A shopper has a jacket in her basket. She is between sizes in every brand she has ever bought, the chart gives a chest measurement she has never measured, and the reviews contain one person saying it runs small and another saying it is generous.
She closes the tab. Nothing about the product was wrong.
The teardown: why size charts fail
They ask a shopper to measure herself with a tape she does not own, against a garment she cannot see, and to trust that the brand measures the same way as the last one. Almost nobody does it.
Failure one: body measurements instead of garment measurements
Garment measurements, flat, per size, are more useful because a shopper can measure something she already owns and compare. This single change helps more people than any chart.
Failure two: no fit description
Whether it is cut close, relaxed or oversized, and what that means in practice. Shoppers want to know how it is meant to sit, not just the numbers.
Failure three: nothing about the model
Height and the size worn. Shoppers calibrate against it instantly and it costs nothing to publish.
Failure four: no guidance for between sizes
Which way to go, and why. Brands know the answer and rarely say it, leaving every between-size shopper to guess or leave.
Failure five: no comparison to known brands
Saying how it compares to widely known sizing is genuinely useful and almost nobody does it, presumably out of caution.
Failure six: fabric behaviour is unmentioned
Stretch, shrinkage, whether it relaxes with wear and how it washes. A rigid fabric and a stretchy one at the same measurement fit completely differently.
Failure seven: reviews are not mined
Customers say whether something runs small in reviews. Summarising that consistently on the product page uses information you already have.
Failure eight: the return policy is not visible at the size selector
A shopper hesitating over a size is reassured by an easy return. Putting the policy where the hesitation happens converts, which is the argument in an AI shopping assistant for product fit.
Failure nine: nothing is said about how it was worn in the photographs
A garment styled loose on a model two sizes below the one photographed tells a shopper nothing useful. Saying which size is shown, and whether it has been pinned or styled, is a small honesty that costs one line and prevents the return that arrives with a note saying it looked completely different online.
What actually works
Garment measurements per size, a fit description, model height and size, between-size guidance, fabric behaviour, comparisons and a visible return policy.
Answer the question live
A shopper hovering at eleven at night will ask if there is somewhere to ask. Answering from the published fit information at that moment is what saves the sale.
Track what they ask
The sizing questions you receive tell you which products are mis-sized and which descriptions are failing, which is far more useful than a return rate.
Launches make it worse
New products have no reviews and no accumulated knowledge, which is the pressure in launch day questions when a DTC brand drops a product.
Fit questions dominate other categories too
They are covered in bike sizing and fit questions before buying online, and order questions follow in where is my order, questions that flood DTC inboxes.
Getting started
Publish garment measurements, fit descriptions, model details and between-size guidance, then answer sizing questions at the moment of hesitation. SpideyChat for ecommerce answers from your published pages. Try SpideyChat on your own site, or read about answers on the features page.