Returns get treated as a cost of doing business and left alone. They are a design problem with three distinct parts, and only one of them is about the policy.
Part one: preventing the return
Most returns are an information failure. The customer learned something after delivery that they should have learned on the product page.
Fashion: fit. Measurements per size against a real body, model height and size worn, fabric stretch. A size guide that says S, M, L prevents nothing.
Anything with scale: a photo against a known object. The number one cause of "smaller than expected".
Anything with a timeline: how long before it works, stated honestly. People return things in week three that take eight weeks.
Colour: multiple lighting conditions, and customer photos above all. Studio lighting lies.
Categorise your returns by reason for one quarter. The distribution is almost always concentrated — one or two causes dominate, and both are usually fixable on the page rather than in the policy.
Part two: the process itself
Self-service. A portal where the customer initiates, selects a reason, and gets a label. Every return handled by an email thread costs more in support time than the item was worth.
Ask for the reason as structured data, not free text. The free text is interesting; the structured reason is what you can act on.
Offer exchange before refund. Same size different colour, different size same item, store credit at a small premium. Each of those keeps the revenue. A meaningful share of refund requests will take an exchange if it is the easier path.
Decide what comes back. For low-value items the shipping, inspection and restocking often exceeds the recoverable value. Refund and let them keep it. This feels wrong and is frequently correct.
Part three: the policy
| Approach | Fits | Cost |
|---|---|---|
| Free returns, 30 days | Fashion, competitive categories | High, often necessary |
| Customer-paid returns | Lower-margin, heavier items | Lowers rate, lowers conversion |
| Exchange or credit only | Consumables, personalised | Low, needs clear disclosure |
| Extended window | Gifting-heavy categories | Moderate, improves conversion |
Whatever you pick: state it on the product page, not only in the footer. A clear policy before the decision lifts conversion more than a generous one discovered after.
The numbers worth tracking
Return rate by SKU. One product at triple the store average is a product page problem or a quality problem, and you want to know which.
Return reason distribution. Where the fixes are.
Exchange rate within returns. How much revenue the process is retaining.
Net margin after returns, by product. Some products are unprofitable once returns are counted. Most brands have never checked.
Talk it through
Send us your top three return reasons.
Two of them are usually fixable on the product page, which is cheaper than any change to the policy.
Get a readA real read within one business day — not a sales call.
About the author
Manpreet Singh
Founder of Proscube. Leads the studio's Shopify, headless and CRO engineering, and works directly with founders rather than through account managers.
More about Manpreet
