Your fashion store gets the visit. Fit is where it loses the sale.
Apparel asks someone to commit to a thing they cannot try on. That is two questions, not one — will it fit me, and will it look like that on me — and the product page has to answer both before the tab closes. On most apparel storefronts it answers neither.
Source: Baymard Institute — apparel UX research, 1,765 hours of usability testing
Apparel carries two uncertainties. Most categories carry one
Someone buying a kettle has one question: is this the right kettle. Someone buying a dress has two, and they are independent — a garment can fit perfectly and still look wrong, or look exactly right in the photo and arrive unwearable. There is nothing to touch and nobody to ask, so both have to be settled by the page. That is a harder job than most storefronts are built for, and the category conversion rates line up with it.
- Arts & crafts5.23%
- Health & wellbeing3.57%
- Kitchen & home3.34%
- Pet care2.95%
- Sports & recreation2.12%
- Cars & motorcycling1.82%
- Fashion, clothing & accessories1.81%
Roughly half of kitchen & home — though up 33% year on year, from 1.36%
- Toys & games1.72%
- Food & drink1.47%
- Baby & child0.55%
Source: IRP Commerce — live ecommerce market data, July 2026
Fashion is not the floor, and that matters
Read the chart honestly: three categories sit below fashion. The claim is not that apparel is uniquely doomed — it is that apparel converts at about half the rate of a category selling comparably priced physical goods to the same people, and the gap is concentrated in a few specific moments. Those moments are findable. If you take one number from this page, do not take 1.81% as your target — take it as evidence that the gap is structural rather than a sign your store is broken.
- The size chart opens in a modal that is unusable on a phone, so it is never opened at all
- The model's height and worn size are not stated, so the photo answers nothing about fit
- A swatch changes the variant but not the photograph, so colour stays a guess
- A sold-out size still looks selectable, and the failure reads as a broken site rather than a stock problem
- Reviews mentioning fit exist, three screens below the point where the doubt arrives
- The returns policy is written as a warning, and it is the last thing read before the tab closes
19.3%
Online sales returned
Nearly one in five, across all online retail in 2025
15.8%
All retail returned
Down from 16.9% in 2024
Online returns run well above the all-retail rate, and apparel is the category that drives it. The instinct is to tighten the policy — and a stricter policy is itself an abandonment cause, because the shopper reads it as “I am carrying the risk of your sizing”. The way out is not a harsher policy. It is answering the fit question early enough that the order was right the first time.
Source: National Retail Federation — 2025 Retail Returns Landscape, with Happy Returns and UPS
Finding your fit leak in a week
This works with any replay and heatmap tooling, including a free one. The order matters more than the tool.
- 1
Split mobile out first
Fashion traffic skews heavily to phones and social. A blended product-page conversion rate averages a working desktop experience with a broken mobile one and reports something in between that describes neither.
- 2
Find your highest-traffic product template, not your favourite product
One template usually carries most of the catalogue. Fixing it fixes hundreds of pages; fixing the product you personally like fixes one.
- 3
Check whether the size guide is ever opened
Put a heatmap on that template and look at the size-guide control. If it gets almost no interaction while sessions still leave from the variant area, the guide is not being rejected — it is not being found.
- 4
Watch ten sessions that selected a size and left
Not ten random sessions. The ones that reached the variant picker and abandoned are the ones carrying the answer. Look for repeated swatch taps, a scroll down to reviews and back, or a tap on a photo expecting it to zoom.
- 5
Read your own returns reasons
If “too small” and “too big” dominate, the fit answer is wrong or invisible on the page — and every one of those returns had a matching visitor who did not risk it at all.
- 6
Move one answer, then measure it
Put the fit answer next to the variant picker rather than further down. Change that alone, and hold it against revenue rather than against a click-through rate.
| With DynoWeb | By hand | |
|---|---|---|
| Seeing the variant moment | Heatmaps are template-aware, so the size and swatch controls are tracked as controls and survive a theme update | Wire custom events onto the variant picker, then rewire after each theme change |
| Isolating the sessions that matter | Filter replays to sessions that reached the variant step and left | Scrub a random sample until you happen to catch one |
| Getting from finding to change | Each finding arrives with a specific storefront change you preview before it goes live | Write the brief, queue it, wait for a developer or a theme sprint |
| Knowing the fix paid | Measured against a holdout and reported as net revenue | Compare this month to last and hope the season did not move |
Where the other one wins: If you already suspect one template and you have an afternoon, go and look — a free replay tool and your own returns spreadsheet will get you to the first hypothesis without paying anyone. What you cannot get that way is the second, third and fourth cycle, which is where most of the money actually is.
How DynoWeb does it
Before you install
My store converts at 1.6%. Is that bad?
On the IRP figures above it is close to the category norm, so it is not evidence that your store is broken. It is also not evidence that it is fine — a category median is computed from a traffic mix, price point and market that are not yours. The number worth chasing is your own segment last year at equal volume, not anybody's benchmark.
We already have a size chart on every product page.
Almost every apparel store does, which is why the presence of one predicts very little. What predicts conversion is whether it is reachable at the second the doubt arrives, whether it uses garment measurements the shopper can compare against something they own, and whether it survives being opened on a phone. Check the interaction data on the control before assuming the content is the problem.
Would a virtual try-on or fit-quiz app fix this?
Possibly, and there are named-brand case studies published by those vendors showing large lifts. Read them as vendor marketing rather than independent research, and confirm the fit step is actually where your store loses people before you buy a tool that only helps if it is. If your drop-off is at shipping cost, a try-on widget changes nothing.
We barely get returns, so fit cannot be our problem.
A low return rate on a low sales volume is consistent with a fit problem rather than evidence against one — the people who were unsure did not buy, so they never returned anything. Returns only measure the shoppers who took the risk.
Find out which moment costs you the sale.
Install from the Shopify App Store and run it on your own product pages.