Return-Adjusted Order Contribution

Subtracting a return percentage from revenue is rarely a complete return model. A returned order can reverse revenue, recover some inventory value, leave outbound delivery costs unrecovered, and add return handling. Build separate economic outcomes for kept and returned orders, then weight them by the probability of each outcome. This approach makes the treatment of product recovery and refund amounts visible rather than hiding them in a broad margin percentage.

Define the kept-order outcome

Suppose a hypothetical order sells for $60. Product cost is $24, outbound delivery and packing cost $8, and payment costs are $2. A kept order contributes $26 before acquisition and fixed overhead. Assume those same delivery and payment costs remain incurred if the order is returned. The invented payment cost is a scenario input, not a claim about whether any processor returns fees.

Define the returned-order outcome

For a full refund, retained revenue becomes zero. Suppose the returned product has recoverable inventory value of $18 after inspection, compared with its original $24 cost. The unrecovered product cost is therefore $6. Return shipping and handling add $7, while the $8 outbound cost and $2 payment cost remain. Returned-order contribution is negative $23. Counting the full $24 product cost again while also recognizing only a $6 loss would double-count inventory expense.

Weight the outcomes

Assume a 10% probability of this full-return outcome per original shipped order and a 90% probability the order is kept. Expected contribution is 90% times $26 plus 10% times negative $23, or $21.10 per original order. Across 1,000 identical orders, the expected contribution is $21,100. Simply reducing the $26 kept-order contribution by 10% would produce $23.40 and miss the additional economic cost of the return outcome.

Model the actual return patterns

Partial refunds, exchanges, store credit, resalable goods, and destroyed goods require different outcomes. Assign each a probability and check that mutually exclusive probabilities sum to 100%. Use the original shipped order as the denominator unless the analysis deliberately models units. Returns may be correlated with products, sizes, promotions, or shipping damage, so a single sitewide rate can obscure the economics of the item being priced.

Reconcile the estimate after the window closes

Measure a mature shipment cohort and connect refunds, recovered stock, shipping charges, and handling time to the original orders. Keep cash timing separate: the refund may leave the bank before the recovered item sells again. Do not value returned stock automatically at its original retail price; use a justified recoverable-cost assumption for this managerial model. Expected contribution may be negative, and a negative result should trigger a review of the assumptions and offer rather than be replaced with zero.

Method references: OpenStax: break-even and contribution analysis; Stripe: balance transaction types; Shopify: customers reports. Accessed October 22, 2026. All dollar figures are hypothetical worked examples.