A refund total looks like the cost of returns. It is the smallest part of it. The money has already been spent acquiring the customer, picking and packing the order, and shipping it out — and a return adds shipping back, handling, and a unit that may not be sellable at full price again.

The refund is the visible part. ORVX measures the rest.

Two return rates, because they answer different questions

Units and money are not the same rate Unit return rate = returned units ÷ sold units Refund-to-revenue = refund value ÷ realized revenue

Quoting one as though it were the other is the most common mistake in returns analysis, and the gap between them is itself a signal. When refund-to-revenue runs well above the unit rate, your expensive items are the ones coming back — a different problem, with a different fix, from a high volume of cheap returns.

ORVX publishes both, at store, product and collection level.

Returned units never became revenue

This is the structural point and it is where most reporting goes wrong. In ORVX, realized revenue counts delivered orders only. A returned unit does not enter revenue and then get subtracted later — it never enters at all.

The difference matters because a margin computed against orders placed is a margin computed against money that did not stay. On a store returning a fifth of its units, every margin in the business is overstated until returns are handled at the revenue line rather than as an expense afterwards.

Return rate by product is where the decisions are

A store-level return rate is a number to worry about. A product-level return rate is a number to act on. Returns are almost never evenly distributed — in most catalogues a small group of SKUs generates a disproportionate share, and they are frequently not the ones anyone suspects.

  • A product whose gross margin survives returns and whose contribution does not — usually shipping in both directions on a heavy or bulky item.
  • A size or variant returning far above its siblings — a sizing problem that a product-level average hides completely.
  • A collection carried by one low-return SKU while the rest come back, which looks healthy in aggregate.

The costs a refund total does not include

Where you record them, ORVX carries shipping and handling as cost components rather than folding them into a single blended number, so the true cost of a return can be separated from the refund itself. Outbound shipping on a returned order is money spent with nothing to show for it, and it does not appear anywhere in a refund report.

What to do with it

Returns intelligence is not mainly about reducing returns. It is about pricing and stocking with the return rate already in the number — which changes three decisions:

  • What you can afford to pay to acquire. A 30% return rate means the CPA has to be recovered from 70 units of every 100 you sell.
  • Which products deserve budget. Ranking by post-return contribution reorders most catalogues.
  • Where a product fix pays for itself. Sizing guidance, better photography or a description change on the worst offenders.