 5.0 rating from 16 reviews, the largest verified review count of any firm on this page. Web design sits under a "performance creative" practice, next to admissions operations work such as lead management and eligibility.

Best fit when paid media is a large part of your admissions engine and you want the site, the ads and the intake operation under one roof. Look elsewhere if your budget cannot support managed media, since the website is positioned as one input into that machine rather than the product.

Lead to Recovery

Lead to Recovery publishes an address in Pompano Beach, Florida and describes itself as an addiction treatment website design agency first. Its Clutch profile lists founded 2019, 10 to 49 employees, $100 to $149 per hour, a $1,000 minimum and no reviews yet. The published service mix is broad for a firm this size and reaches past digital into television, radio, print and direct mail.

Best fit when you want offline and online demand handled together, which is rarer than it sounds in this vertical. Look elsewhere if third-party verified reviews are a hard requirement for your board, because there are none published yet.

Behavioral Health Partners

Behavioral Health Partners publishes a Denver, Colorado address and a service list that goes past marketing into treatment consulting: licensing, accreditation support and feasibility studies for people opening centers and sober living homes. Its Clutch profile lists founded 2024, 10 to 49 employees, $200 to $300 per hour, a $10,000 minimum and a 5.0 rating from one review. Both the rate and the minimum are the highest published on this page.

Best fit when you are opening or expanding and want the licensing conversation and the marketing conversation in the same room. Look elsewhere if you need a long verifiable track record, since the firm publishes a 2024 founding date.

Side by side

Everything in this table is published by the firm on its own site or its Clutch profile, checked August 2026. Blank means the firm publishes nothing on that point, which is itself worth knowing.

What tenure tells you, and what it does not

Tenure tells you a firm has survived, which in a vertical this volatile is not nothing. It does not tell you the site they will build you is fast, that the person who sold you will touch the work, or that the template they used in 2019 has been rethought since.

Read that chart against the pricing one below and the ranking logic falls apart. The firm with the longest published history is a broad healthcare agency where behavioral health is one practice among many. The firm founded most recently publishes the highest hourly rate on the page. Neither of those facts predicts whether your admissions line rings.

A better question than "how long have you been doing this" is "show me a treatment center site you launched in the last twelve months". Recent work is testable. Tenure is not.

What they publish about money

Three of the seven publish no price signal anywhere. That is normal in this vertical, and it is the single biggest reason buyers end up comparing quotes that are ten times apart.

Treat published minimums with care. Dreamscape lists a $1,000 minimum project size next to a 50 to 249 person headcount, which tells you the field is a directory setting rather than a quote. The useful signal in that data is the hourly band, because it maps to seniority and overhead more honestly than a minimum does.

Ours is set out on our offer page: a flat one time build, or an ongoing monthly partnership. Our build fee is a package price for a focused scope, not a from-price that grows during the call. For the wider market ranges, including the compliance line items nobody quotes, we broke the numbers down in what a treatment center website actually costs in 2026.

Specialist or generalist

The specialist argument is real. A firm that only works with treatment centers already knows what an admissions funnel looks like at 2am, how families read a staff page, and which claims will get an ad account suspended. You will not spend the first month explaining your business.

The generalist argument is also real, and it gets dismissed too easily. Vertical specialists tend to converge on a house template, which is why so many treatment center websites look like each other: same hero, same stock photo of a hand on a shoulder, same six trust badges. If everyone in your market hired from the same short list of specialists, the fastest way to look different is to hire outside it.

The honest split we use: hire a specialist when the bottleneck is media buying, compliance or admissions operations. Hire a design-led partner when the bottleneck is that your site does not convince anyone. If you are not sure which it is, our addiction recovery page lays out the four things families check before they call.

Nine questions that separate them

Ask all nine on the first call. The answers take an hour and save a year.

1. Show me a center site you launched in the last twelve months. Not a portfolio grid, a live URL you shipped recently.
2. Who owns the domain, the code, the analytics property and the ad account? The correct answer is your legal entity, on every one of them.
3. What happens to the site if we stop paying monthly? If the answer is that it goes away, you are renting, and that should be priced as rent.
4. Do you also sell leads, run a directory, or represent other centers in my market? Any of those can be fine, but you want it on the record before you sign, not after.
5. Who writes the clinical copy, and who reviews it? Someone at your facility has to own accuracy about levels of care.
6. What tracking goes on our condition pages, and who signed off on it? See the test section below.
7. Can my admissions team edit a page without a ticket? If not, every typo becomes a two-day request.
8. What will you report on, leads or admissions? Lead counts are easy to inflate. Admissions are not.
9. Can I speak to a client who left you? The reaction to this question is more informative than the referral.

The test you can run before you sign

You do not need a technical background for any of this, and it takes about twenty minutes per finalist.

Run their own site, then three client sites, through PageSpeed Insights. Google publishes the thresholds it uses: a Largest Contentful Paint of 2.5 seconds or less, an Interaction to Next Paint of 200 milliseconds or less and a Cumulative Layout Shift of 0.1 or less, all measured at the 75th percentile of real page loads. Passing is not a high bar and most of the web still fails it. The HTTP Archive's 2024 Web Almanac found 43% of mobile websites had good Core Web Vitals. If an agency sells speed and its own site fails, you have learned something free.

Ask what goes on the condition pages. Marketing pixels on pages about detox, dual diagnosis or specific substances have been the noisiest compliance topic in this field, and the picture changed in 2024. On June 20, 2024, a federal judge in the Northern District of Texas vacated the part of the HHS Office for Civil Rights online tracking bulletin that treated an IP address on an unauthenticated public health page as protected health information, holding the agency had acted beyond its authority under HIPAA, and HHS later withdrew its appeal. That does not make tracking a free-for-all, and 42 CFR Part 2 still governs substance use disorder patient records separately. It does mean the decision belongs to you and your counsel rather than to a vendor's default settings. Any agency that cannot tell you exactly what it installs, and why, is answering the wrong question.

Check the domain registration. A public WHOIS lookup on your own domain takes thirty seconds. If your previous agency is the registrant, fix that before you start talking to anyone new.

Four things that should end the conversation

A guaranteed number of admissions. Nobody can promise that, and the promise usually comes attached to a compensation structure worth reading closely, which we covered in how centers fill beds without buying leads.

A refusal to name the platform your site will be built on. You are entitled to know what you will be maintaining.

A proposal with no page-level scope. "A 12-page website" is not a scope. Which twelve, saying what, ranked for what.

A portfolio you cannot click. Screenshots are not evidence. Live URLs are.

Where we fit, and who should not hire us

KhanWork is a design and development partner. We have shipped 200+ projects across the US and Europe, we answer partnership requests within 48 hours, and a focused build typically launches in about two weeks. Cornerstone Healing Center saw 20% more website conversions once their redesign went live. We publish our prices, we build the site to be fast, and the domain, the code and the analytics are yours from day one.

Here is who should not hire us. If your main need is a team to place and manage a large paid media budget every month, hire one of the specialists above, because that is their craft and not ours. If you want a vendor to sit in your weekly census meeting and own admissions operations, same answer. If you want the cheapest possible template, there are shops that will beat us and you will get what you paid for. We are the right call when the website itself is the problem: it is slow, it does not look like anywhere a family would trust their son, and it does not turn the traffic you already have into calls.

How to run this in two weeks

Week one: pick three firms across two categories, one specialist and one design-led at minimum. Run the speed test and the domain check before the calls so you arrive with evidence. Ask the nine questions and write down who dodged which.

Week two: ask each finalist for a small paid first scope rather than a twelve month contract. A homepage, a single level-of-care page, or a paid discovery. You will learn more about how a firm works from three weeks of real work than from any pitch deck, and the cost of being wrong stays small.

Then pick, and give whoever you pick a fair run. The most expensive pattern in this vertical is not choosing the wrong agency. It is changing agency every nine months and starting the compounding over each time.

The lead vendor model is easy to understand and hard to leave. A call center or a directory ranks for the searches you cannot win, screens the caller, and hands you a warm phone call for a fee. You pay per call or per admission, you can turn the tap up when census drops, and you never have to build anything. That is the pitch, and for a while it holds.

What changed is the exposure. In 2025 the Federal Trade Commission stopped treating deceptive treatment advertising as the marketing company's problem alone and named the clinics that received the steered calls. Federal kickback law reaches people paid per referral, and it is criminal law, not a licensing matter. Meanwhile the certification you need to advertise at all became a fixed annual cost per facility that no vendor relationship removes.

So this is a piece about the other path: the channels you own, what the law actually says, what compliance costs, and the order to build in. One caveat first. We could not verify a credible public benchmark for what aggregator calls cost per admission. The ranges circulating on agency blogs do not show their data, so we are not going to repeat them. Every figure below links to a source you can open yourself.

The 30-second answer

You can fill beds without buying leads, and 2026 is a good year to start. Paying anyone per referral runs straight at the Eliminating Kickbacks in Recovery Act, which carries a fine of up to $200,000 and up to 10 years in prison for each occurrence. The FTC now names treatment providers, not only their marketing vendors, in deceptive advertising complaints. And the replacement channels are cheaper to hold: a free FindTreatment.gov listing, a site that answers the four questions every caller has, local and organic search, and referral relationships structured so nobody is paid by the head. Build them in that order, then measure admissions by source rather than leads by source.

