Pourquoi les vêtements reviennent comme « n'allait pas » et comment trouver la vraie cause

8 octobre 2026

Most “didn’t fit” returns are decided on the product page, before the order is placed. To reduce them, fix two places: the return form, so you learn what actually went wrong, and the product page, so shoppers can answer “will it fit me?” and “will it suit me?” before they buy.

If you run a clothing store on Shopify, you have probably seen it in your returns report. Reason after reason says “didn’t fit” or “wrong size”. It confirms there is a problem, but it does not tell you what to change.

This guide walks through how to turn vague return reasons into something you can act on, and how to set up your product pages so fewer wrong sizes get ordered in the first place.

Do not quote a clothing-return rate as if NRF measured dresses. The 2025 Retail Returns Landscape (NRF and Happy Returns) estimates 19.3% of online sales will be returned in 2025 — all categories, not apparel. Retailers expected 17% of holiday sales back. Use those as the floor for a clothing store, not the ceiling.

Apparel is worse. Returnalyze’s 2025 peak-season data (volume) put women’s clothing around 34–39% depending on the week, and men’s around 21–25%. Those are holiday weeks, not your shop. They still tell you the category is not a 5% problem.

On the fix: Zalando’s size-and-fit team has said size-related issues are around a third of returns, and items with size advice had 10% fewer size-related returns than similar items without it. That is Zalando’s own catalogue, not yours. It is still the cleanest public proof that answering “will it fit me?” moves the parcel.

Shopify’s own data: shoppers who can visualize products convert up to 94% better. That is the category of visualization (3D / AR), from Shopify’s 3D Ecommerce writing and changelog (11 May 2022). It is not a promise about AI Frame, and it is not a return-rate claim. It is the look half of the same physics.

Why “didn’t fit” tells you so little

“Didn’t fit” is the shopper’s summary, not the cause. The same reason can mean the waist was tight, the sleeves were short, the fabric had no stretch, or the shopper simply picked the wrong size from your chart.

Size is a big share of clothing returns. Zalando has said that about a third of its returns are size-related. Coresight Research put size and fit behind 53% of apparel returns in 2023.

There is also a second question hiding in many “didn’t fit” returns: the piece fit, but it did not look the way the shopper expected on them. Shoppers often tick “didn’t fit” for that too, because it is the closest option.

So every clothing purchase depends on two separate questions:

Will it fit me? A numeric question: which size matches my body.
Will it suit me? A visual question: how will this look on me.

A size chart answers the first. A product photo on a model answers neither for the person buying. When either question stays open, some shoppers guess, and some order two sizes and send one back.

Step 1: Ask better questions on your return form

You cannot fix what you cannot see, so start with the return form. Replace the single “reason” field with a few quick choices that take under half a minute:

Which size did you order, and which size do you think would have fit? This alone tells you whether the item runs small or large.
Was it too small, too big, or did it fit but not look right? The last option separates the look problem from the size problem.
Where was it off? Only show options that make sense for the product: waist, hips, length and rise for trousers; chest, shoulders and sleeves for tops.

Review the answers per product each month. When one product keeps coming back as “too small at the chest”, the fix is usually its size chart or a fit note on the page, not a new policy.

This step tells you what is wrong. Steps 2 and 3 stop it from happening.

Step 2: Answer “will it fit me?” right next to the size picker

A size chart only helps if shoppers see it and can use it. Many stores hide the chart in a tab, a footer page or a PDF in the description. Others show it, but expect the shopper to measure their chest with a tape measure first. Most people will not do that while shopping on their phone.

A few changes make the fit question much easier to answer:

Put the size chart link directly next to the size buttons. That is the moment the shopper is deciding, so that is where the answer belongs.
Use one chart per fit, not one chart for the whole store. An oversized hoodie and a slim shirt in size M are not the same garment. Reusing one chart is a quiet source of returns.
Show garment measurements, not only body measurements. A shopper can lay their favourite hoodie flat and compare. That is often easier than measuring themselves.
Offer a size from height and weight. People know these numbers without measuring. A size recommender can estimate their measurements and match them to your chart, then show their size next to the picker, with a short reason why.

The goal is that every shopper can see a size meant for them, without leaving the page and without a tape measure.

Step 3: Answer “will it suit me?” with a try-on

Once the size is clear, the remaining doubt is about the look. Will this colour work with my skin? Is this cut flattering on me, or only on the model?

A virtual try-on lets shoppers upload a photo and see the product on themselves before buying. It answers the visual question that a size chart never can.

Two things to keep in mind:

A try-on shows the look, not the size. An AI image does not prove a size will fit. That is why try-on works best next to a clear size answer, not as a replacement for one.
Not every shopper will upload a photo. Many will, many will not. The size answer from step 2 should work for everyone, with or without a photo. The try-on is the extra layer for shoppers who want to see it on themselves.

If you sell in Europe, also check how a try-on app handles photos. Shoppers should give consent, and their photo should only be used to create the try-on image, not to guess their body measurements.

Shoes, swimwear and gifts need a different approach

The steps above work for most clothing, but three cases need extra care

Category	What changes What to do
Shoes	Height and weight say nothing about foot size, and half a size matters.	Use a shoe size chart based on foot length in centimetres, with EU, UK and US sizes side by side.
Swimwear and lingerie	Try-on is often not suitable for these products, so the chart carries the whole decision.	Make the size chart extra clear, with cup and band sizes where relevant, not only S, M and L.
Gifts	The person buying is often not the person wearing it, so a selfie try-on shows the wrong body.	Let shoppers enter the wearer’s height and weight, and make exchanges simple.

What this looks like on a product page

At AI Frame, we built our Shopify app around these two questions, in the order shoppers need them:

Their size next to the size picker. Shoppers enter their height and weight, and see their size with a short explanation of why it fits. No tape measure, no photo needed.
Their look on their own photo. The virtual try-on shows the product on the shopper, for anyone who wants to see it before buying.
Your size chart and their Fit Profile behind both. The recommendation reads your own chart, and shoppers can save their details to use again on other product.

One of our clients, a streetwear label selling heavyweight oversized pieces, saw about 6,900 size estimates in September. Shoppers saved about 960 Fit Profiles, and 438 items were bought by shoppers who had seen a size recommendation or the size chart first.

AI Frame is free to start, needs no theme code and has no per-order fees. If you want to see it on your own store first, we’re happy to set it up for free on one of your products.

Install AI Frame on the Shopify App Store · See how it works

Retour au blog