Ecommerce Intermediate 17 min

Reviews and returns — reading 500 complaints as one operational report

The problem

Your return rate is 14%. You know this number. You do not know what share is 'size did not fit', what share is 'damaged in transit', what share is 'not as described', and what share is genuine product failure.

The workflow
1Export returns with free text
2Strip identifiers
3Cluster by cause
4Categorise into four types
5Cost each cause
6Rank by recovery
7Fix the cheapest wins
Re-measure

What you need

Product: insulated steel lunch box, ₹749. Return rate 11% over 6 months, 340 returns.

The tool

Code mode on your returns export

The prompt

Subscribers. The full prompt is part of the SMEMinds AI Playbook. See plans.

What comes back

1. CLUSTERS (from free text) Leaked in bag — 88 · Smaller than expected — 71 · Dented on arrival — 54 · Does not keep food hot long enough — 43 · Ordered by mistake / changed mind — 39 · Wrong colour sent — 26 · Other — 19 2.

Verify before you use it

Human review required. Check every line below before this leaves your screen.
  • Verify the ₹186 cost per return includes everything. Most sellers omit the unsellable share and the lost contribution.
  • Check the 'about 4 hours' claim yourself. If it is really 3, that is a claim you made that is now costing ₹8,000 every six months.
  • Confirm the reason-code finding by opening ten of those 61 returns yourself. It is the most consequential finding and it should be verified directly.
  • Confirm identifiers were stripped before upload. Returns exports contain names, addresses and phone numbers.