What the law says about paying per referral

Federal criminal law prohibits paying or receiving anything of value in exchange for referring a patient to a recovery home or a clinical treatment facility. That is the Eliminating Kickbacks in Recovery Act, 18 U.S.C. 220, passed in 2018. It applies to any health care benefit program, which means commercial insurance and cash pay, not only Medicare and Medicaid. The penalty is a fine of not more than $200,000, imprisonment of not more than 10 years, or both, for each occurrence.

The part that matters for marketing sits in the exceptions. Payments to an employee or an independent contractor are permitted, but only where the compensation is not determined by and does not vary by the number of individuals referred, the number of tests or procedures performed, or the amount billed to the health care benefit program. Read that twice, because it describes almost every business development pay plan in this industry. A salaried outreach coordinator sits inside the exception. The same person on a bonus per admission sits outside it.

In July 2025 the Ninth Circuit gave the statute its first appellate reading in United States v. Schena. The court held that percentage and commission based compensation for marketers is not automatically a violation, and that the statute is not a blanket prohibition on those structures. That was widely reported as good news. It is narrower than it sounds. Falling outside a safe harbor is not the same as being lawful, and the court's point was that such arrangements become unlawful when paired with undue influence over the referral, such as false or misleading statements about the service. If a vendor's call script is doing the persuading, you are relying on somebody else's honesty to stay on the right side of that line.

State law can be stricter, and often is. Florida's Patient Brokering Act, Florida Statute 817.505, makes it a felony to offer or pay a commission, bonus, rebate, kickback or bribe to induce a patient referral. The fine scales with the number of patients involved: $50,000 for fewer than 10, $100,000 for 10 to 19, and $500,000 for 20 or more, with a felony conviction attached at every tier.

We build websites, we are not lawyers, and none of this is legal advice. If you currently pay anyone per call, per lead or per admission, the useful next step is an hour with a healthcare attorney who works in behavioral health, before your next renewal.

Your vendor's ads are your problem now

The FTC has started naming treatment providers as defendants over advertising their marketing partners ran. That single change is why this article exists.

On June 24, 2025 the Commission filed a complaint in the U.S. District Court for the District of Maryland against Mercury Marketing, LLC and others, alleging they used Google search ads displaying the names of specific treatment clinics while routing the resulting calls to a defendant call center, where telemarketers posed as the searched-for clinic or as a centralized admissions office. The FTC alleged the telemarketers claimed clinical professionals had recommended particular facilities after an objective assessment, when in reality they were working on behalf of those facilities. The defendants named include Malibu Detox, LLC, Malibu Recovery Center, LLC and Aliya Health Group, LLC: the providers on the receiving end, listed alongside the marketing company and four individuals. The Bureau of Consumer Protection's director said the defendants "took advantage of consumers searching online for substance use disorder treatment". The case is a complaint, so the allegations are untested.

Two weeks earlier the Commission had settled with a Florida provider over the same tactic. In the settlement announced with Evoke Wellness, the company ran Google ads keyed to rival clinics' names paired with its own call center number, and its telemarketers posed as a centralized admissions office or an addiction hotline. The order imposes a $7 million civil penalty, suspended to $1.9 million because the defendants could not pay the full amount, bans them from using rivals' names in search-engine ads, and requires a compliance program that monitors their call centers for misrepresentations.

Both actions were brought under the FTC Act and the Opioid Addiction Recovery Fraud Prevention Act of 2018, which the Commission says gives it enhanced remedies including civil penalties and consumer refunds specifically for deceptive marketing of substance use disorder treatment. Its advice to businesses in this sector is blunt and free: search your own name and see who is bidding on it.

The practical read for an operator is not that paid search is forbidden. It is that when you buy calls, you inherit the ad copy, the script and the impersonation risk of a company you do not control, and the Commission has shown it will put your legal entity in the caption of the complaint.

What it costs to advertise legitimately

Advertising addiction treatment on the major platforms requires certification, and it is a recurring per-facility cost. Google restricts the promotion of recovery-oriented drug and alcohol addiction services, and only advertisers that have applied and been approved may run them, in a short list of countries: Australia, Canada, France, Ireland, New Zealand and the United States. Services unrelated to drug and alcohol addiction, such as behavioral addiction or nicotine addiction, sit outside that policy.

The approval most US centers need comes through LegitScript, whose addiction treatment certification is recognized by Google, Meta, Microsoft and Nextdoor. Published standard pricing is an application fee of $1,395 to $1,595 per facility and an annual certification fee of $2,550 to $3,095 per facility, with expedited review adding $2,500 per application. The certification verifies 16 core standards covering licensing, legal compliance, staff qualifications, insurance, accurate website content, privacy and transparent advertising.

Here is the part nobody selling you leads will mention: lead generators and recovery support services can be certified under the same program. Certification makes you eligible to compete, it does not make you win. A multi-location operator pays it per facility, every year, and still has to earn the click.

The demand you are actually competing for

Your census problem is a findability problem, not a demand problem. SAMHSA's 2024 National Survey on Drug Use and Health, as summarized by the National Association of Counties, counted 52.6 million people aged 12 or older, 18.2% of that population, who needed substance use treatment in the past year. It counted 10.2 million, or 3.5%, who received it.

Five people needed care for every one who got it. That gap has causes no website can fix: cost, waitlists, stigma, geography, ambivalence. What it does tell you is that no vendor is selling access to a scarce resource. They are selling a shortcut to attention, at a price that resets every month, in a market that is not short of people.

Start with the listing you are probably not using

The cheapest acquisition channel in this industry is a federal directory that costs nothing to join. FindTreatment.gov, SAMHSA's treatment locator, drew over 5.1 million views in 2025. It is also where the FTC sends consumers in its alerts about search-ad scams, which makes it a rare listing actively promoted by the government to exactly your audience.

Getting listed runs through the National Substance Use and Mental Health Services Survey. Facilities must be licensed, certified or otherwise approved by their state substance abuse agency to be eligible, and the survey has to be completed every year to keep the listing live. It takes an afternoon. We have sat with clinical directors who did not know their listing had lapsed, in the same quarter they approved a five-figure lead spend.

Do the same pass on your state's provider locator, your accrediting body's directory, and the in-network directories of every insurer you take. None of these will fill a wing on their own. They are free, they are permanent, and they rank.

Make the site answer the four questions

A treatment center website has one job: answer the four questions a person has at 11pm, before they will pick up the phone. Will my insurance cover this. What happens in the first 24 hours. Who will I actually be talking to. Can I reach a human right now.

Most sites in this category answer none of them above the fold. They lead with a photograph of a beach, a mission statement, and a phone number in small grey type. And the caller is frequently not the patient. It is a mother, a spouse, an adult child who has been carrying this for months and is finally doing something about it at midnight. Write for that person and the question of tone answers itself.

The mechanics that move the number, in the order we build them:

• Insurance, stated plainly and early. Name the carriers you work with. If you verify benefits, say how long it takes and what you need from the caller. Vagueness here reads as expensive.
• One primary action per page. A phone number that is a real tap target on mobile, plus a short form for people who will not call. Not four competing buttons.
• Program pages that describe the day. Levels of care, length of stay, what family involvement looks like, what the room looks like. Photographs of the actual facility, not a stock library.
• Proof that survives scrutiny. Licenses, accreditations and staff credentials with names and faces. Only what you can evidence.
• Speed, and mobile first. This traffic is overwhelmingly phone traffic, often on a poor connection, often late at night.

After we redesigned the site for Cornerstone Healing Center, website conversions rose 20%. That is one client and one project, so treat it as a data point and not a forecast. If you want the full costing of a build like that, we published a full cost breakdown for treatment center websites. For the layout detail, our guide to recovery website design goes deeper.

Own the searches you are currently renting

Every dollar paid to an aggregator rents a position you could hold. Three assets do most of that work.

Your own name. Search it. If a competitor or a call center is bidding on your brand, you are paying a middleman for demand that already asked for you by name. Defensive brand bidding is cheap, because your quality score on your own name is excellent, and the FTC order against Evoke Wellness shows how long the alternative remedy takes.

Local presence. A complete Google Business Profile for every location, consistent name, address and phone across directories, and real photographs. Someone searching for treatment near a specific city is further down the funnel than any national keyword you could buy.

The questions that come before the decision. Nobody starts at "rehab near me". They start at whether insurance covers residential treatment, what withdrawal actually feels like, how to tell an employer, whether they can keep their phone. Write those pages honestly, without a sales turn at the end, and they compound. They are also what AI assistants quote back, which is worth building for deliberately. We wrote up how to get recommended by ChatGPT and Google AI separately.

Referral relationships, structured so they stay legal

Referrals are the highest-converting channel in behavioral health and the easiest one to structure illegally. The compliant version is simple: relationships built by salaried people whose pay does not vary with the number of individuals they refer, per the federal exception above.

