/ Design · 3D /

3D on a Website: When It Sells, When It Just Costs

An honest look at 3D visualization: which businesses gain conversion from it, what a GLB model weighs and costs, and how to add 3D without killing page speed.

3D on a Website: When It Sells, When It Just Costs
Pavlo Basha

Short answer: 3D on a website sells where the customer buys shape, function or effect — a product you want to hold (tech, furniture, footwear, packaging), a space (real estate, interiors, events), a process (equipment, installation, “how it works inside”). It does not sell where the decision is made by price and reviews (FMCG, simple services), and 3D only adds waiting seconds. The DEXA rule: 3D must load faster than the visitor can doubt it, and pay back with trust — not with “wow for wow’s sake”.

When 3D is a tool, not a decoration

A photo shows the product. 3D shows the client’s decision: “will this sofa fit my alcove”, “how does this mechanism open”, “what do I get a year after installation”. If an interactive answers those three questions, 3D pays off. If not — it is an expensive screensaver.

Who 3D actually lifts: four niche types

  1. The product you want to pick up. Electronics, shoes, tableware, cosmetics, auto parts, equipment. In a physical store the buyer lifts the item off the shelf — 3D gives the same tactile contact online: rotation, zoom, color/configuration switching. It builds the “not a pig in a poke” trust.
  2. Space and real estate. Houses, apartments, interiors, exteriors, object tours. The client buys something that does not exist yet — a walkable 3D model removes the main fear “what if I do not like it”. Architecture in 3D sells the idea before the foundation — examples live in our portfolio.
  3. Complex product/process. Equipment, installation, system internals: a model exploded by layers shows what no video can — how a boiler, a filtration station or a CRM screen works inside.
  4. Premium positioning. Investment firms, boutique brands, studios. Here 3D is the language of status: a site with an animated 3D scene reads as “this company has level”. The measurable effect is not day-one conversion but trust over time (see the Wrenhaven Capital case — Lighthouse 100 and doubled investor leads).

Who does not need it: three honest “no”s

  • Mass goods decided by “price + review” (pet food, household chemistry, socks): the audience does not spin socks, it reads “5 stars and delivery tomorrow”.
  • Sites with weak mobile traffic on slow networks in your region — if 70% arrive on 3G from old Androids, the 3D scene must weigh like a photo, or it should not exist.
  • “Because a competitor has it” — 3D for 3D’s sake on a five-page brochure never pays off. Fix the first screen and speed first (why your website does not convert: 12 reasons).

The technical price: weight, speed, money

Modern web 3D is not a “heavy plugin” but an optimized GLB model file plus Three.js or the native <model-viewer>:

ParameterGood practiceBad practice
Model weight per page1–3 MB (hero decorations even less)10–40 MB “as the render engine exported it”
Loadingonly when the block approaches the viewport, with a poster placeholdereverything at once on page open
Interactionrotation, zoom, hotspots, color/config switching“it spins prettily by itself” with no purpose
Mobilethe same experience, simplified lighting“does not work on phones, sorry”
Lighthouse impactdoes not move CWV (our projects hold 95–100)kills speed and rankings

On budget: one interactive product/object on a ready 3D model is an affordable entry (from a few tens of thousands of hryvnias); an architecture fly-through tour with hotspots is a mid-level project; a site-wide 3D hero (like our homepage) is part of a full build. Timelines and scope are on the animation & 3D website development page; the general pricing logic — how much a website costs in 2026.

Where 3D models come from and what the start costs

Three sources: (1) the manufacturer already has one (catalog GLB/OBJ for appliances and furniture are often free on request), (2) stock 3D models, (3) custom modeling from photos/drawings — the priciest path but your asset. DEXA starts with an audit of what already exists in 3D with you and your suppliers — often a client’s “3D content” turns out 60% ready.

The mini-test: will 3D pay off for you specifically

  • One product/object. Pick the most margin-heavy or the most “incomprehensible from photos” item.

  • One scene. Rotation + 2–3 hotspots (material, mechanism, configuration) — without “a whole world”.

  • A 2-week measurement. Compare time-on-page and cart adds/requests: 3D page vs plain page. Card conversion usually grows on “hard to picture” products. If it grows — scale the catalogue; if not — 3D is not your tool, which is also a good result for minimal money.

What we can do and where to see it

DEXA builds 3D scenes as part of the site, not a “glued-on widget”: viewport-based preloading, a Web Worker for heavy math, degradation to a static poster on weak devices. The living example: the hero scene on dexa-web.com — an interactive 3D composition that loads at ~100 Lighthouse — and Solomia Beauty, where 3D presentation carries the premium feel of the catalogue. Want the same for one of your products — send photos and CAD/dimensions; in two days we will assemble a prototype and show it on your screen.

The common mistake

Do not ask for “3D like Apple” without Apple’s content budget. Their scene is hundreds of hours of modeling, lighting and animation plus tens of millions in traffic that can appreciate it. The right SME brief is “3D like the competitor who lifted leads with it”: one model, three points of interest, speed without losses. That scale is exactly the one that pays off.

FAQ

Will 3D hurt my SEO through speed?

No, if the scene loads lazily, weighs 1–3 MB and does not block rendering. We always compare Core Web Vitals before and after — 3D must not move LCP/INP. If the budget cannot fit, we ship a poster image activated by a “view in 3D” click.

We have no 3D model of our product. Is that a verdict?

No. A good photo/drawing can be modeled; for furniture and appliances models often already exist at the manufacturer. The most expensive part is bespoke objects (a brand symbol, a “cutaway” of a complex machine) — and those usually sell the hardest.

Is a plain `<model-viewer>` enough without Three.js?

For “spin the product” — yes, and it is cheaper. Three.js is needed when 3D is part of storytelling (scenes, animation, scroll integration, exploded disassembly). The correct start is model-viewer on one product; scale the engine when the measurement shows an effect.

In short

3D on a website is not a “portfolio trend” but a trust tool wherever the client buys shape, function or space. Start with one product and a two-week measurement: if card conversion grows you have a new revenue source; if not — you saved the “wow” budget. Tell us your product name — we will honestly say whether 3D sells in your niche or it is just pretty money in the wind. Read also: web design trends 2026 and animations that convert.

corporate · Lite from $1,325 (₴55,000) · corporate · Standard from $2,145 (₴89,000) · corporate · Premium from $3,494 (₴145,000)

Need the same solution for your business?

A company website with bespoke design and motion: services, cases, team, languages and an admin panel. We build the structure around what your clients search for on Google.

  • From 5 pages: services, cases, team, prices
  • Lite — a custom layout without complex motion
  • Standard — bespoke design, GSAP, two languages, an admin panel
Discuss a project → Browse services

Get a free website audit in 48 hours

Speed, SEO, conversion — a 12-page report with a prioritized fix plan. No obligation, no spam: we show what to improve even if you never hire us.

We reply within 2 hours during working hours (Mon–Sat 9:00–20:00). Your contact is used to answer you only — no third parties, no mailing list.