3D for e-commerce4 min read

Where AR helps in e-commerce, by product category, and where it does not

Furniture, eyewear, fashion, beauty and electronics get very different value from AR. A category by category breakdown, with the weak cases included.

By CharpstAR · 21 Mar 2023 · Updated 22 Sept 2026

Part of Augmented reality for furniture retail: the complete guide

MAR2023

Where AR helps in e-commerce, by product category, and where it does not

"AR benefits e-commerce" is too coarse a statement to act on. The benefit depends almost entirely on what you sell, because AR answers one specific question well: what does this object look like, at its true size, in my space or on my body. If that is the question blocking your shoppers, AR is worth the money. If it is not, you are buying a nice viewer and calling it a strategy.

Here is the breakdown by category, weak cases included.

The table

Category	What AR answers	Strength of the case	The honest limit
Furniture and large home goods	Does it fit, does it suit the room, how big is it really	Strong	Model accuracy matters more than anywhere else; a wrong colour is a return
Eyewear	Does this frame suit my face, is the width right	Strong	Face tracking is good, but frame weight and nose fit are still unknown
Heating, appliances, fittings	Scale in the room, clearance, how it looks installed	Strong	Buyers often need installation advice more than a preview
Fashion, garments	Roughly how a cut and colour look on a body	Weak to moderate	Drape and true size are not simulated reliably
Beauty	Shade on skin, in the shopper's own light	Moderate	Colour accuracy varies by camera and lighting; texture and finish do not show
Consumer electronics	Physical size of a device, how it sits on a desk	Moderate	Most purchase anxiety is about specification, not appearance
Small commodity goods	Almost nothing	Weak	Nobody needs to place a phone cable in their kitchen

Furniture and large home goods: the strongest case

This is where AR does the most work, because the blocking question is literally spatial. A shopper who cannot tell whether a three-seat sofa leaves room for the door either measures carefully or does not buy.

Sweef sells modular sofas, where the number of possible layouts makes photography impossible and the fit question is acute. Contura sells wood-burning stoves to people who have mostly never owned one and cannot picture the scale. In both cases, placing the real-size model in the room is not a flourish, it is the answer to the objection.

Returns matter here too. The National Retail Federation, with Appriss Retail, put the online return rate at 17.6 percent of sales in 2023, 247 billion dollars. Furniture returns are expensive to process, so avoided returns are worth more per unit than in most categories.

Eyewear: strong, and the technology is mature

A frame is rigid and a face is tracked well by modern phone hardware. Shopify's write-up cites Fittingbox research from 2025 that 54 percent of customers who were uncertain about buying glasses online wanted to check the fit of the frame, and that 29 percent of glasses shoppers have used virtual try-on at least once. That is a well matched problem and solution.

What AR still does not tell the shopper is how the frame feels, how heavy it is, or how it sits on their nose bridge over an afternoon. Those remain reasons for returns.

Fashion: the weakest of the "obvious" cases

Garment try-on is a styling preview. The camera does not measure shoulder width or arm length, fabric drape is approximated, and the output shows a colour and a cut on something roughly like the shopper's shape. That has value for confidence and sharing. It does not solve size returns, and vendors who claim otherwise are ahead of the technology.

If you sell clothes, the higher-yield moves are a proper size chart, fit notes drawn from your own returns data, and 3D for the accessories in your range.

Beauty: moderate, with a lighting problem

Shade matching in AR works well enough to narrow a choice and badly enough to be a bad final authority. The shopper's camera and room lighting change the result. Shopify cites Benchmarking Company data that products judged hardest to buy online are facial complexion products at 41 percent, self-tanners at 38 percent and body makeup at 38 percent, which is a fair map of where try-on is most wanted and least reliable.

Electronics: moderate, for size only

AR answers "how big is this monitor on my desk" and nothing else. That is a genuine question for large or oddly shaped devices and irrelevant for most of the category. A good 3D model with dimensions may do the same job without the camera.

How to decide for your own catalogue

Ask three questions about a product. Is size or spatial fit a reason people hesitate? Is the object rigid, so a model can represent it honestly? Is the order value high enough that a percentage point of conversion pays for the asset? Two yeses make it worth testing. Three make it worth doing at scale.

Our pricing starts at 10 dollars per product per month on the Basic tier, twelve-month term with a minimum of 100 variants, and 20 dollars per model per month on Value with a twenty-four month term. That arithmetic is easy for a 900 euro sofa and hard for a 19 euro accessory, which is really the same advice as the table above.

The cheapest way to find out where your catalogue sits is to test one product. Send us one and we will build the model for free, or see what we build if the answer is already clear to you.