Your category's COD reality, quantified.

6×–20× on Meta, under break-even delivered — our live pilot ran 5.6× booked, 1.56× delivered after 50% RTO.

A drop books 6×–20× on Meta; after apparel's 25–35% RTO the delivered number is a fraction of it — often under break-even.

The most-returned category online (~31%) on 60–80% COD — see profit per delivered pair, model by model, after every RTO.

57% gross margin, roughly zero net — see delivered ROAS and per-design profit after millimetre returns, RTO and plating write-offs.

Beauty pays back on the repeat order — but a quarter of COD first orders never deliver, and a returned serum is a write-off, not a restock.

Subscription LTV assumes the first jar delivers — 26% of COD orders don't, and a returned supplement is destroyed stock, not a restock.

Decor pays freight on the air in the box and breaks on the last mile — a 3.5× ROAS on ~3% contribution is one cracked mirror from a loss.

At food margins break-even is 4.5× — so the '4× winner' is losing, and every returned box is destroyed stock, not a restock.

Your buyer inspects at the doorstep before she pays — ~26% of COD baby orders RTO and refused diapers write off in full. See the ROAS that survives delivery.

A ₹250 RTO on a ₹600 wallet is the whole order's profit — and 28–35% of impulse COD accessory orders bounce. See the ROAS that survives the round trip.

Your repeat engine only compounds if the bags actually land — over 25% of COD orders fail, and a bounced 5kg bag bills freight both ways. See LTV on delivered revenue.

Your 3.0× target ROAS was set without returns in the formula — at 18–25% RTO, campaigns 'hitting target' lose money per order. See the break-even that includes the bounce.