What an architecture firm should pay for a website in 2026

As a rule of thumb, an architecture practice in Europe pays between 4,000 and 25,000 euros for a website in 2026. A template set up in-house costs a few hundred a year. A custom build from one designer-engineer sits mid-range. Agency work with a team runs higher. Scope drives the number, not craft.

Where do those bands come from?

They come from quotes I see circulating in this market, and they are rules of thumb rather than survey data. There is no reliable public dataset for what architecture practices pay for websites, so any article that gives you a precise average has made it up. What can be checked is what each band contains, and that is the useful part.

Platform template run by you: under 1,000 euros a year

Squarespace, Cargo and similar platforms publish their prices in the open; plans run from a few euros to a few tens of euros a month on the Squarespace pricing page. Add a domain and you are still under 500 euros a year. The cost that does not appear on the invoice is your own time: choosing the template, cropping images, writing the project texts, and then living with the layout decisions someone else made for a different kind of business.

One designer-engineer, custom: 4,000 to 15,000 euros

One person designs the site and writes the code. There is no account manager, no strategy deck, and no handover between a designer and a separate developer, which is where most of the cost and most of the drift lives in larger teams. At the lower end of this band you get a homepage, a practice page, a contact page and six to eight project case studies. At the upper end you get a larger archive, a content system your office can use, two languages, or a build that has to preserve the search rankings of an old site.

Studio or agency with a team: 15,000 to 60,000 euros

You are paying for a producer, a strategist, a designer, a developer and a project manager, plus the overhead that keeps them in a room together. That structure earns its money when the site is genuinely large: a practice publishing forty projects in three languages with a press archive and a jobs board is a coordination problem before it is a design problem. MVRDV is the public example everyone in this trade knows, and the scale of that archive is the reason it is not a five-thousand-euro job.

What actually moves the price?

Project pages. Almost everything else is fixed cost. A homepage is one design problem solved once. Eight project case studies, each with a brief, a set of photographs in a deliberate order, drawings, credits and press, is eight rounds of editing and layout, and the editing is the slow part. If you send me forty projects and we publish twelve, the twelve will do more for you than the forty and the invoice will be smaller.

After that, in rough order of impact: whether your photography is ready or still to be shot; whether someone in the office will write the project texts or the writing is bought in; how many languages; whether an old site has to be migrated without losing the URLs that already rank; and whether you want anything unusual, such as an assistant that answers first enquiries.

What does the money buy that a developer can see?

Speed, first. Google documents 2.5 seconds as the threshold for a good Largest Contentful Paint, measured at the 75th percentile of real visits. An architecture site is mostly photographs, so hitting that on a phone on a train is engineering work: image formats, sizing, loading order, and a hard budget per image. Template sites usually fail it by default because nobody set the budget.

Second, typography and grid that match your drawings rather than the demo content the theme shipped with. Third, project pages structured the way you would present in a pitch, so a planning officer or a private client reaches the answer without scrolling through a mission statement. I wrote about the way the strongest sites sequence that material in what makes an architecture website win awards.

What should I budget that is not the website?

Photography is normally the largest line and it is not part of the web fee. Commission it before you commission a site, and check the licence covers web use in perpetuity. Then: typeface licensing, which most foundries price per monthly pageview rather than as a single purchase; copywriting, if nobody in the office wants to write; hosting and a domain, which for a well-built static site is small money; and the hours your own team spends gathering drawings, credits and consent from clients. That last one is the cost people forget, and it is the one that delays launches.

Is a template ever the right answer?

Yes. If you are two years old, have three completed projects and no professional photography, a template is the correct decision and a custom site is not. Spend the money on the photographer. Come back when you have eight buildings worth showing and enquiries you are losing. The same logic applies in other trades; I made the case for spending on proof before polish in the home builder site that books site visits.

How do I compare two quotes without reading code?

Ask six questions and compare the answers rather than the totals.

One, is the price fixed against a written page list, or is it an estimate that bills by the hour. Two, who writes the project texts. Three, what happens to my photographs, and what performance target are you working to. Four, who owns the code and where does it run after launch. Five, what does adding a project cost me in a year, in money and in effort. Six, what exactly is in the handover.

A quote that answers those in a page is usually from someone who has done this before. A quote that is a single number and a mood board is a negotiation, not a plan.

What does it cost to keep running once it is live?

Hosting for a static architecture site is close to nothing, in the range of zero to a few tens of euros a month. Budget separately for change: as a rule of thumb, ten to fifteen per cent of the build cost per year covers adding projects, small edits and the occasional structural change. If a proposal has no answer for year two, that is a cost that will arrive later as a rebuild.

When is the wrong time to spend?

When the photography is not shot. When the practice is mid-rebrand and the name or the logo may still move. When you have no capacity for the work a working site brings, because an enquiry form that nobody answers within a day is worse than no form. Everything else is a scheduling problem rather than a reason to wait.

If you want to see what the money buys in practice, the scope, the timeline and the two paths are set out on my page for web design for architects.