The relationships worth the calendar time are discharge planners at local hospitals, emergency departments, therapists and psychiatrists in private practice, employee assistance programs, drug courts and probation officers, sober living operators, and your own alumni and their families. None of them need to be paid. They need to know what you actually treat, what you do not, how fast you can accept someone, and that you will call back the same day.

Alumni deserve their own line. A center with a real alumni program has a referral engine no lead vendor can sell, because the person recommending you has nothing to gain. Ask for the review while the gratitude is fresh, make it effortless, and never offer anything of value in exchange for one.

Measure admissions, not leads

Lead counts are the metric a vendor prefers, because they are the metric a vendor controls. Track two things instead. First, cost per admission by channel, calculated across a full quarter so one long stay does not distort it. Second, a single question at intake: how did you first hear about us, asked before anyone mentions a website.

Those two numbers usually tell an uncomfortable story in the first quarter and a clarifying one by the second. Owned channels look terrible in month one because they carry no volume yet, then quietly overtake, because their cost stays flat while their output rises.

What this costs and how long it takes

Launch Partner is a one time build, and Growth Partner is a monthly partnership covering ongoing design and development. We scope and quote both on a call. Fourteen business days is the typical launch window for a focused build, counted from the day content and brand assets land, though a multi-location treatment center with program pages, staff bios and insurance content usually runs longer, because content approvals are the slow part and never the code. We have shipped 200+ projects, and we reply to partnership requests within 48 hours. The full detail is on our pricing page, and our addiction recovery page explains how we approach this sector.

Set that against the annual floor you already pay: certification per facility, plus whatever your current vendor bills. The comparison that matters is not agency cost against zero. It is cost per admission from an asset you own against cost per admission from an asset you rent.

Who should keep buying leads

We would rather say this than take on a client we cannot help. Keep your vendor, at least for now, if any of these describe you.

• You have weeks of runway and need heads in beds this month. Owned channels do not work on that timescale, and anyone who tells you otherwise is not being straight with you.
• You are not licensed, certified or otherwise approved by your state authority. Fix that first. It gates the free directories and the ad certification both.
• Your admissions line is not answered by a trained human seven days a week, including evenings. Every channel in this article ends in a phone call. Improving the front of the funnel while the back of it leaks is the most expensive mistake in this business.
• You want a partner who will guarantee an admission count. We will not, and an agency that does is either buying leads on your behalf or inventing the number.

A 90-day sequence

If you are starting from a site nobody has touched in three years, this is the order we would run it in.

Days 1 to 30. Search your own brand name and document who is bidding on it. Confirm or restore your FindTreatment.gov listing and complete the annual survey. Claim and complete a Google Business Profile for every location. Add the how-did-you-hear-about-us question to intake. Put your current vendor contracts in front of a healthcare attorney.

Days 31 to 60. Rebuild the pages that matter: home, each level of care, insurance and admissions, staff. Publish real photographs and named credentials. Put insurance information where a phone user sees it without scrolling twice.

Days 61 to 90. Publish the first five question-led pages, one for each real question your admissions team hears every week. Start the referral calls, five conversations a week, from a salaried person. Pull the first cost-per-admission-by-channel report and hold it next to your vendor invoices.

None of it is clever. All of it is yours, and next quarter it will still be there without a renewal.

The confusion is not your fault. Almost no agency in the addiction treatment space publishes pricing. The few public numbers range from $3,500–$6,000 entry-level builds to $50,000–$120,000 multi-location custom projects. Both can be legitimate. They are simply different products wearing the same name.

This guide gives you the real numbers: published agency tiers, itemized line costs, the compliance fees unique to addiction treatment (LegitScript alone is $4,690 in year one), and the admissions math that tells you what a website is actually worth. Every market figure links to its source. Where we state our own pricing, we say so plainly.

The 30-second answer

In 2026, most US addiction treatment centers pay $5,000–$15,000 for a professional single-location website, $15,000–$50,000 for a custom conversion-focused build with CRM, call tracking, and insurance verification, and $50,000–$120,000+ for multi-location operators. Template builds start around $2,000–$5,000. Add rehab-specific overhead: LegitScript certification ($1,595 application plus $3,095 per year, per facility), HIPAA-aware forms ($2,000–$8,000 setup), and $150–$800 per month in upkeep. One admission is worth $15,000–$50,000, so a site that produces even one extra admission has paid for itself.

Treatment center website costs at a glance

You have four ways to get a website built, and each occupies a different price band with a different risk profile. The table below is the honest version of the market in 2026, built from published prices and buyer surveys rather than sales pages.

One reference point for the top band: specialist agency Webserv publishes tiers at $8,000 (10 pages, one location), $15,000 (20 pages, three locations), and $25,000 (40+ pages, five or more locations). At the other end, DrugRehab.Agency publishes $3,500–$6,000 builds delivered in 30 days. Both are real. The difference is everything in the next section.

What actually determines the price

Six factors explain nearly every dollar of difference between a $5,000 quote and a $50,000 quote. If a proposal cannot tell you where it sits on each of these, it is a guess, not a quote.

1. Who builds it

Labor is the biggest input. Freelance designers charge a median of $92.75 per hour, agency surveys by GoodFirms show about half of development agencies quoting $3,000–$15,000 for small projects, and enterprise firms bill $225–$350+ per hour. The same 100-hour project costs $9,000 with one seller and $30,000 with another before scope changes at all.

2. Locations and levels of care

Every location adds pages, local SEO work, and structured data. Every level of care (detox, residential, PHP, IOP, outpatient) adds a content cluster that has to be clinically accurate. Published specialist tiers nearly double from one location to three ($8,000 to $15,000) and climb again past five locations ($25,000).

3. Integration depth

CRM integration, admissions call tracking, and an insurance verification (VOB) form workflow each add roughly $3,000–$10,000 per integration. A center running paid media usually needs all three, which alone can add $10,000–$30,000 to a build.

4. Compliance scope

HIPAA-aware implementation runs $2,000–$4,000 for basic measures and $5,000–$8,000 for comprehensive setups. Anything that collects protected health information (PHI, meaning identifiable health data) needs encrypted handling and a vendor who will sign a Business Associate Agreement. Stock WordPress and stock form plugins do not qualify.

5. Content and photography

Professional copywriting runs $60–$300 per page, and commercial photography runs $800–$3,000 per day. In this niche, real photos of your facility and staff are close to mandatory, because families deciding where to send a loved one can smell stock photography instantly.

6. Conversion engineering

A brochure shows information. An admissions engine moves a frightened family from a 2 a.m. search to a phone call in as few steps as possible. That work (crisis-sensitive UX, insurance-first messaging, sticky call bars, tested page layouts) is the real product a specialist sells, and we break it down in our guide to recovery website design that converts.

Where the money goes in a professional build

On a typical custom healthcare build, roughly 60% of the budget goes to design and development, with the rest split across strategy, content, compliance, and testing. Here are the published phase ranges from Patient10x's healthcare cost breakdown, which span everything from a small practice to a hospital system:

• Strategy and planning: $2,500–$15,000. Sitemap, admissions journey mapping, competitor and keyword research.
• Design: $5,000–$35,000. Brand application, page layouts, mobile-first UX, trust and credential presentation.
• Development: $8,000–$60,000. The actual build, CMS setup, speed optimization, tracking.
• Content: $3,000–$20,000. Clinically reviewed copy for every level of care, plus photography.
• Compliance setup: $2,000–$8,000. HIPAA-aware forms, encryption, accessibility (WCAG), audit trails.
• Integrations: $3,000–$25,000. CRM, call tracking, VOB workflow, analytics that do not leak PHI.

Quality assurance rarely appears as its own line, but it is why timelines run 8–16 weeks instead of 8 days: cross-device testing, accessibility checks, redirect mapping so you do not lose existing rankings, and a compliance review before anything goes live. When a quote is dramatically cheaper, one or more of these lines has quietly been deleted. Your job is to find out which one.

Treatment-center costs nobody quotes you for

Addiction treatment carries five cost lines that generic website guides never mention, and they are the reason a rehab site costs more than a dentist's site. Budget for these separately from the build itself.

LegitScript certification

If you want to run Google or Meta ads for addiction treatment, LegitScript certification is mandatory, and its fees are published: $1,595 to apply plus $3,095 per year per facility at the 1–9 facility rate, or $535 plus $1,070 per year for an individual practitioner. Expedited two-day review costs an extra $2,500. That is $4,690 in year one before you spend a dollar on ads.

Here is why it belongs in your website budget: LegitScript reviews your website itself, including clinical claims, staff credentials, ownership transparency, and required disclosures. A site written without those requirements in mind gets rejected and rewritten at your expense.

HIPAA-compliant forms

Any form that asks about someone's substance use collects PHI, and the free tier of every form builder is off the table. A form vendor that signs a BAA costs real money: Jotform gates HIPAA features to its $129 per month Gold plan, and Formstack's BAA-covered tier starts at $99 per month. That is $1,200–$1,550 per year, forever.

Insurance verification (VOB) workflow

A "verify my insurance" page is the highest-intent conversion point on a rehab site, and building it properly (secure intake, routing to your admissions team or VOB software) runs $3,000–$10,000 on the build side, with verification software adding $300–$800 per month depending on CRM integration.

