A Fit Analytics alternative that starts with the garment

What is Fit Analytics?

Fit Analytics is a size recommendation service that retailers install on their own product pages. Its shopper-facing component is called Fit Finder, and it asks the shopper for details about themselves before returning a recommended size for the item being viewed (fitanalytics.com, retrieved 27 July 2026). Snap Inc. announced its acquisition of the company in March 2021.

Two things follow from that design, and both are descriptions rather than criticisms. First, the tool lives on the retailer's site, so you only meet it in shops that have bought it. Second, it reasons about the person: the inputs are about your body, and the output is a size for one product on one page.

Everything on this page about Fit Analytics comes from what the company publishes about itself. We do not quote its pricing, its integration options or its internal accuracy, because we cannot verify those from the outside, and a comparison page that guesses at a competitor's numbers is worth nothing.

How is ShirtMath different?

ShirtMath starts from a garment instead of a body. You take a shirt you already own and like the fit of, lay it flat, photograph it next to a scale card, and the scanner reads its flat measurements. Those measurements are then compared against published garment measurements across the catalogue, so the question shifts from "what size am I?" to "which of these shirts is shaped like the one already in my drawer?"

The deeper split is not the interface, it is the unit of measurement. One approach models the body and infers a size; the other measures the clothing and compares like with like. We have written that difference up on its own page: body measurement versus garment measurement. Our own working is on the methodology page.

Why do published garment measurements matter?

Chest is measured flat, across the garment one inch below the armhole, and doubled if you want the circumference. Here is what a Medium means across five styles in that set.

The gap that matters is the last row. A ladies' G500L Medium publishes 17.5" against 20.0" to 20.5" for the unisex Mediums, a three inch difference on the same letter. The gap widens at Large, where G500L publishes 19.0" against 22.0" to 22.5". "Medium" is not a dimension. It is a label that names a different garment depending on the cut, which is the whole subject of our note on vanity sizing.

None of this is an argument against body-based recommenders. It is an argument for knowing the number underneath whichever recommendation you are given. If you want the raw figures, they are laid out in t-shirt chest measurements and compared style by style in which blank runs biggest.

When Fit Analytics is the better choice

There are real cases where a retailer-side recommender is the right tool and ShirtMath is not, and pretending otherwise would waste your time.

• You are a retailer, not a shopper. Fit Analytics is sold to shops that want a size recommender running on their own product pages. ShirtMath has no consumer-facing widget for third-party stores today. Brands who want to talk to us about catalogue work should start at ShirtMath for Brands.
• The shop already has it installed. If a recommender is sitting right there on the product page, use it. It is free to you, it takes seconds, and it knows things about that specific product that an outside tool may not.
• You have no garment to measure. ShirtMath's starting point is a shirt you already own and already like. First purchase in a category, first shirt after a change in body shape, buying a gift: none of those give you a reference garment, and a body-based tool has something to work with where we do not.
• The item is not a shirt. Our live scanner reads shirts laid flat. Anything else, we would be guessing.
• You would rather answer questions than handle a tape or a camera. Typing your height and weight is faster than laying out a garment and photographing it. That is a genuine trade, and speed sometimes wins.

What can ShirtMath do today, and what can it not?

ShirtMath is pre-launch, and the honest boundary is narrow. What is live: photograph a shirt laid flat with a scale card, get its flat measurements back, and match those against published garment measurements across the catalogue. Roughly half of scanner readings land within an inch of a tape measure. Sleeve opening and bicep are the least reliable readings. Anything the scanner cannot read confidently is left blank rather than filled with a guess, which is why some results show fewer fields than others.

What is not live, and what we will not imply: any form of body scanning, shoe identification or shoe fit from a photograph, sunglasses measurement from a photograph, and face-shape analysis. Where this site covers those subjects, it covers them as reference material about how to measure by hand.

If you want to check the scanner's work, or skip it entirely, the tape method is written out in how to measure a shirt and the single most useful number is explained in pit to pit measurement.

How do you decide between them?

1. Ask who the tool is for — Fit Analytics is bought by a shop and shown to you. ShirtMath is used by you and shown to nobody. If your problem is running a store, that answers it immediately.
2. Ask what it starts from — body details, or a garment on a table. If you own a shirt that fits you well, the garment route removes a whole layer of inference.
3. Ask whether it travels — a retailer-side widget helps on the sites that carry it. A garment measurement you hold yourself is the same number on every site.
4. Ask what you can check — a published chest figure can be verified against a tape in your own hands. Keep the number, not just the recommendation.
5. Use both where you can — take the shop's recommendation, then sanity-check it against the flat chest of a shirt you already like. Where they agree, buy with confidence. Where they disagree by more than the tolerance band, look harder.

Common questions

More comparisons live on the compare hub, including our write-up of the True Fit alternative question.