Size & fit recommendation

Recommend the right size, every time

StyleProfile’s fit engine combines what a shopper tells us about their body with each garment’s own measurements, and returns a size while they are still deciding. It runs on the same catalogue sync as try-on, so the size answer arrives in the same panel as the render.

Get a demo Try it live

A tailor's table from above: a cloth measuring tape curled beside a folded camel trouser leg, with chalk, scissors and pins in a dish. — The engine reads the garment's own measurements, not its size label.

The size problem

Fit is the reason clothes come back

Roughly half of apparel returns are fit-driven — the garment arrived and it was the wrong size, not the wrong item. That return is paid for twice: once in shipping and handling, and again in the margin lost on a garment that has to be restocked or discounted.

Cardboard shipping boxes and a grey mailing bag stacked by a doorway, one box open with a folded knit half lifted out.

20–30%of online apparel orders are returnedIndustry aggregate (NRF, Narvar, Happy Returns)

~50%of those returns are driven by fit and sizingIndustry aggregate (NRF, Narvar, Happy Returns)

15–30%reported return reduction where virtual try-on is deployedPublished industry range, cited per page

Ranges, not promises. Your own result depends on category, catalogue quality and how early shoppers reach the fit step.

How the engine works

Two inputs and a feedback loop

The engine compares a shopper’s body data against the measurements of the specific garment, then learns from what happens to the order. Nothing here depends on the shopper trusting a size label.

Body data from the shopper

Height and weight, a usual size in a brand they already trust, or the measurements they know. Every field is optional, and the recommendation says what it used.

Input one

The garment’s own measurements

Your size chart, read per garment rather than per size label. An M from one brand and an M from another are treated as two different garments, because they are.

Input two

Return outcomes, fed back

Kept and returned orders tell the engine which recommendations held up. Sizing that runs small on a particular cut gets corrected for the next shopper.

Feedback

Per garment, not per label

An M in one brand is not an M in another

The engine reads each garment's own measurements, so two mediums that fit differently are treated as two different garments — because they are.

Where it appears

Three places a shopper meets the recommendation

The recommendation shows up at the three moments a size decision is actually made: on the product page, inside the try-on panel, and in the cart before checkout.

Product page

A size pill next to the size selector

The recommendation sits inline with the sizes a shopper is already looking at, with one line explaining what it is based on. It is a suggestion, not a lock: any size stays selectable.

Product page

Satin midi skirt · $96

XSSMLXL

Your size: L — based on your height, weight and this garment’s measurements.

Try-on panel

A fit note beside the render

While the shopper is looking at themselves in the garment, the panel says which size that render represents and what the alternative would change. Fit stops being an abstract chart.

Try-on panel

Size L looks right for this cut.

Take M if you want it closer to the body. This render is showing L.

Cart

A nudge before the order is placed

If the size in the cart disagrees with the recommendation, one quiet line offers the swap. No modal, no interruption, and never more than once per item.

Cart

You picked M.

Shoppers with your measurements usually keep L in this style. Switch to L

Illustrative interface, drawn here in plain HTML — not a screenshot.

With try-on

Better together, on purpose

Try-on answers “does this look right on me”. The fit engine answers “which size do I order”. Shoppers ask both in the same breath.

One set of measurements
The size shown beside a render is the size that render represents.
One feedback loop
Every kept or returned order teaches the engine about that specific cut.

How the render works Model the return impact

Trouser · true to size
Skirt · size down
Knit · true to size
Heel · half size up

Illustrative guidance per garment — the size answer belongs beside the render, not on a separate size-chart page.

FAQ

Size and fit questions

Accuracy tracks how much the engine knows. With a garment size chart and a few body inputs from the shopper it performs well on standard fits; with neither, it falls back to your own sizing guidance instead of guessing. Every recommendation shows what it is based on, and a shopper can always overrule it.

No. A shopper can give height and weight, name their usual size in a brand they know, or skip the question entirely and still use try-on. More input produces a sharper recommendation, but none of it is mandatory.

Yes, and that is the normal case for multi-brand stores and marketplaces. The engine reads each garment’s own measurements rather than trusting the size label, so an M from one brand and an M from another are handled as different garments.

Next step

Answer the size question before the order ships

Try the live demo now, or take 30 minutes with us and we will run try-on and fit on your own catalogue.

Try it live Get a demo

Cancel anytime · No long-term contract · Live on your store in minutes