Who hurts
DTC brands doing $1M–20M with a return rate above 15%.
The problem
Return reasons are collected as a dropdown and analysed never. Brands treat returns as a logistics cost instead of the highest-signal product feedback they own.
What you build
Cluster return reasons and free-text comments per SKU, quantify the revenue behind each cluster, and output specific fixes — this size chart, that photo, this description claim — ranked by recoverable dollars.
Why now
Free-text return comments were unreadable at scale and are now trivially clusterable, which is where the real reasons live.
Validate it this week
Ask one brand for a returns CSV export. Send back the analysis. The dollar figure closes the deal by itself.
Why you'd keep winning
Cross-brand benchmarks per category — knowing what a normal return rate is for that product type — which no single merchant can compute.
The honest risk
Fixes require merchandising work the brand may not do, and then the tool gets blamed for the number not moving.
The prompt is written to make an AI argue with you before it writes code — that first round of pushback is worth more than the scaffold.