3D on a business website: when it pays, when it's a toy, and what it costs to do properly

Our homepage is a galaxy rendered in text characters. A client's site is a browser game running on an engine we wrote ourselves. So we are the last people who will tell you 3D on the web is a gimmick. We are also the people who will tell you when it is.

When it pays

Three-dimensional content pays when the thing you sell is three-dimensional and the decision depends on it.

• Cars, bikes, furniture, watches, kitchens. A configurator that shows this colour on this trim in this light answers the question a photo gallery cannot. The buyer stays, plays, and arrives at the showroom already decided.
• Property that does not exist yet. A walk through an unbuilt apartment sells off-plan better than a render on a billboard, and it lets a buyer in Dubai see the view from the fourth floor in Amsterdam Zuid.
• Products with a mechanism. An exploded view that the visitor rotates explains an industrial valve or a hearing aid faster than a page of text — and it keeps the engineer who is doing the evaluation on your site instead of your competitor's.
• Brands whose product is the feeling. Fashion houses, spirits, hospitality. Here the experience is the message, and a scroll-driven scene that stages the product can carry a launch. This is also where most of the budget gets wasted, so read the next section twice.

When it is a toy

• When it replaces the content instead of framing it. A visitor who has to fly through a tunnel to find the opening hours will not find the opening hours.
• When it does not work on the phone in the visitor's hand. Most of your traffic is mobile; a scene tuned on a designer's MacBook that stutters on a mid-range Android is a scene most people experience as broken.
• When crawlers cannot read the page underneath. A canvas is invisible to search engines and to AI answer engines. If the 3D layer is the site, the site does not rank.
• When it is there because the agency wanted an award.

The rules we build by

The show never blocks the content. Every 3D experience we ship is a layer on top of a page that is complete without it: rendered on the server, readable by every crawler, usable with the layer switched off. The galaxy on our homepage sits above nine crawlable case-study pages fed from the same data. Switch off JavaScript and you still get an agency website.

A performance budget, in numbers, from day one. Frame time, memory, download size, battery. Our text galaxy renders the 3D scene at 60 fps but only converts it to characters at 22 fps, because the eye does not need more and the phone does. Everything pauses when it is off-screen or the tab is hidden. We wrote up the engineering — the numbers are real.

Progressive, not all-or-nothing. A capable device gets the full scene; an older one gets a lighter version; a device with reduced-motion switched on gets stills. Nobody gets a spinner.

Own the engine or know it cold. For WAV JPG we built a browser game on our own engine, because a Unity export would have been 30 MB of someone else's decisions. For a configurator or a product viewer, Three.js is the right tool and we have shipped enough of it to know where it breaks.

Design the fallback first. The stills, the poster image, the sequence of screens without motion — if that is not good, the 3D version will not save it.

What it costs

A product viewer or a small scroll-driven scene on an otherwise conventional site starts around €6.000–€12.000 on top of the site. A configurator with real options, pricing and a lead form is €15.000–€40.000 depending on the product data. A full interactive world — a game, a virtual space, an engine-level build — is scoped per project and rarely under €25.000. Included in every one of those: the performance budget, the fallback, the crawlable page underneath, and a device lab test on real phones, not simulators.

Those are honest numbers, and they are why we say no to 3D more often than yes. When the product deserves it, nothing else comes close. Tell us what you sell and we will tell you which side of the line it is on — before anyone draws a polygon.