St Andrews University Shop: try-on that paid for itself twice over in month one

The official shop of the University of St Andrews sells crested apparel, knitwear and gifts to students, alumni and visitors worldwide — most of whom buy online, sight-unseen, often from another continent. In June 2026 it became the first store to run Rendered Fits in production. In its first month live, try-on-influenced revenue came to more than double the monthly plan price — from under £20 of consumed try-on credits.

The numbers

Figures from Rendered Fits attribution (try-on-session-linked orders only) and the store's live fit-data instrumentation, 10 June – 10 July 2026. Return multiple uses the Starter plan's per-credit rate — the conservative case.

Why a university shop is a hard test

This is not a boutique with thirty dresses. The catalogue spans hoodies, quarter-zips and knitwear alongside mugs, whisky, stationery and shortbread — over 900 products where only the wearable ones belong in a try-on experience. Buyers range from an 18-year-old fresher to alumni ordering graduation gifts overseas, across a campus network where hundreds of users share one IP address. It exercised the product harder than a typical fashion store would.

What went live

• Photorealistic try-on on the product page, generated from the shop's existing product photography — no reshoots, no 3D digitisation.
• Complete the Look — shoppers build a full outfit from the catalogue and see it rendered on themselves in one image, with non-wearable giftware automatically excluded.
• AI size recommendation — fit guidance at the moment of decision, now feeding an outcome-linked fit dataset for the store.
• Selfie-only mode — a clear face photo is enough; no full-body photo required.

Timing that mattered

The rollout landed just before graduation season — the shop's highest-intent weeks of the year, when families and graduates buy crested pieces they cannot try on in person. That is precisely the moment "how will this look on me?" hesitation costs the most revenue, and the month-one result reflects it.

What happens next

The store's fit dataset is now instrumented end to end: every size recommendation is linked to what was subsequently kept or returned, against a measured baseline return rate of 9.4% across ~25,000 historical units. As that dataset matures, the shop's size guidance gets sharper — and the return-rate effect becomes directly measurable rather than estimated.