An exchange has a net cost, not just a shipping cost
Exchange cost = reverse or replacement shipping + handling − retained contribution. The retained contribution is what the business keeps because the customer takes a replacement instead of a refund.
When retained contribution is larger than shipping and handling, the exchange cost is negative: it is a net benefit relative to losing the sale. That is not automatic; it depends on the replacement product and the customer accepting it.
A refund has a loss
Refund loss includes the contribution lost on the original order plus the cost of processing the return: shipping, handling, payment fees that are not returned and inventory recovery.
The two numbers should use the same definition of contribution. If the exchange retained contribution includes the replacement product cost, do not add that cost again.
The formula
Expected cost per return = exchange share × exchange cost + (1 − exchange share) × refund cost. Monthly expected cost = returns × expected cost per return.
Break-even retained contribution = exchange shipping + handling − refund loss. Above that retained contribution, exchange is cheaper; below it, refund is cheaper.
Worked example
200 returns, 30% exchange share, $6 exchange shipping, $3 handling, $20 retained contribution and $25 refund loss. Exchange cost is 6 + 3 − 20 = −$11. Refund cost is $25.
Expected cost per return is 0.30 × (−11) + 0.70 × 25 = $14.20, or $2,840 per month. All-exchange would be −$2,200; all-refund would be $5,000. Exchange is cheaper by $36 per return.
Reading the break-even
The break-even retained contribution is 6 + 3 − 25 = −$16. Because it is negative, exchange is cheaper even if retained contribution is zero: the shipping and handling are lower than the refund loss.
If refund loss were only $5, the break-even retained contribution would be $4. Exchange would need to retain more than $4 of contribution to beat a refund. The sign and size of the break-even tell you which lever matters.
Where the model stops
It excludes customer lifetime value, brand effects, replacement inventory constraints and any future purchase from a retained customer. It also assumes the exchange share is independent of the refund policy.
Do not conclude that exchange is universally better. Use the comparison for your inputs, then test the policy on a cohort and measure repeat purchases as well as immediate cost.