Returns & CX Apr 2, 2027 5 min read
How to Use Past Purchases to Recommend Sizes
Past-purchase data is the strongest size predictor. 'You wore M in our shirt' → 'we recommend M in our hoodie too'. Same brand high accuracy. Cross-brand requires fit calibration data.
The Chart
| Data Available | Recommendation Accuracy |
|---|---|
| Same brand + product type | 95%+ — same fit baseline |
| Same brand, different product | 85-90% — slight calibration |
| Different brand (with calibration data) | 75-85% |
| No history | 65-75% from chart + body measurements |
FAQs
Cross-brand size logic?
Need brand fit data for both. 'You wore Levi's 511 W30' → can recommend 'M in our slim shirt' if we know our shirt fits like Levi's. Tailor Size Guide offers brand calibration data.
What about returned past items?
Strong signal. If customer returned size M for being too small in similar product, recommend L next time. Loop returns through the recommender.
Need this on your store?
Tailor Size Guide ships pre-built size charts for Shopify.