Accreditation trust signals

Joint Commission and CARF badges convert because families and referents look for them, but they are not free to earn: Joint Commission behavioral health applications run $1,700–$8,000 with annual fees on top, per published estimates, while CARF's own fee schedule lists a $995 application. Displaying them correctly, above the fold and verifiable, is part of the design job. Our guide to healthcare trust principles covers how.

Real photography

A half-day shoot of your facility and team runs $800–$2,000, a full day $1,500–$3,000 or more. It is the highest-leverage $2,000 in this entire article, because trust is the product a treatment center website sells.

Ongoing costs after launch

Plan on $150–$800 per month after launch for a professionally maintained treatment center site, before any marketing retainers. The build is a one-time cost. These are not: Budget reviews are worth calendaring: set a reminder every quarter to check that hosting, plugins, and form tools are still on current versions, because most emergency rebuild invoices we see started as a $30 per month maintenance gap that nobody owned.

• Hosting: $20–$100 per month. Managed WordPress hosting starts at $20–$35 per month (WP Engine, Kinsta); modern Next.js hosting sits in a similar band with better speed.
• Maintenance and updates: $95–$400 per month is the typical small business band, covering software updates, backups, security patches, and small content changes.
• Compliance monitoring: $200–$800 per month for practices handling PHI, per Patient10x.
• HIPAA form tooling: $99–$129 per month, as covered above.
• Optional growth retainers: $500–$2,500 per month for ongoing SEO and content, which increasingly includes showing up in AI answers. We cover that shift in getting recommended by ChatGPT and Google AI.

Run the five-year total: a $12,000 build plus $300 per month is $30,000 over five years. That number should not scare you. It should change which question you ask, from "what does the site cost" to "what does the site produce." Which brings us to the math.

The ROI math: one admission pays for the website

A treatment center website pays for itself with a single admission, and that is not a slogan, it is arithmetic. Here are the two numbers that matter.

First, what an admission costs to buy with ads. Webserv's 2025 State of Rehab Marketing, an agency benchmark across $16.5 million in managed ad spend, 40,652 leads, and 994 admissions, found an average cost of $406 per lead and $16,608 per admission, with only 2.4% of leads becoming admits.

Second, what an admission is worth. Published figures put the value at roughly $15,000 for outpatient and up to $50,000 for residential, consistent with consumer-side pricing of about $12,500 on average for a 30-day inpatient stay.

Now connect them. Your website is the conversion endpoint for every dollar of that spend: the ad click costs the same whether your site converts 2% or 4% of visitors, but the second site cuts your cost per admission nearly in half. A $15,000 site that produces just one extra admission per quarter returns its full cost in the first quarter, then keeps producing.

Two more anchors to kill the sticker shock. Rehab digital marketing budgets commonly run $3,000–$25,000 per month, so a full custom website equals a few weeks of the ad spend you already accept, paid once. And websites are multi-year assets: Orbit Media measured leading marketing sites lasting about 2 years, while well-built sites averaged over 6. The cheap site is a 2-year rental. The good one amortizes across 5+ years of admissions.

Cheap website risks and agency red flags

The biggest budgeting risk is not overpaying, it is buying a cheap site that quietly burns your ad spend for two years. The failure modes are concrete: template layouts that cannot hold VOB forms and admissions flows, PHI collected through non-BAA form plugins, accessibility gaps that invite ADA demand letters, and redesigns that wipe out rankings through botched redirects.

The agency-side red flags are just as consistent across every buyer forum and every burned marketing director we have talked to:

• Guaranteed rankings or guaranteed admissions. Nobody can promise either. Walk away.
• 100% payment up front. Standard structure is a deposit with milestone payments.
• The agency keeps ownership of your domain, site, or code. This is the hostage clause, and it is the most expensive mistake on this list.
• No discovery questions. If they quote before asking about your levels of care, locations, and admissions process, they are selling a template.
• No itemized quote. A single mystery number means you cannot compare, negotiate, or audit.
• Screenshot-only portfolio. Ask for live URLs and a verifiable result.

How KhanWork prices it

We publish our pricing, which is rare enough in this niche to be a differentiator by itself. KhanWork is a US-focused design, development, and growth agency specializing in addiction treatment and recovery centers, with 200+ projects shipped, and every site built on Next.js for speed. Here is where each tier fits:

• Launch Partner, a one time build. Full branding plus a high-converting landing page, live in 14 business days with 3 revision rounds. Fits new or pre-launch centers that need a credible, compliant presence before the full site.
• Growth Partner, a monthly partnership. Full website design and build with 48-hour turnaround on requests, plus growth, social, and collateral design, on a 3 month minimum and month to month after that. Multiple parallel workstreams are available when you need more at once. Three months of Growth Partner sits below most single-location agency quotes while including ongoing iteration instead of a handoff.

The proof point we lead with: after our redesign, Cornerstone Healing Center saw a 20% increase in website conversions. At the admission economics above, a 20% conversion lift is worth multiples of any tier's annual cost. You can see the work in our case studies and the full breakdown on our pricing page.

Honest fit note: if you need a 40-page multi-location build with deep Salesforce integration on a fixed enterprise contract, a $50,000+ specialist project may fit better. If you need a site that converts, ships fast, and keeps improving monthly without a five-figure invoice up front, that is what we built our model for.

5 questions to ask any agency before signing

These five questions expose 90% of bad website deals before any money moves. Ask them in writing and keep the answers.

1. Who owns the domain, site, and code the day we part ways? The only acceptable answer: you do, in full, on final payment.
2. Can you itemize this quote? Design, development, content, compliance, and integrations should each carry a number you can compare against the ranges in this guide.
3. Will the site be LegitScript-ready at launch? Disclosures, credential displays, and compliant claims should be deliverables, not afterthoughts.
4. Which form vendor holds the BAA, and where does PHI live? If the answer is vague, your intake forms are a liability.
5. What measurable result did your last treatment center site produce? A real answer sounds like "20% more conversions," with a name attached. A fake one sounds like "great feedback."

The bottom line: expect $5,000–$15,000 for a strong single-location site, $15,000–$50,000 for a custom admissions engine, plus roughly $5,000–$7,000 in first-year compliance costs unique to this industry. Judge every quote against what one admission is worth, and never sign with anyone who will not put ownership and itemization in writing.

The reflex is to buy more traffic. That is the expensive answer. If your site turns 0.8% of visitors into enquiries, doubling traffic doubles a small number and doubles your ad bill. Fixing the conversion path costs less and keeps paying, because every visitor from that point forward lands on the better version.

What follows is a diagnostic in priority order. Each section is one leak: how to tell whether you have it, what the research says, and what to change. Every market number links to the source we read. Where we use our own numbers, we say so plainly.

The 30-second answer

If you have traffic but no leads, the cause is almost always one of six things: the traffic has the wrong intent, the page loads too slowly to hold anyone, the message never makes a clear ask, nothing on the page proves you are safe to hire, the form demands too much too early, or the lead arrives and nobody answers fast enough. Test in that order, because traffic quality and page speed sit upstream of everything else. For scale, the median landing page converts at 6.6% (Unbounce, Q4 2024), while whole-site conversion runs far lower: 1.7% for new visitors and 2.9% for returning ones (Contentsquare, 2026). Find the cheapest leak, close it, measure again.

Where leads actually leak

Leads leak at six specific points, and you only ever see the number at the end. The path a stranger walks is: arrive, understand, believe, decide, submit, get answered. A drop at any one of those steps shows up in your inbox as the same symptom, which is why "our website does not work" is a diagnosis nobody can act on.

So measure each step separately. You want the conversion rate of your top five landing pages, split by device, split by channel. If one page converts at 4% and another at 0.2%, you no longer have a website problem. You have a page problem, and it takes an afternoon instead of a rebuild.

Leak 1: the traffic was never going to buy

The most common cause of high traffic and low leads is traffic that arrived with no intention of hiring anyone. Someone searching "what does a rehab cost" and someone searching "outpatient rehab near me accepting insurance" are two different people, and only one of them is ready to call. If most of your sessions come from the first kind of query, your site is behaving normally and your keyword mix is the problem.

Channel tells you almost as much as intent does. Contentsquare's 2026 Digital Experience Benchmark, built on 99 billion sessions across more than 6,000 sites, puts paid search at 2.8%, AI-referred traffic at 1.3% and organic social at 0.7%. Traffic from AI assistants is small but it was the only channel whose conversion rate grew, up 55% year over year.

Two structural shifts make this worse than it used to be. First, a majority of searches now end without a click at all: SparkToro's 2024 zero-click study, built on Datos clickstream data, found 58.5% of US Google searches produced zero clicks. Second, the traffic that does reach you has changed shape. Semrush's study of more than 10 million keywords found AI Overviews appearing on 15.69% of queries by November 2025, and the intent mix behind them moved sharply: informational triggers fell from 91.3% in January to 57.1% in October, while commercial triggers rose from 8.15% to 18.57%.

