Flipkart: the return that costs twice
Reducing return rate without quietly making returns harder.
Reframes 'reduce returns' into 'reduce the returns caused by bad information' — protecting the trust that drives the next purchase.
The trap
You can cut returns tomorrow. You'll regret it next quarter.
Returns in fashion are a margin killer — you pay shipping twice and often can't resell. The tempting fixes all add friction: stricter windows, harder return flows, restocking fees. They work, and they're a trap. The return experience is where a marketplace earns the confidence to let you buy clothes you can't touch. Break it and you don't reduce returns — you reduce purchases.
So the real question isn't 'how do we return less?' It's 'which returns are we causing, and which are we insuring against?'
Segment the returns
Not all returns are the same problem
- Fit/size returns — the big one in fashion. Caused by bad size information, not bad intent. Preventable upstream.
- 'Didn't match the photo' — a listing-quality and seller problem, concentrated in specific sellers.
- Wardrobing / abuse — small in volume, real in cost; a fraud problem, not a UX problem.
- Genuine defects — these returns are doing their job. Blocking them is how you lose a customer for good.
Tradeoffs
Where I'd spend, and where I wouldn't
Attack fit returns with better size signal
Fit is the largest, most preventable bucket and the fix (size guidance, fit data from past purchases, true-to-size signals) improves the buying experience instead of degrading the return one. You prevent the return before it happens.
Hold sellers accountable for 'not as described'
These returns are caused by the supply side. Routing the cost and the fix to the seller corrects the actual source; punishing buyers corrects nothing and erodes trust.
Keep defect and genuine returns frictionless
A return-rate target that includes legitimate returns optimizes against your own customers. Some returns are the cost of the trust that earns the next order.
Reflection
The metric was the mistake
The whole case turns on refusing the metric as handed over. 'Reduce return rate' is the wrong target; 'reduce information-caused returns while protecting trust-building ones' is the right one. Picking the right metric is most of the product work. [PLACEHOLDER: add a real Flipkart-specific data point if you have one.]