How Much a Website Costs in 2026, and What Actually Drives the Number

Website design cost is the sum of eleven separate line items, not one number. For most business sites in 2026 the total lands between $2,000 and $30,000, and where you land inside that band comes down to page count, how many systems the site has to talk to, and who writes the content.

That range is useless on its own. So this guide does two things: it puts dated market figures next to each project type, and it gives you the arithmetic to place your own project inside them before you get on a call with anyone.

One promise about the numbers. Every figure here carries a source or a date. None of them are averages pulled from our own client files, because one agency's book isn't a market.

Here are the market bands as of July 2026. A single landing page runs $500 to $3,000. A small business site of five to fifteen pages runs $2,000 to $9,000. A custom corporate site starts around $5,000 and passes $20,000 once integrations and a second language are involved. A customer portal or web app begins at $25,000 and can reach $100,000 or more.

Those bands hold up against the guides that rank for this question. According to OuterBox's 2026 website pricing guide, a base small business site costs $3,000 to $10,000 and up, a custom lead generation site $10,000 to $50,000 and up, and enterprise work starts at $100,000. DesignRush puts custom mid-market builds at $5,000 to $30,000 and up. Forbes Advisor sits lower because it counts DIY builders in the same table, which is worth knowing before you compare two numbers that were never measuring the same thing.

Our own published prices sit inside those bands rather than above them: €1,500 for a landing page, €4,500 for a corporate site, €12,000 for a portal. Each is a starting price for a defined scope.

Honest caveat: any single number you read online, ours included, is the opening of a conversation. The spread inside each band is wider than the gap between bands.

Project type sets the floor, and nothing else moves the number as much. A landing page and a customer portal are different products that happen to arrive through a browser. The table below puts our published starting prices next to US market ranges for 2026, so you can see where a quote sits against the market instead of against the last quote you got.

Two rows deserve a note in the text, because table cells hide things. The landing page range starts at $500 because a stock template with your logo on it is technically a landing page. Ours starts at €1,500 and covers the conversion structure, the responsive build, and a form that reports into something you actually check, which is the scope described on the landing pages page.

The portal row is the one people argue with. $25,000 to $100,000 and up isn't padding: accounts, roles, permissions and data sync are software, and portal budgets and timelines behave like software budgets, not design budgets.

Most B2B buyers land in the middle, on a corporate website of 10 to 30 pages. If your scope sits between two rows, price the higher one. Scope creeps up, never down.

Same brief, two agencies, quotes that differ by a factor of five. It happens constantly, and it usually isn't one of them being dishonest. Three things drive the gap: the hourly rate behind the estimate, what each agency assumed was included, and who is expected to produce the content. Rates alone account for a threefold spread before anyone opens a design file.

Published 2026 rates give you the arithmetic. digitalapplied reports US agencies billing $125 to $300 an hour, with UK and EU agencies at $95 to $225. A 120-hour build at $95 an hour costs $11,400. The same 120 hours at $300 costs $36,000. Nobody padded anything.

Then come the assumptions, and this is where the gap usually lives:

• One quote includes copywriting. The other assumes you'll hand over finished text.
• Two revision rounds are in scope, or they're billable. Both are normal. Only one is written down.
• A CRM integration is a line item, or it's a change order you meet in week five.
• Post-launch fixes run for a month, or they end the day the site goes live.

Ask both agencies to itemize against the same list and a five-times gap often shrinks to under two. The remaining difference is usually seniority, and that one is a real choice rather than a trick.

Content is the quiet variable. It's the line both sides assume the other is handling, and it stalls more projects than any technical problem. Worth saying plainly: a cheap quote isn't automatically a bad quote. It's an incomplete one until you know which of these it left out.

A quote that reads "website design, $12,000" gives you nothing to negotiate with. A quote broken into line items tells you what to cut, what to defer, and what to bring in house. Eleven items cover almost every business site. Not every project needs all eleven, and working out which ones you can drop is most of the budgeting work.

Three of these have published prices you can check rather than guess. On our pricing page, copywriting starts at €20 a piece, motion and animation work starts at €100, and rush delivery is a flat 25% uplift on the project. That last one is the transparent version of what a lot of agencies handle with a shrug and a bigger round number.

Two more get underestimated almost every time. Content migration is priced per page and it's tedious, so it gets quoted either high or optimistically low, and the truth surfaces in week three. An additional language isn't a translation invoice either. It's a second content set, a second QA pass, and hreflang work, which is why multilingual builds sit above single-language ones at the same page count. We run five locales on this site, so that line is not theoretical for us.

Cut discovery and you haven't saved money. You've moved the cost to week six, when somebody notices the site doesn't do the one thing it was built to do.

You can get inside a band in about five minutes with two inputs: how many pages the site needs, and how many systems it has to talk to. Page count drives design and build. Integrations drive engineering. Everything else, content, languages, migration, prices as an add-on you can decide on later. The matrix below is our reading of the 2026 market ranges above, not a quote.

Read a cell as a starting band, then adjust it:

• Add roughly the cost of a page for every extra language content set.
• Add a migration line if you're moving more than about twenty existing pages.
• Add the rush uplift if your deadline is shorter than the timeline in the first table.
• Subtract if you're supplying finished copy and photography.

Then the step most people skip. Put 15% aside as a reserve. Not because agencies are unreliable, but because you'll change your mind about something once you see the design, and you'd rather have that money sitting there than have the change-order argument.

The worksheet below is the copy-paste version. Drop it into a spreadsheet, or straight into the email you're about to send three agencies. Making all three fill in the same rows is the fastest way to compare quotes that were never written to be comparable.

Every quote below the market floor is buying its price somewhere, and the useful question isn't whether corners were cut but which ones. The order is remarkably consistent: discovery goes first, then user research, then testing, then anything scheduled after launch. None of those are visible on launch day. All of them are visible in month six.