Read that second number the right way. Fewer informational visits will reach you, and more of the people who do arrive are further along. That is good news for lead volume per visit and bad news for anyone whose traffic report is mostly blog posts answering definitions. We wrote about the mechanics of that shift in getting recommended by ChatGPT and Google AI, and about intent-led keyword structure in our guide to SEO for service businesses.

The test: sort your landing pages by sessions, then look at conversions per page. If your highest-traffic pages are informational and your highest-converting pages get almost no traffic, you do not have a conversion problem. You have an inventory problem, and the fix is to build and rank the commercial pages you are missing.

Leak 2: the page is too slow to close

Speed changes conversion rates by more than most owners expect, and the effect is largest on exactly the traffic you paid for. Portent's 2022 study of more than 100 million page views across 20 sites, 14 of them B2B lead generation, found that a site loading in 1 second converts about three times better than one loading in 5 seconds.

The tighter number comes from Google's own commissioned research. In Milliseconds Make Millions, run by 55 and Deloitte across 37 European and American brand sites and more than 30 million sessions, a 0.1 second improvement in mobile site speed raised form submission progression on lead-generation sites by 21.6%, with page views up 7%. Retail conversions rose 8.4% and travel conversions 10.1% off the same tenth of a second.

Now add the device split. Contentsquare's benchmark puts 69.9% of traffic on mobile while desktop converts 74% higher. Some of that is behavioural, but a large part is that mobile visitors experience your slowest, heaviest version on the worst connection. If you test one thing this week, open your own contact page on a phone on cellular data and count the seconds.

The test: run your top landing page through PageSpeed Insights on mobile and look at field data, not the lab score. Then check what is actually loading: chat widgets, tracking scripts, oversized hero images and unused page-builder plugins account for most of the weight we find on the sites we audit.

Leak 3: the page never makes the ask

Plenty of sites with real traffic and real credibility simply never ask for the business. The visitor reads three sections of pleasant description, cannot tell what happens if they get in touch, and leaves to compare you against someone clearer.

Four things fix most of it, and none of them require a rebuild:

• Say what you do, for whom, in the first line. Not your positioning statement. The thing a stranger would type into a search box.
• Put one primary action above the fold and repeat it. Two competing buttons of equal weight is the same as no button.
• Make the phone number a link. On mobile, a phone number that is not tappable is a decision you have asked a stressed person to type out by hand.
• Tell them what happens next. "We reply within one business day with two or three questions and a rough price" removes more hesitation than any amount of adjectives.

We laid out the full structure, section by section, in the anatomy of a high-converting landing page. The short version: every page should be answerable in one sentence, and that sentence should end in an action.

Leak 4: nothing proves you are safe to hire

Your website is rarely the first place someone checks, which means it is where belief either gets confirmed or quietly falls apart. BrightLocal's 2026 Local Consumer Review Survey of 1,002 US consumers found 97% read reviews for local businesses, and after reading positive ones, 54% go and check the business's website while 20% contact the business directly. Standards have also tightened: 68% will only use a business rated four stars or higher, and 31% now hold out for 4.5 or better.

So a third of your prospects arrive already half-convinced, and the site's only job is not to undo it. Stock photography, unnamed testimonials and a decade-old copyright date undo it. What works is specific and verifiable: named clients, real photographs of real people, a result with a number attached, and credentials shown where someone would look for them rather than buried on an about page.

One concrete example from our own work, since we ask clients for the same specificity: after our redesign, Cornerstone Healing Center saw a 20% increase in website conversions. That is a sentence a stranger can weigh. "Great feedback from our clients" is not.

Leak 5: the form interrogates before it helps

Every field you add is a small reason to leave, and most contact forms are built for the sales team's convenience rather than the visitor's. Baymard Institute's checkout research found the average flow in 2024 carried 11.3 form fields where most sites need only 8, and that 17% of users have abandoned a purchase because the checkout was too complicated. That is e-commerce checkout rather than lead capture, and the numbers do not transfer directly, but the mechanism does: length itself costs completions, independent of how reasonable each question is.

For a service business, a first-contact form needs a name, a way to reach them, and one open field. Budget ranges, dropdown menus of service categories, "how did you hear about us" and a required company size all belong in the reply, not the gate. If your sales process genuinely needs qualification, ask on the call or use a two-step form where the low-friction question comes first.

Leak 6: the lead lands and nobody answers

The leak that costs the most is the one after the form works. Chili Piper's 2022 study of B2B vendor response times found an average reply of four hours and fifty minutes among companies that responded at all, only 20% replying within five minutes, 7% inside a minute, and roughly 30% never responding.

Those companies all paid to acquire the lead. Roughly a third then paid nothing to follow it, which is the most expensive line item in this entire article and the only one you can fix this afternoon without touching the website. Route enquiries to a phone that someone actually carries, set an internal target you can measure, and send an immediate acknowledgement that names the person who will reply and when.

The same applies to phone calls. If your admissions line, front desk or main number rings out during business hours, no amount of conversion design upstream will help.

You cannot fix what you never counted

A surprising share of "we get no leads" turns out to be "we get leads and do not count them." Before you rebuild anything, verify all five of these:

1. Form submissions fire a tracked event. Not a thank-you page view that also fires on refresh. A real, deduplicated event.
2. Phone calls are counted. For local and healthcare businesses, calls often outnumber form fills. If they are invisible in analytics, your conversion rate is fiction.
3. Spam is excluded. Bot submissions inflate the numerator; bot sessions inflate the denominator. Both make the report useless.
4. Channels are tagged. Untagged campaign links land in direct traffic, which is where attribution goes to die.
5. You look at pages, not the site. One blended number cannot tell you where to spend the next hour.

Do this first. A meaningful share of the emergency redesigns we get asked to quote turn out to be a tracking problem, and it is a much better week when the answer costs nothing.

What a healthy number actually looks like

There are two numbers people confuse, and comparing yourself against the wrong one causes real damage. Unbounce's Q4 2024 dataset of 41,000 landing pages, 464 million visits and 57 million conversions puts the median landing page at 6.6%. That is a dedicated page, usually receiving campaign traffic, with one job.

Whole-site conversion is a different animal. Contentsquare's 2026 benchmark has new visitors converting at 1.7% and returning visitors at 2.9%, across all traffic including people who came to read one blog post and leave. Holding your homepage to a landing-page benchmark will make a healthy site look broken.

Use benchmarks as context, never as a target. The number that matters is your own, last quarter, on the same pages and the same channels. If a page moved from 1.1% to 1.9%, you got 73% more enquiries from traffic you were already paying for, and no industry average can tell you whether that was a good quarter.

The 30-day fix, in order

Work top to bottom. Each step is cheaper than the one below it, and skipping ahead is how people spend a redesign budget solving a tracking bug.

• Days 1 to 3: verify the plumbing. Submit your own forms, confirm delivery, check that submissions and calls are tracked, and exclude spam.
• Days 4 to 7: set the response standard. Decide who answers, on what device, within what window, then send an immediate acknowledgement automatically.
• Days 8 to 14: fix the top page. Take your highest-traffic commercial page. Clarify the first line, cut the form to three or four fields, make the phone number tappable, add one specific proof point.
• Days 15 to 21: fix mobile speed. Compress hero images, remove unused scripts and page-builder bloat, and retest on a phone using cellular data rather than office wifi.
• Days 22 to 30: fix the traffic mix. Identify the commercial searches you should rank for and do not, and plan the pages that answer them.

Then wait a full cycle before judging. Most service businesses do not get enough weekly conversions for a week-over-week comparison to mean anything, and reading noise as signal is how good changes get reverted.

When the site itself is the problem

Sometimes the honest answer is that the pages are not worth optimising. The signs are consistent: the site cannot be edited without a developer, mobile layouts break, page speed is structurally poor because of the platform rather than the images, and the design predates the way the business now sells. At that point every fix above is being applied to a foundation that will not hold it.

That is the work we do. KhanWork is a design and development partner for US and European service businesses, with 200+ projects shipped. Our Launch Partner package is a one time build covering full branding plus a high-converting landing page, and Growth Partner is a monthly partnership for continuous design and build. A focused build typically goes live in 14 business days from the day content and brand assets land, though scope and content approvals move that. You can see the full breakdown on our pricing page, the results on our case studies, and how we approach conversion for local service businesses specifically.

Who should not hire us: if your traffic is genuinely all informational and your commercial pages simply do not exist yet, buy content and SEO before design. If your close rate on qualified calls is the actual bottleneck, a better website will send you more calls you are not converting. And if you have never once measured your conversion rate by page, do that first. We would rather you spend a free afternoon on the diagnostic than a paid month on the wrong fix.

The pattern behind almost every version of this problem is the same. Traffic is the input everyone reports on, leads are the output everyone judges, and the six steps in between are the part nobody instruments. Instrument them, and "we get traffic but no leads" turns into a specific, cheap, fixable sentence about one page and one number.

The same thing is happening in every service category. A founder asks Claude for "the best web agency for healthcare brands." A homeowner asks Perplexity who to call about a foundation crack. A CFO asks Google's AI Mode which bookkeeping firm to shortlist. The AI answers with names, and those names win the customer before a traditional search ever happens.

