Week 7: The Cost of Coming Back - Returns Economics

An e-commerce return isn't free once it leaves the warehouse — the return-rate and reverse-logistics numbers behind the other half of the throughline claim

5 April 2027

Every week so far has priced the outbound half of a shipment: the box, the drop test, the freight formula. Half of the throughline claim’s cost side has been missing. This week fills it in: what happens, and what it costs, when the product comes back.

The numbers#

Roughly 19.3% of e-commerce orders are returned overall — but that average hides a wide spread by category: apparel 20-40%, footwear 17-30%, electronics 8-15%, beauty 4-12%, jewelry 12-15%. Processing a return isn’t free: reverse-logistics and restocking for electronics runs about $30-65 per item, and for furniture or other oversized goods, the reverse-logistics cost can exceed the entire margin on the item. About half of all apparel returns come down to fit and sizing alone, not a defect or a change of mind. Some brands now issue a “keep-it” refund — refunding the customer without requiring the item back — once the cost of collecting, inspecting and restocking it would exceed what the item is worth.

What this does to week 1’s lululemon case#

Apparel sits at the top of the return-rate table, and half of that is fit alone — something no amount of packaging can fix, since the product itself is the variable. But packaging still touches this cost in two ways. First, generous sizing options mean more units in circulation that are statistically likely to come back, which is itself a reason to keep the outbound packaging cheap and standardised (week 4’s oversized-box logic) rather than premium — a returned item that gets reshipped in the same generic box costs less to process than one that needed custom packaging undone and redone. Second, an easy-to-repack box lowers the labour cost of that return cycle, which is a real, if less visible, line item next to the freight cost week 4 already put a number on.

The “keep-it” refund is the sharpest version of this logic: once you’ve built a return-cost model like the one above, refusing the item back and eating the loss on the product is sometimes cheaper than collecting it. This isn’t generosity. It’s the same cost-minimising logic as every other decision in this course, applied to the one direction of shipping we hadn’t priced yet.

This week’s lab#

Returns Cost Model — build a small return-cost model for a product category, and work out where a “keep-it” refund would actually make sense.