The whole set came back because four pieces were missing

A 480-piece set arrives four bricks short and the customer returns all 480. That is rational behavior — the alternative on offer was a half-built model — but it is a catastrophic outcome for the seller: the set cannot be resold as new, the refund is full price, and the customer usually leaves a one-star review on the way out.

Seven in ten returns named missing or broken pieces, and the review page repeated the complaint loudly enough that new shoppers were reading it before buying. Nothing on the listing, in the insert or in the brand's messaging offered any route other than “return it”, so every shortfall of four bricks cost the brand a full set and a point of rating.

We built the alternative to a return and put it where the customer looks.

• A replacement-parts request route, printed on the insert and repeated in the listing, promising the missing pieces within days at no cost.
• Spare-piece packs listed as their own ASINs, so a buyer who simply lost a brick has somewhere to go instead of the return button.
• Count verification at the packing line for the four sets that generated most complaints, plus a small surplus of the most-lost pieces in each box.
• Expectation set in the image stack — the exact piece count and the replacement promise shown as a frame, not buried in the A+.

Returns fell from 11% to 4%, and net revenue rose 49% on modest order growth — about a sixth of that is product that no longer comes back unsellable, and the rest is what a 4.1 rating does to conversion.

The rating moved 3.7 → 4.1 because the customer whose set was short now gets replacement bricks in the mail instead of a refund, and that customer does not write the review the old process guaranteed. Over-packing five small parts costs pennies per set; a single return cost the full retail price plus the review it came with.