Size-aware personalization: showing the right fit first without a quiz
Most apparel shoppers know their usual size and still get burned by brand variance. Size-aware personalization fixes the default: instead of showing every product in a generic size order, the store leads with the size that fits this shopper, based on what they bought and kept before.
Why fit defaults matter more than fit content
Stores invest in size guides, fit notes, and model measurements, then present a size selector defaulted to medium or sorted in a way that ignores everything they know. Shoppers under time pressure take the default, and the return label gets printed a week later.
Changing the default is a small intervention with outsized effect. Pre-selecting the likely size, ordering the selector around it, and flagging low stock in that size first all reduce the chance a shopper checks out with a guess.
The signals that predict fit
Kept purchases are the strongest signal: a size bought and not returned is close to ground truth. Returned-for-fit purchases are nearly as valuable, because the reason code tells you the direction of the error, not just that it happened.
Category context matters. The same shopper may take different sizes in tailored and knit categories, or in footwear versus apparel. Keep size profiles per category rather than collapsing everything into one number that is wrong everywhere politely.
When the store has no history
For new shoppers, start from the product side: this style runs small, this fabric has no stretch, most buyers size up. Aggregate return reasons per product are a fit signal no individual history can match, and they improve with every order.
Offer the two-minute version of help, not the twenty-question quiz. Height, usual size in a known category, and fit preference can set a provisional profile, clearly editable, that the first kept purchase then confirms or corrects.
Handling gifts, shared accounts, and changes
Size profiles break in predictable ways: gift purchases pollute the history, shared household accounts mix two bodies into one profile, and real bodies change over time. Weight recent kept purchases most, let shoppers view and edit their fit profile, and never treat a gift order as fit evidence.
Transparency prevents the creepy version of this feature. A small note such as sized from your past orders, with a way to change it, turns a silent inference into a service the shopper controls.
Measuring whether it works
The scoreboard is fit-related return rate and size-related support contacts, not click-through on the size selector. Compare shoppers with size-aware defaults against a holdout, by category, because fit problems concentrate in categories like denim and tailoring.
Watch bracketing as a secondary signal. If confident defaults reduce the buy-two-sizes behavior, average return cost per order falls even before the headline return rate moves.
Frequently asked questions
What if a shopper has never bought from us before?
Use product-level fit data instead: return reasons aggregated across buyers, runs-small or runs-large notes, and fabric stretch. A short optional fit profile can fill the gap until the first kept purchase.
Should the size profile follow the shopper across categories?
Only cautiously. Fit varies by category, so keep separate profiles for categories with different sizing logic and let each be confirmed by kept purchases in that category.
How do we stop gift orders from corrupting fit data?
Flag gift orders at checkout and exclude them from fit evidence. Orders shipped to a different name or address with gift options selected should never update the buyer size profile.
Is showing a recommended size a privacy concern?
It uses purchase history the shopper already has with your store. Naming the source on the page and letting shoppers edit or clear the profile keeps it transparent and controllable.