Ecommerce Return Cost per Order: Price Refunds Into Your Margin

Published by MyBreakeven. Report a calculation or content issue to support@mybreakeven.com.

Published September 27, 2026 · Updated September 27, 2026

Calculate return cost per order from refund losses, shipping, processing, inspection and resale value, with worked examples and a break-even check.

E-commerce · Returns · Contribution Margin

A return does more than reverse a sale. Outbound freight may already be spent; payment processing, a return label, inspection and repackaging can remain even when the product comes back. Measure the loss on a returned order, then spread expected return losses across every order you price. The numbers below are an illustrative model, not a claimed average return rate.

Quick answer: Expected return burden per placed order equals return probability times the contribution lost when a customer returns, compared with a kept order. If a kept $80 order contributes $29, a returned order loses $31, and 10% of orders are returned, expected contribution is 90% × $29 + 10% × (−$31) = $23. The return burden is $6 per placed order.

Decide what "cost of a return" means

There are two useful views. Cash loss on the returned order counts unrecovered outbound freight, handling, card fees, return freight, inspection, damaged stock and refunds, less any resale proceeds. Incremental cost versus a kept order measures the contribution you expected but lost. The second view is what belongs in an expected-contribution comparison across all placed orders. Do not subtract both the entire refund and the forgone contribution from the same order: that double counts the lost sale.

Whether a processing fee is returned depends on the payment provider and contract. For example, Shopify Payments says its original card fee is not refunded when a merchant refunds a transaction; that is a provider-specific rule, not a universal rate. Check your own gateway, return policy and fulfillment invoices. Return shipping may be paid by you, the customer or split. A restockable item and one that must be discarded have very different economics.

A complete $80 order example

Assume a product sells for $80. If kept, its product cost is $28; fulfillment and outbound shipping cost $12; payment processing and platform charges cost $4; and acquisition spend attributable to that order is $7. Kept-order contribution is $80 − $28 − $12 − $4 − $7 = $29. Keep acquisition cost in the model even when it was spent days before the purchase; it is still money the sale must recover.

If the customer returns the product for a full refund, assume the unit can be sold again at full price later. Restore the $28 unit to inventory rather than treating it as destroyed, but record $6 return freight and $2 inspection and repackaging. The spent $12 outbound freight, $4 fees and $7 acquisition remain. That return produces a $31 loss on the order: −$12 − $4 − $7 − $6 − $2 = −$31. The product cost is recovered as sellable inventory under this assumption. If the unit cannot be resold, account for the unrecovered stock value too.

The difference between a kept order's $29 contribution and a returned order's −$31 is $60. At a hypothetical 10% return probability, expected contribution per placed order is 0.90 × $29 + 0.10 × (−$31) = $23.00. The return burden relative to the no-return $29 case is $6 per placed order. On 200 orders, that is $4,600 expected contribution before fixed overhead. Use observed returns by category; an order-level storewide rate can hide a costly product line.

If the returned unit has only $12 recoverable value instead of $28, write down $16 in the return case. Returned-order contribution becomes −$47 and expected contribution at the same 10% return assumption becomes $21.40. Check the actual recovered value when the item resells; do not book a hoped-for resale as guaranteed cash.

Three practical return scenarios

Lower return volume, same $80 order

At a 5% return probability with $29 kept contribution and −$31 returned contribution, expected contribution is 0.95 × $29 + 0.05 × (−$31) = $26.00. Compared with the 10% assumption above, the extra $3 per order across 200 orders amounts to $600 of monthly contribution. That does not mean lowering returns is free; product photos, fit guidance and service may themselves cost money.

Merchant pays a more expensive return label

Keep the 10% probability but raise return freight from $6 to $16. Returned-order result drops from −$31 to −$41. Expected contribution is 0.90 × $29 + 0.10 × (−$41) = $22.00. The extra $10 on each return costs $1 per placed order at this return probability. Model the difference before promising free returns on every SKU.

Partial refund or exchange

Suppose an $80 order is exchanged for another $80 item, with $14 for reverse and replacement freight and $3 extra handling. It is not equivalent to a full refund: revenue stays $80 if the replacement is paid for and the original unit is saleable, but contribution falls by $17 from $29 to $12. If the original cannot be resold, add its unrecovered value too. Exchanges need a separate line in your data, and repeat customer value is a separate decision from immediate order contribution.

How returns affect break-even orders

Suppose fixed overhead and owner-pay target total $4,600 a month. With no returns and $29 contribution, $4,600 ÷ $29 means 159 whole orders. With the 10% return scenario's $23 expected contribution, $4,600 ÷ $23 means 200 placed orders. That is 41 additional orders before accounting for changes in traffic quality or capacity. Neither count guarantees that the projected return mix will occur, so monitor actual outcomes weekly.

How to run your own numbers

Export a recent cohort of placed orders by SKU and channel. Mark kept, refunded, exchanged, canceled before shipment and partially refunded orders separately. For each returned item, record cash refunded, outbound and reverse freight, nonreturned fees, handling, acquisition spend and actual recoverable inventory value. Calculate contribution for kept and returned orders with the same cost boundary. Feed your weighted average product price, variable costs, expected returns, overhead and fulfillment capacity into the ecommerce break-even calculator. It is a planning aid; replace modeled return assumptions with actual cohorts as they mature.

Refunds and exchanges are not the same as fraudulent chargebacks, which may add other fees. Read the ecommerce shipping cost guide for the outbound and subsidized-shipping side, the product pricing guide to rebuild price, and the orders-per-day guide to test demand and fulfillment.

Common mistakes

• Calling the full refund a second expense: if you already remove the sale from revenue, subtracting the same $80 again makes the loss look twice as large.
• Assuming all returned products can be resold: inspect actual condition and recovery value before giving inventory full credit.
• Using one return percentage for all SKUs: channel, product and customer cohort can change both probability and loss severity.
• Ignoring time: a refund in a later month can leave a cash gap today even if the model balances eventually.
• Assuming gateway fees behave alike: verify the refund terms for your own provider.

Key takeaways

• Compare kept-order and returned-order contribution using one consistent cost definition.
• Weight those outcomes by observed return frequency for each product and channel.
• Price shipping, processing, handling and unsellable inventory into the decision.
• Convert expected contribution into required orders and check fulfillment capacity.

FAQs

How do I calculate ecommerce return cost per order?

Compute contribution for a kept order and a returned order. Multiply the difference by your observed return probability. In the worked example a $60 difference at a 10% return probability reduces expected contribution by $6 per placed order.

Should returned inventory be counted as lost?

Only to the extent that value cannot be recovered. If the item can be resold, record a supportable recoverable value; if it is damaged or obsolete, include its write-down.

Are card processing fees refunded on returns?

Policies vary by provider and contract. Shopify Payments says its original credit card fee is not returned when a refund is issued. Verify your own gateway terms before forecasting.

Is a 10% return rate a benchmark for my store?

No. Ten percent is only an illustrative assumption in this article. Measure your own mature orders by SKU, channel and period before using a return probability in pricing.

How do exchanges affect the model?

Track exchanges separately from full refunds. If revenue is retained, extra shipping and handling reduce contribution; if the original item cannot be resold, add its unrecovered value.

Browse the MyBreakeven guide library for more planning guides.