Returns Are a Product Problem, Not a Policy Problem

When return rates hurt, the reflexive fixes are policy fixes: shorten the window, charge for the label, tighten the rules. Occasionally justified, usually backwards. Policy friction suppresses orders (return-policy anxiety is a documented conversion factor — generous policies raise sales more than they raise returns in most categories) while the returns keep coming, because most returns are the product's fault. Or rather: the listing's.

Read the reasons like research

Every marketplace and most platforms capture coded return reasons, and most integrations discard them. Aggregated per SKU, they are the cheapest product research that exists:

• "Doesn't fit" clustering on one SKU → the size chart is wrong or the cut runs small. Fix: measured garment specs and a one-line fit note ("runs small — most customers size up"). We have watched this single line cut a SKU's return rate by a third. Ecommerce-wide, sizing drives the plurality of apparel returns — it is the biggest lever in the building.
• "Not as described / looks different" → photography is flattering to the point of fiction, or colour management is off. Fix: honest detail shots, texture close-ups, colour-calibrated images, video drape for garments. Conversion barely moves; returns drop; contribution rises. That trade is nearly always worth it.
• "Arrived damaged" → packaging engineering, not customer service. One client's ceramics line: a €0.30 corner insert removed a double-digit damage-return rate. The fix was in the warehouse, and the data was in the returns file nobody opened.
• "Changed mind" concentrated in one channel → look at that channel's advertising promises and its delivery time. Slow delivery inflates regret returns measurably; the cure is the delivery promise, not the policy.
• Serial returner patterns (wardrobing, bracket-buying) → a segment strategy, not a policy for everyone. Bracket-buyers of sizes are telling you, again, about the size chart.

The operating loop

Route returns reasons into the same data layer as sales and margin; compute return-adjusted contribution per SKU (the only margin number that is real); rank the offenders; fix listings, packaging or the product itself in that order of cost; watch the cohort after the fix. Monthly cadence, owned jointly by ecommerce and product — because the split ownership is exactly why nobody currently does it.

Where policy does belong

Fraud and abuse are real and rising (refund-scam services are an industry now); handle them with targeted controls — verification on flagged patterns, receipt-required exceptions, velocity limits — not blanket friction that taxes the honest majority. And in the EU, remember the 14-day withdrawal right is statutory: your policy competes on generosity above that floor, never below it.

Returns are margin leaking through the gap between what the listing promised and what the box delivered. Close the gap with truth, not terms — the customers you keep will be the profitable kind.