24.4%
of online apparel is returned
Coresight Research's estimate for US online apparel, with size and fit the most-cited reason among the brands and retailers it surveyed (53%).1
Size and fit are the reason apparel brands most often give for online returns. Give your size logic, made-to-measure workflow or product team real body measurements from two smartphone photos — inside your own site or app.
The cost of a wrong size
Shoppers can't try clothes on online, so many guess — or buy two sizes and send one back.
Coresight Research's estimate for US online apparel, with size and fit the most-cited reason among the brands and retailers it surveyed (53%).1
The same report's estimate of what handling US online apparel returns cost in 2023 — before counting merchandise that can't be resold.1
Buying several versions of an item, such as two sizes, intending to return some, in NRF and Happy Returns' 2024 survey. Across US retail, 19.3% of online sales were expected to come back in 2025.2,3
1 Coresight Research, “The True Cost of Apparel Returns” (April 2023); survey of 100 US apparel brand and retailer decision-makers.
2 NRF and Happy Returns, 2024 Consumer Returns in the Retail Industry.
3 NRF and Happy Returns, 2025 Retail Returns Landscape.
One scan, four uses
The same scan feeds the storefront, the cutting room and the product team.
Feed the shopper's measurements into your size charts or recommendation engine. With confidence ranges, a shopper who sits between two sizes can be told so — instead of buying both.
The measurement set your patterns use, plus an OBJ 3D body model for pattern and CAD software — captured at home instead of at a fitting appointment.
Anonymised measurements from your own customers show how real bodies compare with your size chart and fit models — evidence for grading, fit-model choice and range decisions.
The 3D body model, joint positions and a body-part map for avatars, visualisation and virtual try-on pipelines — with or without a measurement list.
In the customer journey
Your site keeps the customer's measurements, so one scan answers the size question on every product after it.
Size chart, body measurements in cm
| Size | Chest | Waist | Sleeve |
|---|---|---|---|
| S | 88–94 | 76–82 | 61–63 |
| M | 94–100 | 82–88 | 63–65 |
| L | 100–106 | 88–94 | 65–67 |
| XL | 106–112 | 94–100 | 67–69 |
| You | 100.5 | 88.6 | 64.2 |
Illustrative — the size logic is yours
Who it's for
Integration
Two photos and a height in; your measurement set and an optional 3D body model out. The developer docs cover the request, outputs and data handling.
FAQs
Many size-related returns start with a shopper guessing their size from a chart. A two-photo scan replaces the guess with the shopper's own measurements, which your size logic compares with each garment's chart. Shoppers who sit between two sizes can be told so — instead of ordering both and sending one back.
Average accuracy is 97% in everyday clothing, across 18 standard measurements, tested on people photographed in varied clothing and compared with 3D body scans. For context, reliability studies of trained measurers using a tape report waist measurement errors of 1.3–6.5 cm, and the body itself changes through the day. Results are best in fitted clothing, with arms slightly away from the body. The best test is your own: try the free live demo, or run a pilot that compares scans with your size charts.
No. Fitted clothing gives the best results, but everyday clothes such as jeans and T-shirts work too; only coats and bulky layers come off. The scan runs inside your own site or app on any smartphone, with nothing to install.
Yes. Define the measurements your patterns use — any circumference or length between points on the body — and every scan returns that set, plus an OBJ 3D body model for pattern and CAD software. Customers can be measured at home instead of booking a fitting.
It can feed one. The scan returns body measurements, not a size: your own size logic, or the recommendation provider you already use, turns them into a size for each garment. Optional confidence ranges show when a shopper is close to a size boundary.
Nothing that identifies them. Photos are processed in memory and deleted once the result is returned, and a scan needs no name or email. The measurements go to your system, where you can save them to the customer's profile and reuse them across your catalogue.
Yes. The capture flow and the optional 3D viewer are white-label, or your team can build its own interface and call the API directly — see the developer docs.
Send us your size chart or pattern measurements, and we'll set up a measurement set to test against it.
Or write to us directly: info@snapmeasureai.com