VISUOSOFTS · AUGMENTED REALITY
Furniture is the product category AR was made for: big, expensive, and held back by two questions — will it fit, and will it look right? A customer who places a true-to-scale sofa in their living room answers both in thirty seconds. This buying guide covers what UK furniture and homeware brands need to build, what it costs, how the 3D models get made, and how to know it worked — honestly priced, no hype.
WHY FURNITURE IS AR'S BEST-FIT VERTICAL
Most AR retail use cases are nice-to-have. Furniture AR is need-to-have, because it attacks the two most expensive problems in the category: purchase hesitation and returns. A £1,200 sofa bought online and returned costs the retailer delivery both ways, inspection, and often a discount to resell as ex-display. Anything that lets the customer check fit and look before ordering earns its keep fast.
The psychology helps too: furniture buyers already measure rooms and photograph spaces — AR slots into behaviour that exists. Compare that with convincing someone to point their camera at a £12 candle. The friction-to-reward ratio in furniture is the best in retail AR, which is why it tops every serious AR in retail roundup.
One honest caveat: AR will not fix a bad product page. If your dimensions, materials and delivery information are vague, customers will still hesitate — they will just hesitate in AR. The retailers who get the most from room placement are the ones whose product data was already solid.
WHAT YOU ACTUALLY NEED: 3D MODELS THAT LOOK RIGHT
The AR experience is the easy part. The 3D models are the project. Every product you want customers to place in their rooms needs an accurate, good-looking, lightweight 3D model, and there are three ways to get one:
- From CAD files. If manufacturing already produced CAD, this is your cheapest route — the geometry exists, it just needs converting, cleaning and texturing for real-time use. Always ask your factory first; many retailers pay for models they already own in another format.
- Photogrammetry / 3D scanning. Photographing the physical product from dozens of angles and reconstructing it. Good for existing catalogue items with no CAD, and for capturing real fabric textures. Needs a controlled setup for consistent quality.
- Modelled from scratch. A 3D artist builds it from photos and dimensions. Most flexible, most expensive per unit, and the route for anything with complex materials or configurable options.
Whichever route you take, the models must be true to scale — a sofa rendered 5% too small will look perfect in AR and wrong in the customer's room, which is worse than no AR at all. Insist on a scale-verification step in the pipeline: model against stated dimensions, check in-app against a known reference, sign off per SKU.
Do not model the whole catalogue on day one. Prioritise: your top 20% of SKUs by revenue, the products with the highest return rates, and anything where "will it fit" is the known objection. Thirty excellent models beat 300 mediocre ones — expand once the numbers justify it. Our 3D product models for ecommerce AR piece covers the production pipeline in full.
Materials matter more than polygon counts. Customers judge a fabric sofa by the fabric — get the texture, sheen and colour right under warm indoor lighting, because that is where the model will be viewed. A geometrically perfect model with flat, lifeless materials reads as "video game" and undermines trust instead of building it.
ROOM PLACEMENT: WEBAR VS A NATIVE APP
You have two delivery routes, and the right one depends on where your customers already are.
WebAR — opens in the phone browser from a link, QR code or product-page button. No download. Right when AR is a feature of the buying journey: a "View in your room" button, a QR code on the swing tag. Friction is minimal, which matters — most customers will try AR once per purchase and will not install an app for it.
A native app earns its place when AR is a destination rather than a feature: saved rooms, multiple products placed together, sharing a room design with a partner, or integration with a loyalty programme. If you are asking customers to come back repeatedly and build something, an app's persistence and performance justify the download. For a single "will it fit" check, it does not.
Most UK furniture retailers should start with WebAR on the product page and graduate to an app only when usage data demands it. Under the hood, both routes now use the same plane-detection fundamentals — the phone finds the floor, the model sits on it at true scale. Phones with LiDAR (recent iPhone Pro models) place more accurately in tricky lighting, but a well-built experience works acceptably on standard phones, which is most of your audience.
COST BANDS: WHAT UK RETAILERS TYPICALLY PAY
Honest ranges from the UK market, so you can budget before you brief:
- 3D models: £300–£800 per SKU for straightforward products from existing CAD or clean photography; £800–£1,500+ for complex, configurable or highly detailed pieces modelled from scratch. Configurators (fabric/colour/leg options) multiply the work — each variant needs its materials built.