Most businesses have no idea whether they appear in these answers, and the data says most do not. This guide explains exactly how ChatGPT, Perplexity, and Google's AI features choose which businesses to recommend, and gives you a concrete 7-step plan to become one of them. We will use addiction treatment centers as the running example because it is the highest-stakes version of this problem, but every tactic applies to any US service business.

The 30-second answer

To get recommended by ChatGPT, Perplexity, and Google AI, you need four things: a consistent, unambiguous business identity across the web (same name, address, services, and positioning everywhere), answer-shaped content on your site that states facts directly with statistics and sources, strong third-party proof (reviews, directories, press mentions, and "best of" list inclusions, which AI trusts more than your own site), and a technically clean, server-rendered website indexed in both Google and Bing. There is no paid placement. Visibility is earned through entity clarity and third-party trust.

Why AI search changes how customers find you

AI search changes discovery because the answer now happens before the click, and most of the time the click never comes. ChatGPT crossed 900 million weekly active users in February 2026, double its February 2025 figure, and 49% of US adults now use AI chatbots, up from 33% in 2024. About 37% of consumers say they now begin searches with AI tools, and 42% prefer a chatbot over a search engine for multi-step research, which is exactly what hiring a service business is.

On the Google side, the change is just as sharp. A randomized field study found organic clicks drop 38% on queries that trigger an AI Overview, Search Engine Land reported a 61% organic CTR decline, and roughly 60-70% of Google searches now end with zero clicks. Small businesses feel this hardest: sites with 1,000-10,000 daily pageviews saw search referrals fall about 60% over two years per Chartbeat data reported by Axios.

Here is the part that should get your attention: the demand side is enormous and the supply side is nearly empty. Around 45% of consumers now ask AI assistants to find local businesses, yet only 1.2% of business locations ever get recommended by ChatGPT for local service queries. Almost nobody has optimized for this. The businesses that move first inherit a discovery channel with almost no competition.

And the visitors who do arrive from AI are better customers. Semrush's analysis values the average AI search visitor at 4.4x a traditional organic visitor, ChatGPT ecommerce traffic converts 31% higher than non-branded organic, and brands cited inside AI Overviews earned 35% more organic clicks than uncited brands. The honest caveat: AI referrals are still only around 1% of most sites' traffic, though that slice grew 206% year over year. Treat AI visibility as the fastest-growing layer on top of SEO, not a replacement for it. Our guide to SEO for service businesses covers the foundation this builds on.

How AI engines actually pick businesses

AI engines pick businesses by cross-referencing what the wider web says about you: your entity data, your reviews, the third-party pages that mention you, and how extractable your own content is. Each engine has different plumbing, and knowing the plumbing tells you where to invest.

ChatGPT: Bing is the gatekeeper

ChatGPT's live web answers are retrieved through Bing's index. A Seer Interactive study found that 87% of SearchGPT citations matched Bing's top results for the same query. The practical consequence is blunt: a site that is invisible in Bing is invisible in ChatGPT's web answers, no matter how well it ranks on Google. For recommendations drawn from training data rather than live search, ChatGPT leans on high-trust sources it saw repeatedly: Wikipedia, major directories, review platforms, and editorial "best of" lists.

Perplexity: the most winnable engine

Perplexity runs a real-time retrieval pipeline that pulls roughly 5-10 pages per query and cites 3-4. It rewards pages that state the answer in the opening lines, contain verifiable specifics rather than vague marketing, and are recently updated: a large share of its citations come from current-year content. Because every Perplexity answer carries citations and its re-ranking favors relevance over raw domain authority, it is the engine where a well-structured page from a smaller site has the best odds.

Google AI Overviews and AI Mode: normal SEO, looser rankings

Google's official documentation says AI Overviews and AI Mode source from the standard Google index with no special markup, files, or AI-specific requirements. Standard SEO is the on-ramp. But the link between ranking and citation is loosening fast: Ahrefs data shows the share of AI citations coming from top-10 ranking pages fell from about 76% in mid-2025 to roughly 38% in early 2026. That kills the comfortable assumption that "we already do SEO, so we are fine." It also means well-structured pages outside the top 10 now have a citation path that never existed in the blue-links era.

The signals all three share

Across engines, the strongest quantitative evidence points off-site. Ahrefs' 75,000-brand correlation study found branded web mentions correlate with AI visibility at 0.664 versus 0.218 for backlinks, with YouTube mentions the single strongest signal at roughly 0.737. Semrush's 150,000-citation analysis found Reddit cited in about 40% of LLM answers and Wikipedia in about 26% (treat the exact shares as directional; they shift month to month). And the only peer-reviewed controlled experiment in this field, the Princeton and Georgia Tech GEO study, found that adding statistics, quotations, and source citations to a page boosts AI visibility 30-40%, while promotional tone underperforms. Being talked about matters more than what you say about yourself, and when you do speak, evidence beats adjectives.

The 7-step fix

The fix is a sequence, not a menu: make your business unambiguous to machines, make your content quotable, then make the rest of the web vouch for you. Work the steps in order. Each one compounds the next.

Step 1: Lock down your entity and NAP consistency

AI models recommend businesses they are confident about, and confidence comes from agreement across sources. Your name, address, phone, service list, and one-sentence positioning must be identical across your website, Google Business Profile, Bing Places, Yelp, BBB, industry directories, LinkedIn, and every other public profile. Contradictory data across listings lowers the model's confidence and quietly removes you from consideration. Claim Bing Places even if you have never thought about Bing: it feeds the engine 900 million people use weekly.

Step 2: Publish answer-shaped content

Restructure your key pages so a machine can lift a complete answer from them. Lead every section with a direct 2-3 sentence answer under a question-form heading, then elaborate. Add the three elements the Princeton GEO experiment proved move citations: statistics, quotations from credible sources, and linked references (statistics addition scored a 41% gain on the study's position-adjusted visibility metric). Write like a reference document, not a brochure. Analyses of ChatGPT's citation behavior show it favors pages whose claims are self-contained and easy to extract verbatim.

Step 3: Add structured data that describes the entity

Implement JSON-LD schema for Organization, LocalBusiness (with hours, geo, service area), FAQPage on question-heavy pages, and Article on editorial content. Be honest about what this does: Google says no special schema is required for AI Overviews, but Microsoft has stated that schema helps its LLMs understand content, which matters because Bing feeds ChatGPT. Schema is entity disambiguation for machines: it tells every crawler exactly who you are, what you do, and where. It is a confidence multiplier, not a magic switch.

Step 4: Build third-party proof where AI already looks

This is the highest-leverage step and the one most businesses skip. AI engines lean on a small set of trusted surfaces: review platforms, Reddit, Wikipedia, YouTube, LinkedIn, and editorial sites. Review platforms function as AI's trust layer: engines cross-check multiple review sites before recommending a vendor, so maintain complete, actively reviewed profiles on the 2-3 platforms that matter in your category. Pursue inclusion in the "best X in [city]" articles that already rank in your niche, because recently updated list posts are the most-cited page format in ChatGPT sources. Pitch podcasts and local press; earned media is the delivery mechanism for the branded mentions that correlate with AI visibility three times more strongly than backlinks. Participate genuinely in relevant subreddits, with real answers under a transparent account. Expect this step to compound over 12-18 months, not 30 days.

Step 5: Build direct-answer pages for your money queries

List the 10-20 questions a buyer asks an AI right before choosing a provider: "best [service] in [city]," "how much does [service] cost in 2026," "[option A] vs [option B]," "does [service] take [insurance/payment type]." Build one page per question that answers it completely, includes the current year in the title, gives concrete numbers, and compares options fairly, including ones you do not sell. Yes, publish your own comparison lists: Ahrefs found vendor-published "best X" lists among the most prominent ChatGPT sources, provided each entry has distinct, concrete rationale. Then keep these pages fresh with substantive updates: AI-cited content is measurably fresher than organic-ranked content (about 25.7% fresher in Ahrefs' 17-million-citation analysis), and much of that freshness comes from maintained older pages, not new ones.

Step 6: Fix the technical foundation

Three checks, in order. First, verify your site in Bing Webmaster Tools, submit your sitemap, and confirm your key pages are indexed; this is a hard prerequisite for ChatGPT citations. Second, confirm your content exists in the initial HTML response. The Vercel and MERJ dataset of 500 million GPTBot fetches found zero evidence of JavaScript execution by AI crawlers: anything your site renders client-side simply does not exist for ChatGPT, Claude, or Perplexity. If your site is a client-rendered React or Vue app, move to server-side rendering or static generation; this is one of the reasons we build on Next.js rather than page builders, a tradeoff we break down in WordPress vs Next.js in 2026. Third, check robots.txt and your CDN or firewall logs to confirm the search-time agents (OAI-SearchBot, ChatGPT-User, Claude-SearchBot, PerplexityBot) are not being blocked. Cloudflare bot rules silently blocking AI agents is the most common invisible killer we find in audits.

