What a 3D website really costs (and where the money goes)

Nobody publishes this number, which is why the first call with a 3D studio is usually spent circling it. So here is the honest version: a 3D website costs more than a template site, less than most people fear, and almost all of the difference is decided by three facts you already know before you talk to anyone.

The three things that move the price

Whether the models already exist. If your product team has CAD files, or you have a 3D model from a manufacturer or an architect, the most expensive part of the work is already paid for. We optimise what you have — which is real work, but a fraction of modelling from scratch. If the scene has to be built from photographs and drawings, modelling and texturing become the single largest line on the quote.

How many distinct scenes there are. One hero scene that the whole site lives inside is one budget. A homepage scene, a product configurator and a walkthrough are three, and the third one is not much cheaper than the first — each has its own assets, its own performance budget and its own fallback frames. Scope creep on a 3D site is almost always a fourth scene that “would be nice”.

What proportion of the site is 3D. A brochure site with a striking 3D front door and twelve ordinary, fast pages behind it is mostly an ordinary site. A product experience where every page is a scene is a different kind of project, closer to a small game than a website, and priced like one.

Send a studio those three facts and you should get a range back, not a discovery-call invitation. If you get a form instead, that tells you something too.

Where the money actually goes

On a typical single-scene project, the scene itself is a third to a half of the total. People expect it to be more. The rest is the website — content, pages, CMS, hosting, the same things a good non-3D site needs — plus a category of work that 3D sites specifically pay for and template sites do not:

• Performance engineering. Getting a scene to load in two seconds on a mid-range Android is not a setting you turn on. It is a weight budget agreed up front, compressed geometry, textures in the right formats, and a renderer that loads after the page has already painted. This is the part that separates a 3D site that ranks from one that scores 40 on PageSpeed.
• Fallbacks. Every scene needs a version for phones, low-power laptops and people who have asked their system for reduced motion — usually a still frame captured from the real scene, plus the logic to decide who gets which. If a quote does not mention this, the site will not have it.
• Motion and interaction design. How the scene responds to scrolling, hovering and clicking is designed, not improvised. It is also where most of the memorable part lives.
• Testing on real devices. Not emulators. The phones your visitors have.

The trade-offs nobody puts in the proposal

Cheaper 3D exists, and it is usually worth exactly what it costs. A pre-made scene from a marketplace, an embedded viewer from a no-code tool, a video pretending to be interactive. Each is a legitimate choice for the right brief, and each fails in the same way: the moment you want it to do something specific to your product, you are paying to work around it.

Ongoing cost is real but small. A well-built 3D site has no special hosting requirement — it is static files and a CDN, like any fast site. What does recur is the occasional browser or library change that needs a small fix, and the cost of new scenes when the product line changes.

The biggest hidden cost is a scene that hurts conversion. If your site’s job is to be read quickly and acted on, a 3D layer can cost you more in lost sign-ups than it ever cost to build. That is a real trade-off and the honest studios will raise it before you do — we did, here.

What to ask for in a quote

A weight budget in kilobytes and a target Lighthouse mobile score, written down. A description of the fallback for phones. Who owns the models and the source when it ships. And a prototype milestone early enough that you are looking at a real scene in a real browser before most of the money is spent.

If you want a range for your own project, the three facts above are enough to start — that is the whole brief we need, and it is what our 3D website development page asks for.

Sources and further reading

• Three.js documentation
• Draco 3D data compression — Google
• KTX 2.0 texture container — Khronos Group