A lehenga is not a pair of socks

Most return-rate reports stop at one line: "fashion: 25–40%." That's like quoting the average temperature of India. True, and useless for planning anything.

Inside one fashion catalogue, a wedding lehenga and a three-pack of socks live in different universes. One comes back almost half the time. The other almost never does.

This post zooms in. Here are the 2026 return benchmarks for each fashion category in India, the one pattern that explains nearly all of them, and how to decide where to fix returns first.

(For the cross-category view, fashion versus electronics, beauty and the rest, see our companion post on India return-rate benchmarks by category.)

The benchmarks, category by category

India-specific data by fashion category is thin, so the chart below combines both kinds: the India 2026 figures where they exist, and global data where they don't. The global numbers aren't India's, but the ranking is strikingly consistent.

The one number to remember: ethnic and occasion wear runs at 30–45% RTO in India, but the best brands get it down to 20–28%, and the brands Base names there are the ones with occasion-delivery guarantees and virtual try-on (Base). Same category, same shopper, up to half the returns.

A caution on reading this: the India figures are RTO, not customer returns, and the global ones mix order-level and item-level rates. Compare the order of the bars, not the decimals.

The pattern: the more there is to imagine, the more comes back

Line the categories up and one rule explains almost all of it. Returns rise with how much the shopper has to imagine:

• Low imagination: socks, basics, innerwear, jewellery. Size is standard or forgiving, and the look is obvious from the photo. Returns stay in single digits or low teens.

• Medium: shirts, tees, kurtas in standard cuts. Some fit risk, little drape risk.

• High: dresses, denim, footwear. An inch or a half-size changes everything.

• Highest: occasion and ethnic wear. Drape, embroidery, colour against skin, length and fall, all judged from one photo of one model. Add a wedding date and the stakes double.

Notice the order: looks first, then fit. Fit matters most where the garment is fitted; looks matter everywhere, and most where the outfit is the point.

Category by category

Ethnic and occasion wear (sarees, lehengas, festive kurtas). The highest-risk category in India, at 30–45% RTO (Base). Shoppers can't picture drape or colour on themselves, and some will wear it once and send it back. Fix first: looks. Show it on the shopper's own body, and photograph the drape, not just the front.

Dresses and co-ords. The most-returned item in most global datasets, from about 36% in Italy to over 60% on European marketplaces (Prime AI). Fix first: looks, then length and waist fit.

Denim and trousers. Pure fit risk: rise, waist and inseam. US jeans returns ran above 45% during the 2025 holiday peak (Prime AI). Fix first: fit, with size advice per style.

Footwear. 20–35% in India, with the best brands at 12–20% thanks to better size guidance (Base). Fix first: fit.

Tops, tees and basics. Around 12% globally (EightX). Fix first: colour accuracy. "Looked brighter online" is the usual complaint.

Innerwear. About 13% for intimates brands in one large dataset (Prime AI, citing Loop), and often non-returnable in India for hygiene reasons. Fix first: fit.

Jewellery and accessories. 4–12% in India (Base). Low risk, but scale matters: earrings and necklaces are perfect for try-on, because shoppers want to see them with an outfit.

Where to fix returns first

Not every high-return category deserves your attention first. Plot your catalogue on two axes, return rate and share of sales, and the order becomes obvious:

Fix first: high returns, big sales. For most Indian women's brands that's ethnic sets and dresses. Try-on pays back fastest here.

• Test try-on: high returns, smaller sales. Denim and footwear. Pilot on your top styles.

• Protect: low returns, big sales. Tops and basics. Don't break what works; just fix colour accuracy.

• Leave alone: low returns, small sales. Jewellery and accessories, though they're great add-ons inside try-on.

Then do the work in the right order: looks first (let her see it on herself), fit second (size advice per style), and data third. Every try-on tells you which styles excite shoppers and which disappoint them, so use it for retargeting, merchandise planning and how deep to buy next season.

Flickd builds persistent AI avatars for fashion: virtual try-on on your Shopify store, AI Mirrors and QR Smart Tags in store. Book a demo to start with your fix-first categories.