Step 7: Measure your AI visibility monthly

Only 14% of marketers actually track AI citation visibility, even though 43% call AI optimization an active 2026 strategy. Tracking is simple to start: run your 10-20 money queries through ChatGPT, Perplexity, and Google AI Mode each month, log who gets named, and note the sources each engine cites. Those cited sources are your target list for Step 4. Watch AI referrals in your analytics (chatgpt.com, perplexity.ai referrers) and treat them as a conversion cohort, not a traffic line: their value shows up in lead quality.

AI visibility for recovery centers: the highest-stakes version

For addiction treatment centers, AI visibility is not a marketing curiosity; it is where families now start. One in three US adults has used AI chatbots for health information and 16% have used them for mental health questions, per KFF polling. OpenAI reports 40 million people ask ChatGPT health questions daily, with over 40% of health users researching treatment options and roughly 70% of those conversations happening outside clinical hours. That is the 1 a.m. mother in Phoenix, at national scale.

The economics make organic AI visibility unusually valuable in this niche. Paid search clicks run $25-$150 with "alcohol rehab near me" reaching about $185 per click, and a residential admission typically costs $4,000-$8,000 to acquire through PPC. Meanwhile, per the 2023 NSDUH, only 15.6% of the 48.5 million Americans with a substance use disorder received treatment. Demand is not the constraint. Findability and trust are, across roughly 17,561 competing facilities.

The playbook above applies, plus four niche-specific moves. First, get every trust credential machine-readable: Joint Commission or CARF accreditation stated in plain text and schema on your site, because these are the signals families, payers, and now AI models use to separate legitimate programs from bad actors. Second, claim and complete your FindTreatment.gov listing. It is SAMHSA's official locator and the government-endorsed starting point that other agencies route help-seekers to. Third, maintain LegitScript certification. It has been mandatory for treatment ads on Google, Meta, and Microsoft since 2018, takes 4-8 weeks, and doubles as a third-party legitimacy signal. Fourth, make your insurance-verification page a first-class asset: industry benchmarks report that around 70% of admissions stall on unclear costs, and "Verify Insurance" CTAs outperform generic CTAs by 32%. "Which centers take my insurance" is precisely what families ask ChatGPT, so publish clear, current payer information in crawlable HTML. We cover the conversion side in depth in our guide to addiction recovery website design that converts.

One ethical line, stated plainly: everything above is about making true information findable. In behavioral health, exaggerated claims, fake reviews, or gaming tactics are not just brand risks, they can harm people at their most vulnerable and violate the compliance regime (LegitScript, state regulators, FTC) that governs this industry. Accuracy is the strategy.

What not to do

The fastest way to waste an AI visibility budget is to buy tactics that sound technical but do nothing. These four come up in almost every audit we run.

Do not treat llms.txt as a ranking tactic

Despite tens of thousands of sites publishing llms.txt files, no major AI engine has confirmed using them. Google's John Mueller said "no AI system currently uses llms.txt", and Ahrefs' log analysis found about 97% of published llms.txt files receive zero AI-bot requests. Adding one is harmless, cheap future-proofing. Anyone selling it as the fix is overclaiming.

Do not mass-produce AI-generated content

Flooding your blog with thin AI-written posts optimizes for a 2019 version of search. The Princeton study showed evidence density wins and promotional, keyword-stuffed content underperforms. Ten pages with real statistics, sourced claims, and specific local knowledge will beat two hundred generic ones.

Do not expect schema alone to do anything

Structured data on top of thin content, inconsistent listings, and zero third-party mentions is a label on an empty box. Schema clarifies an entity that already has substance; it cannot create trust that does not exist elsewhere on the web.

Do not astroturf Reddit or buy reviews

Marketers are actively gaming Reddit because its visibility correlates with LLM mentions, and platforms crack down on exactly this behavior. Fake reviews and sockpuppet threads get deleted, get you banned, and in regulated industries like treatment, invite FTC and regulator attention. Earn the mentions; the compounding is worth it.

Your 30-day action plan

You can complete the foundation in 30 days with a few focused hours per week. Here is the sequence we run for clients.

Week 1: Baseline and entity. Run your 10-20 money queries through ChatGPT, Perplexity, and Google AI Mode; screenshot who gets recommended and which sources are cited. Audit every public listing for NAP and positioning consistency, and fix mismatches. Claim Bing Places and verify the site in Bing Webmaster Tools.

Week 2: Technical. Confirm key content is in the raw HTML (view-source test). Check robots.txt and CDN logs for blocked AI agents and unblock the search-time bots. Submit sitemaps to Bing, fix indexing gaps, and add Organization, LocalBusiness, and FAQPage schema.

Week 3: Content. Rewrite your top 5 service pages answer-first: direct answer under each heading, statistics with linked sources, an FAQ section built from real customer questions. Add or update the current year and a substantive refresh to your two most important pages. If you are a treatment center, ship or upgrade the insurance-verification page this week; nothing else on the site matters more.

Week 4: Third-party proof. Launch a systematic review-request flow on your two most relevant platforms. Identify the "best of" lists and directories your baseline research showed AI citing, and start outreach to the top five. Pitch one podcast or local publication. Then put a monthly 30-minute recurring block on the calendar to re-run the baseline queries and log movement.

When to hire help

Hire help when the checklist stalls on execution: the site needs re-platforming to server-rendered HTML, nobody owns content production, or three months of self-serve effort has not moved the baseline. The 30-day plan above is genuinely doable in-house, and if you run it and start showing up for your money queries, you may not need an agency at all.

Where an agency earns its fee is the compounding work: engineering a fast, server-rendered site AI crawlers can actually read, producing evidence-dense content on a schedule, and running the digital PR that builds mentions for 12-18 months. That combination is what we do at KhanWork: conversion-focused sites built on Next.js, plus SEO and AI search optimization for US service businesses, with deep experience in healthcare and recovery brands. For Cornerstone Healing Center, our flagship treatment-center client, that approach lifted website conversions 20%; the full breakdown is in our case studies, and our pricing is public because cost transparency is advice we take ourselves.

Either way, start the baseline this week. Ask ChatGPT, Perplexity, and Google AI the questions your customers ask, and see whether you exist. In a market where 45% of consumers ask AI for recommendations and 1.2% of businesses show up, the gap will not stay open forever. Right now, it is wide open for whoever moves first.

This guide skips the tribalism. It breaks down performance, SEO, AI search, cost, security and editing in plain language, with diagrams, so you can choose with confidence.

The 30-second answer

Choose WordPress if a non-technical team needs to publish and edit content every day on their own, or you depend on a specific plugin such as WooCommerce, memberships, an LMS or a booking system.

Choose Next.js if speed is tied to revenue, you want a custom, brand-led design that is not boxed in by a theme, or you are building something interactive that needs to scale.

Want both easy editing and elite speed? Use them together in a headless setup. We cover that below.

WordPress vs Next.js at a glance

Here is the whole comparison in one view. The rest of the article explains the why behind each row.

What each one actually is

WordPress: a content management system

WordPress is a CMS written in PHP. It stores your content in a database and, on request, assembles a page using your theme and plugins. Its superpower is the editing experience and a plugin for almost anything: shops, forms, memberships, bookings, courses. That flexibility is why it powers a huge slice of the internet.

Next.js: a web framework

Next.js is a framework built on React. Instead of assembling pages on the server for every visitor, it can render them ahead of time into static HTML, or on demand at the edge, and ship only the JavaScript a page truly needs. It is a developer tool rather than a point-and-click editor, which is both its strength (control and speed) and its trade-off (you need code or a connected CMS to publish).

Performance & Core Web Vitals

This is where the two diverge most. A typical WordPress install ships a theme, a page builder such as Elementor or Divi, and a stack of plugins. Each one adds its own CSS and JavaScript, and much of it loads whether the page needs it or not. Without careful optimisation, Core Web Vitals suffer, and Core Web Vitals are a confirmed Google ranking signal.

Next.js renders pages to static HTML at build time, ships minimal JavaScript, and gives you granular control over exactly what loads and when. In practice, a well-built Next.js site routinely scores in the high 90s on mobile Lighthouse, while a plugin-heavy WordPress site often lands far lower until it is tuned.

How they are built (and why it matters)

The performance gap is not magic, it comes straight from architecture. WordPress is coupled and dynamic: content and presentation live in the same system, and the server rebuilds the page on almost every request. Next.js is usually decoupled and prebuilt: pages are generated once, then served as static files from a global CDN close to the visitor.

Prebuilt-and-served-from-the-edge is why a good Next.js site feels instant almost everywhere in the world. WordPress can approximate this with aggressive caching and a CDN in front, which is exactly the kind of tuning a lean build needs.

SEO & AI search visibility

For traditional SEO, both platforms can rank. Google cares about crawlable, fast, well-structured pages and genuinely useful content, none of which is exclusive to one framework. The technical baseline, fast Core Web Vitals, clean HTML, a valid sitemap, canonical URLs and structured data, is easier to hit by default in Next.js and achievable in WordPress with the right setup.

AI search is the newer battleground. Assistants like ChatGPT, Perplexity and Google AI Overviews read and cite pages, and they favour content they can parse quickly: server-rendered HTML, clean semantic markup, and structured data (FAQ, Article, Product schema). Content that only appears after heavy client-side JavaScript is a risk, because some crawlers may not execute it.

