← Back to Blog

The five return metrics that deserve your attention

The overall return rate tells you almost nothing. Five finer measures you can act on: by product, by reason, by processing time, by value recovered and by repeat returners.

Heylio Team August 19, 2026 3 min read

Key takeaways: An aggregate return rate hides everything. Five metrics let you act: the rate per product, the breakdown by reason, the processing time, the value recovered per return and the share of repeat returners. Each one points to a different action.

Why the overall rate is useless

Take a sentence like "We're at 18% returns." It supports no decision. It doesn't say which products, for what reasons, or whether things are improving.

Worse, that figure moves for reasons unrelated to the quality of your offer: a change in product mix, a new sales channel, a promotion. You end up commenting on the swings of a number that doesn't measure what you think it does.

Five finer metrics, on the other hand, let you decide.

1. Return rate per product

It is the one that pays off most, and often the most neglected.

Returns are never spread evenly across the catalog. A handful of products account for a disproportionate share of the volume. Finding them takes one query; fixing them takes little time and has a quick effect.

The causes are almost always the same and almost always fixable: a photo that doesn't show the real color, an item that runs small without saying so, a description that promises a material or a dimension the product doesn't have.

The right reflex: sort your catalog by return rate, look only at the top twenty products, and rework the product pages one by one.

2. Breakdown by reason

Provided you collect usable reasons. "Doesn't suit me" teaches nothing; "too small", "color different from the photo", "quality below expectations" and "defective product" each point to a different action.

The distinction that matters most is between avoidable reasons and unavoidable reasons. A sizing mistake is avoidable: a size guide, reviews from other customers on display. A change of mind is not, and trying to eliminate it means degrading the buying experience.

Tracking the share of avoidable reasons over time teaches you more than tracking the total rate.

3. Processing time

From the moment the customer registers their return to the moment the money goes back. Split it into three segments: registration → shipping, shipping → receipt, receipt → refund.

The first segment depends on how clear your instructions are and how easy the label is to get. The second depends on the carrier. The third depends entirely on you, and it is the one that generates complaints.

Looking at the average is not enough: the tail of the distribution is what costs you. For example, ten returns processed in two days and one processed in three weeks give a reassuring average and one furious customer.

4. Value recovered per return

How many euros of merchandise actually go back on sale, divided by the number of returns.

This metric connects logistics to finance. It rises when sorting at receipt is fast and rigorous; it falls when parcels sit waiting, when items get damaged in storage, or when refurbishing decisions are made by gut feeling.

It is also the only metric that puts a money value on a logistics investment, which makes it useful for trade-offs.

5. Share of repeat returners

A minority of customers accounts for a large share of returns. That is not necessarily fraud: they are often shoppers who order several sizes to keep only one, a behavior encouraged by free returns.

Measuring it lets you choose a proportionate response, and avoid confusing the two groups. A customer who orders a lot and returns a lot can still be very profitable; a customer whose net basket after returns has been negative for a year is not.

A metric without a time cut is worthless

One last point, valid for all five. A return is tied to an order placed weeks earlier. Comparing one month's returns to the same month's sales gives a false ratio, especially after a sales campaign.

Tie returns to the order cohort, not to the month they were received. It takes a little longer to set up, and it is the difference between a metric people comment on and a metric people decide on.

This article first appeared on the Hey’ Return blog, the returns management tool of the Heylio suite. Discover Hey’ Return

Automate your customer support with AI

Get started