Customer Return Rates: Measure Your Store and Model the Cost

Choose the orders before calculating the rate

Use orders delivered in a defined period and allow enough time for the return window to close. Divide returned orders by delivered orders in that same cohort. Counting this month's refunds against this month's new sales mixes different customers and can distort the result. Track returned units separately if customers often buy multiple items.

Compare like with like

Split your report by SKU, variant, delivery country, sales channel and reason code. A size-related return suggests a different action from transit damage or an item not matching its description. Keep cancellations before shipment separate from delivered-order returns. This page does not publish category averages: we do not have a comparable, sourced dataset for those claims.

Measure the change in profit from a refund

Start with the profit the completed order would have earned. A full refund reduces it by the refunded revenue, plus extra handling and return postage, minus eligible fee credits and the value of recovered inventory. Do not subtract the original product cost twice: it was already included in completed-order profit.

Worked sensitivity example

Hypothetical inputs: 200 orders, $30 revenue per order, $4 additional return handling, $8 recovered inventory and no fee credit. Each full refund changes profit by $30 + $4 - $8 = $26. At assumed return rates of 5%, 10% and 15%, the expected monthly reduction is $260, $520 and $780. These are scenarios, not measured industry rates.

Record an intervention and wait for comparable cohorts

If you change a sizing guide, packaging or product photos, compare mature cohorts with the same refund definition. Record order volume alongside the rate so a few returns in a small sample do not look like a stable trend.

Method: these examples use stated hypothetical inputs. Review our calculation methodology before applying them to your business.