Here's what each cut buys you later.

• No discovery means the sitemap is a guess. You find out when marketing asks for a page type nobody planned.
• Template UX with no user flows means traffic arrives and leaves. The site works fine. It just doesn't sell anything.
• Skipped performance work means slow pages, and page speed feeds into ranking, so the SEO budget quietly ends up subsidizing the build savings.
• No CMS training turns every text change into a support ticket with a lead time.
• No post-launch plan turns the first plugin conflict into an emergency instead of a Tuesday.

The arithmetic turns at the rebuild. Rebuilding a site that was structured wrong costs more than building it right, because you pay for the new build, plus migrating content that was never structured for migration, plus the months the old site spent underperforming while you decided. chilledsites' 2026 comparison prices UK agency builds at £5,000 to £25,000 and up with £1,000 to £3,000 a year of maintenance on top. Pay a build twice and the cheap start was the expensive route.

Now the unpopular half, said out loud: plenty of businesses don't need the full build. A one-person consultancy with five pages and no integrations isn't the buyer this article is written for, and spending five figures there would be a bad call. The next two sections are about telling those two situations apart.

Industry barely changes the code. It changes the compliance work, the content volume and the integrations, and those are the three things that move a quote. A law firm site and a dental practice site can carry the same page count and land $6,000 apart, entirely because of what has to connect to what and who has to sign off the copy. The bands below come from the matrix above, not from a survey.

Regulated sectors carry a review cycle you can't compress. Medical and legal copy gets read by somebody whose job is risk, and that person is never on the project plan. Book two extra weeks and price two extra revision rounds.

Booking and scheduling integrations are the other multiplier. A hotel site that has to talk to a property management system and a channel manager is a portal wearing a brochure site's clothes, and it should be priced from the web portal column rather than the corporate one.

The general market bands still apply underneath all of this: $2,000 to $9,000 for a straightforward small business site, $5,000 to $20,000 and up once it's genuinely custom. Sector mostly decides where inside those you land, and whether the top of the band turns out to be your floor.

Buy-versus-build is a four-way choice now, and the only honest comparison runs over three years rather than at signature. An AI builder at $30 a month is $1,080 over three years before you count your own time. An agency build plus maintenance is a different order of magnitude. Both can be the right call, and which one is right depends on whether the site has a job beyond existing.

Published 2026 figures for the AI column: realistic business plans run €15 to €50 a month, with vendor plans spanning $0 to $200 a month across the market. Simple AI-assisted builds of five to ten pages are quoted at $300 to $3,000. On the agency side, chilledsites puts UK builds at £5,000 to £25,000 and up, with £1,000 to £3,000 a year of maintenance.

Two columns in that table carry no figures on purpose. We couldn't find a dated 2026 source for freelance totals or for classic builder totals, so those cells describe the shape of the cost instead of inventing a range. An undated number would look more authoritative and be worth less.

When the builder wins, and it genuinely does: you need something live this month, the content is five pages, there are no integrations, and nobody is spending ad money to send traffic there. Validate the offer first, buy the build once traffic justifies it. That's a sane sequence and we'd say so on a sales call.

When it loses: the moment the site has to connect to something, rank for something competitive, or be edited by three people without breaking. Those are engineering problems, and builders price the hosting rather than the engineering. One more thing a price comparison never shows you is switching cost. Content that lives inside a proprietary builder doesn't move cleanly, so ask what an export looks like before you need one.

A quote is a scope document with a number attached to it. Read the scope. Most disputes trace back to a line nobody wrote down, and the fastest way to find that line is to check what the quote says about revisions, content, integrations and the day after launch. Four questions catch most of it, and all four are fair to ask before you sign.

1. How many revision rounds, and what counts as one? A round is a consolidated set of feedback. Two people sending contradictory notes on a Friday isn't a round, it's two.
2. Who writes the copy? If the answer is you, ask for the deadline in the same breath. Content is the most common reason a fixed timeline slips.
3. What's a change order, and what's already included? Get the out-of-scope hourly rate in writing now rather than in week five.
4. What happens the day after launch? A month of fixes is normal. A year of updates is a separate agreement, usually $50 to $500 a month.

Then the add-ons that surprise people. Rush delivery is real work and gets priced as an uplift, ours is 25%. Stock photography, licensed fonts and premium plugins are pass-through costs and belong on the quote as such. Consent and analytics tooling is its own line in the EU. If you're weighing up several firms, our guide to choosing a web design agency puts these questions into checklist form.

Starting-at prices are real, and they're also the floor of a defined scope. What sits outside that floor is predictable: tax, the payment schedule, third-party licenses, and who owns what at the end. None of it is hidden. It's just rarely on the first page of a proposal, so ask before you sign rather than after.

Tax first. European quotes are usually net of VAT, and 19% to 21% on top of a €12,000 project is real money that needs to be in the same budget line. US quotes are normally pre-tax too. Find out which one you're reading.

Payment terms move cash flow more than most buyers expect. Milestone splits are standard, and so is a deposit. What varies is whether the final payment falls due at launch or after a fix window, and that difference is worth negotiating harder than a few percent off the headline.

Ownership is the one to get in writing. Who holds the code, the design files, the domain and the analytics account on the day the relationship ends? The answer should be you. If a proposal is quiet about it, the quiet is the answer.

And the running cost, which isn't a build cost but comes out of the same wallet. Hosting, maintenance and content updates sit on separate agreements from the build, which is how our own pricing is structured and how most agencies work. Market maintenance runs $50 to $500 a month depending on how much changes. A realistic launch timeline helps here too, because a compressed one is exactly what triggers the rush uplift.