- WebAR room-placement experience: £5,000–£15,000 for a polished product-page integration covering your launch SKU set, including the viewer UI, placement logic and analytics. Multi-product room planning pushes it higher.
- Native app with AR: £40,000–£100,000+ depending on everything else the app does — at that point AR is a feature of an app project, not the project.
- Ongoing: new-SKU modelling as the range changes, plus platform/hosting fees that are usually modest. Budget per-season model production like you budget photography.
The cost driver nobody budgets for is product data hygiene: accurate dimensions, weights, materials and variant structures for every SKU going into AR. If your PIM data is messy, clean it first or pay your AR vendor to work around it — the second option costs more. Our WebAR cost guide breaks down where the money goes in more detail.
HOW TO MEASURE WHETHER IT WORKED
AR measurement should be boring and commercial. Track these, in this order:
- Conversion rate, AR users vs non-AR users on the same products — the headline number. AR users are self-selecting, so compare carefully over a proper period.
- Return rate on AR-visualised SKUs. "Will it fit" returns should fall. If they do not, your scale accuracy or product data has a problem, not your marketing.
- Engagement depth: placements per session, time in AR, shares. Useful for optimising the experience, not for justifying the budget — keep the money metrics primary.
- Time to decision: do AR users buy faster? Shorter consideration means less remarketing spend per sale.
Run it as a proper comparison, not a victory lap. Launch AR on half your top SKUs, hold the other half back, and compare over eight to twelve weeks. Seasonality, promotions and stock levels all move the numbers — a clean test design is what separates "AR works" from "we think AR works". For the returns side of the equation specifically, see our analysis of whether virtual try-on actually reduces returns.
QUESTIONS TO ASK ANY AR VENDOR
- How do you guarantee true-to-scale placement, and what is your verification process per SKU?
- Who owns the 3D model files, in what format are they delivered, and can we take them to another vendor?
- What is your per-SKU cost for new products after launch, and what is the turnaround time?
- WebAR or native — which are you recommending for us, and why? (If they only sell one, get a second opinion.)
- What analytics come as standard, and can we see AR-attributed conversion in our own tools?
- Show us a furniture client and their measured results — not a demo reel, numbers.
If a vendor cannot answer the scale question precisely, walk away. In furniture, wrong scale is not a minor bug — it is the entire value proposition failing.
FURNITURE AR — FAQs
How much does AR cost for a furniture retailer?
Budget £300–£1,500 per SKU for 3D models depending on complexity, plus £5,000–£15,000 for a WebAR room-placement experience integrated with your product pages. Most UK retailers start with 20–50 priority SKUs, putting a sensible pilot in the £15,000–£40,000 range all-in.
Do customers need an app to view furniture in AR?
Not necessarily. WebAR lets customers tap "View in your room" on the product page and place the item with no download. A native app only earns its place when you want persistent features like saved rooms and multi-product room planning.
How accurate is AR room placement — will the sofa really fit?
When models are built true-to-scale and the experience is properly calibrated, placement is accurate to within a few centimetres on modern phones — good enough for fit decisions. The risk is not the technology, it is sloppy modelling: insist on per-SKU scale verification.
We have CAD files from our manufacturer. Can they be reused?
Usually yes, and it is the cheapest route to 3D models. CAD geometry needs converting, cleaning and texturing for real-time AR use, but starting from CAD typically cuts modelling cost significantly versus building from scratch. Ask your factory what formats they can supply.
Does furniture AR actually reduce returns?
It reduces fit-related returns — the "it looked smaller online" category — when scale is accurate and product data is clean. It will not fix returns caused by quality issues, late delivery or vague product pages. Measure return rate on AR-visualised SKUs against a control group to see your real number.
How many SKUs should we launch with?
Start with your top 20% by revenue plus the products with the highest return rates — typically 20–50 SKUs. A tight library of excellent models beats a broad library of mediocre ones, and expansion is easy once the pilot numbers justify it.
