How return rate is measured
Count the orders returned, refunded or refused during a period and divide by the orders placed. Two details decide whether the result is worth anything. The first is what you allow to count. A parcel sent back after delivery is obvious, but an order refused at the door, cancelled before dispatch or lost in transit costs money in much the same way, and those are usually recorded somewhere else entirely, if at all.
The second is timing. A return happens after the sale, sometimes weeks after, so the rate for a recent period always looks better than it will eventually turn out to be. Compare like with like by measuring returns against the period the order was placed, not the period the return arrived.
Why return rate matters
Every advertising figure you look at is gross. The platform records a purchase the moment checkout completes, so returned and refused orders remain in reported revenue and in return on ad spend for good unless something goes back and removes them. Two campaigns with identical reported results can be a long way apart once returns are taken off, and the one that looked better is often the one that sold hardest to people who were least sure.
The cost is also worse than the lost sale itself. You paid to acquire the customer, to pick and pack, to ship out and often to ship back; the item may not be resellable; and where payment is collected at the door, a refusal means the whole journey happened with no revenue at all. That is why cash on delivery deserves its own line in the margin calculation rather than being treated as a payment detail.
Where return rate goes wrong
Most returns are a promise the page did not keep. Photographs that flatter, measurements missing, a vague description, the wrong variant chosen because the selector was confusing, or a delivery that arrives long after it was wanted. Heavy discounting and urgency-led advertising add their own share, because a purchase made in a hurry is the one most likely to be regretted.
The reporting mistake is treating returns as the warehouse’s business. Nobody in marketing sees them, so campaigns keep being optimised towards products and audiences that return heavily, and the account gets steadily worse while every dashboard insists it is improving.
How to act on it
Break the rate down before acting on it, because returns concentrate rather than spread evenly. Look at it by product, by variant, by campaign and by delivery area, and deal with the few lines creating most of the cost. Fix the page first — better photographs, real measurements, honest delivery times, plain terms — since those changes lower returns without lowering sales, which is rarely true of the alternatives.
Then get the number into the decisions that matter. Send refunds back to the advertising platforms, or at the very least adjust reported revenue before judging performance, and assess campaigns on contribution margin rather than gross sales. For a retailer that discipline is the difference between growth and busy loss-making, and it is the least glamorous part of ecommerce and retail marketing that separates shops that last from shops that scale into trouble.