In short: the platform is a means, not the message. The winning sites pair a solid technical base with content worth citing.

Cost & maintenance over time

WordPress has a lower upfront cost and an enormous ecosystem, there is a plugin for almost everything. The trade-off is ongoing maintenance: plugin and core updates, security patches, and the occasional morning where an update broke the layout. Because WordPress is so widespread, it is also a bigger target for bots and vulnerabilities, so that maintenance is not optional.

Next.js sites are typically deployed as static files or serverless functions to a CDN or a host like Vercel. There is almost no attack surface, near-zero maintenance, and hosting is cheap or free at small scale. The trade-off shows up at build time and in content edits, which usually need a developer or a connected headless CMS.

Security

Security follows from the architecture too. A dynamic WordPress site exposes a database, a login, and dozens of third-party plugins, each a potential entry point that must be kept patched. Most WordPress hacks trace back to an out-of-date plugin, not WordPress core itself.

A static Next.js site has almost nothing to attack: there is no database or admin login sitting on the public page, just prebuilt files on a CDN. This is not a knock on WordPress, a maintained WordPress site is perfectly safe, but it is a real reason regulated and high-traffic businesses lean toward static or headless builds.

Editing & developer experience

This is WordPress at its best. Marketers can log in, write a post, drop in an image, and publish, no developer required. For teams that ship content constantly, that self-service editing is a genuine competitive advantage.

Next.js flips the trade-off. Developers get version control, previews, instant rollbacks and a clean codebase, which makes the site a joy to extend and scale. But publishing a simple copy change means either touching code or wiring up a headless CMS so editors get a friendly interface again.

When WordPress is the right call

• A marketing team publishes weekly and needs full self-service editing.
• You depend on a specific plugin ecosystem: WooCommerce, memberships, an LMS, bookings.
• Budget is tight and you need to launch fast on a familiar, well-documented tool.
• Your site is mostly standard pages, blog posts and forms.

When Next.js is the right call

• Page speed is directly tied to revenue: lead generation, SaaS, high-ticket services.
• You want a custom, brand-led design that is not constrained by a theme.
• You are planning to scale and want a maintainable, secure, developer-friendly codebase.
• You need custom interactivity: calculators, dashboards, complex multi-step forms.
• AI search visibility and technical SEO are strategic priorities.

The hybrid: headless WordPress

You do not always have to choose. In a headless setup, WordPress keeps doing what it is best at, being the editor and content store, while Next.js renders a fast, custom front-end from that content over an API. Your team keeps the WordPress dashboard they know; your visitors get static, edge-served speed.

It is the most popular middle ground for companies that publish often but refuse to compromise on performance. The cost is a more involved initial build, which is why it suits teams that have outgrown a standard WordPress theme but still need daily editing. Our headless CMS migration guide covers when this pays off and how to do it safely.

How to choose: a simple framework

Match the platform to how you will actually run the site, not to what sounds modern. Three questions settle most decisions: Who edits the site and how often? Is speed tied to revenue? Do you need custom interactivity or unusual design?

Migrating without losing rankings

Switching platforms is safe when you protect your SEO through the move. Whether you go from WordPress to Next.js or to a headless setup, the rules are the same: keep every URL identical where you can, 301-redirect anything that changes, preserve titles, metadata and structured data, and re-submit your sitemap the day you launch.

Do this carefully and visitors notice only one thing: a faster site. Skip it and you can lose hard-won rankings overnight, so treat migration as its own project with a rollback ready. The step-by-step plan lives in our migration guide.

The verdict

There is no universally better option, only the better fit for your goals. If content velocity is your priority, start with a lean, well-optimised WordPress build. If performance, design freedom, security and conversion lead, Next.js will serve you for years. And if you want both, headless gives you the editing of one and the speed of the other.

At KhanWork we build all three, and we recommend based on your goals, not our preferences. That is the only honest way to answer this question.

What "headless" actually means

A traditional CMS like WordPress couples content and presentation: the same system stores your posts and renders your pages. A headless CMS only stores and serves content (via an API). A separate front-end, usually a framework like Next.js, decides how it looks. The "head" (the front-end) is decoupled from the "body" (the content).

Signals it's time to migrate

• Your site is slow and plugin bloat is fighting your Core Web Vitals.
• You publish to multiple channels (web, app, email) and want one content source.
• Editors and developers keep stepping on each other in one system.
• Security and maintenance overhead is eating your team's time.

If none of these hurt, you probably don't need to migrate yet. Headless is a tool, not a trophy.

The Sanity + Next.js + Vercel stack

Each piece does one job well:

• Sanity, a flexible, real-time headless CMS. Editors get a customisable studio; developers get structured, queryable content.
• Next.js, renders fast, SEO-friendly pages from Sanity's data, statically or on the edge.
• Vercel, deploys and serves it globally with previews, instant rollbacks and near-zero ops.

Together they give editors a clean writing experience and visitors a site that loads in a blink.

What it costs

A migration is not a like-for-like swap; you're rebuilding the front-end. For a typical marketing site, budget for design/build of the new front-end plus content modelling in Sanity. The upside is that ongoing hosting and maintenance costs usually drop afterwards, and speed gains often pay for the project in conversions and rankings.

A safe migration plan

1. Audit & inventory. Catalogue every URL, template and content type on the current site.
2. Model content in Sanity. Design schemas that match how your team actually writes.
3. Rebuild the front-end in Next.js. Match URLs exactly to preserve SEO.
4. Migrate content. Script the import so nothing is lost or mangled.
5. Preserve SEO. Keep URLs identical where possible; 301-redirect anything that changes; re-submit your sitemap.
6. Launch behind a preview, then flip DNS. Test on a staging URL, then cut over with rollbacks ready.

Do this carefully and visitors won't notice anything except a faster site, which is exactly the point.

Why recovery sites are different

Most visitors arrive anxious, skeptical and in a hurry. They're comparing you to other centers and asking one question: can I trust these people with someone I love? Your site has seconds to answer yes.

(Planning the budget side too? Our 2026 treatment center website cost guide breaks down real prices, compliance fees and the ROI math.)

Design for trust first

• Real photos of real people and facilities. Stock imagery reads as hiding something.
• Accreditations and licensing up front, Joint Commission, LegitScript, state licensing badges.
• Staff bios with faces and credentials. People trust people, not logos.
• Outcome data and genuine testimonials (with consent), not vague promises.
• A calm, human tone. Clinical-but-warm beats corporate or salesy every time.

Clear conversion paths

Give anxious visitors one obvious next step on every screen. In practice that means:

• A sticky "Verify Insurance" and call now button, insurance is the #1 anxiety.
• A short, reassuring contact form (fewer fields = more submissions).
• 24/7 phone number prominent on mobile, click-to-call enabled.
• Clear pages for the questions people actually ask: cost, insurance, what to expect, admissions.

Compliance & privacy

Recovery sites handle sensitive health information, so privacy isn't optional. Use HIPAA-aware forms and a Business Associate Agreement with any provider that touches that data. Be careful with tracking pixels on pages where people submit health details, the wrong analytics setup can create real compliance exposure. Publish a clear privacy policy and honour it.

Local & organic SEO

Most admissions are local or regional, so local SEO is where the leads are. Claim and optimise your Google Business Profile, build location-specific pages, gather reviews consistently, and publish genuinely helpful content, on treatment options, insurance, and recovery, that answers what families search for. This is the same SEO foundation every service business needs, applied with extra care for a sensitive audience.

Trust is the conversion

In most industries, design drives conversion through clarity and persuasion. In healthcare, it drives conversion through credibility. Get trust right and the booking follows.

The 10 principles

1. Lead with credibility. Credentials, accreditations and affiliations near the top.
2. Show real faces. Photos of your team and space, not stock models.
3. Make booking obvious. A persistent "Book appointment" CTA on every page.
4. Answer the money question. Insurance, pricing and coverage, clearly.
5. Use social proof. Reviews, patient stories and outcomes (with consent).
6. Write for humans. Plain language beats medical jargon.
7. Be fast. Speed is trust; a slow medical site reads as unprofessional.
8. Design for mobile first. Most patients arrive on a phone.
9. Reduce form friction. Ask for the minimum; explain why you need it.
10. Keep it current. Outdated info erodes trust instantly.

Accessibility is non-negotiable

Healthcare serves everyone, including people with disabilities, and in many regions accessibility is a legal requirement. Meet WCAG AA: sufficient colour contrast, keyboard navigation, proper labels and alt text, and clear focus states. Accessible sites are also better for SEO and for every user, not just those who need assistive tech.

Measure what matters

Track the metrics tied to revenue: appointment requests, calls, insurance checks and completed bookings, not just pageviews. Then improve the weakest step. Small, evidence-led changes to your highest-intent pages usually beat a full redesign.

Intent over volume

A keyword with 200 searches a month from buyers beats one with 20,000 from browsers. Prioritise commercial-intent queries: "[service] agency", "[service] for [industry]", "[competitor] alternative", "how much does [service] cost". These are the searches people make when they're close to hiring.

The technical foundation