Calculate the effect of refunds
Choose one measurement scope. Use revenue after discounts and before refunds, excluding tax and shipping. Match refunds to those same orders. This scenario runs only in your browser.
After-refund ratio = (revenue − refunds) ÷ spend. All three inputs must describe a consistent scope. The refund share concerns revenue, not the number of returned orders.
This is not profit or evidence of incremental sales. Product costs, fulfilment, payment fees and return handling are not deducted. An immature return window can overstate the retained revenue.
Define the numerator before calculating the ratio
Decide which discounts, taxes, shipping charges, cancellations and refunds the starting amount already includes. Do not subtract an adjustment twice. Shopify distinguishes gross sales, discounts, sales reversals and net sales. Sales reversals include more than physical returns. Check the exported field definition before using it in your calculation.
If attributed revenue cannot be linked reliably to refunded orders, label the limitation. Applying the company’s average return rate to every campaign creates an estimate, not an observed campaign result. Keep estimated adjustments separate from order-level reconciliations.
Match refunds to the order cohort
A refund processed this week may relate to an order from an earlier campaign period. A management analysis based on order cohorts should retain that relationship and state the observation cutoff. A cash or refund-processing-date view answers a different question. Both can be useful if they are labelled clearly.
Choose a review cadence that suits the business and record how mature the cohort is. Mark recent orders as provisional while further returns can arrive. Do not compare a fully observed previous season with a recent campaign as though both have had the same opportunity to generate refunds.
Read revenue efficiency alongside the remaining costs
A revenue-to-spend ratio still leaves product costs, fulfilment, payment costs and return handling outside the calculation. A higher adjusted ROAS is therefore not automatically a higher contribution margin. Keep the commercial measure used for the final decision next to the marketing ratio.
Review whether the apparent difference comes from acquisition quality, product mix, discounting or fit-related returns. Use those questions to create separate briefs for paid media, merchandising and conversion work. The analysis should help identify what to investigate rather than collapse several operational issues into a budget instruction.
Keep the reconciliation inspectable
Store the reporting period, spend source, attribution rule, order definition, refund treatment and cutoff with the calculation. Preserve both the reported amount and the adjusted view so that another team member can reproduce the change. Explain partial refunds and exchanges explicitly when they affect your source data.
Ask the Ads Agent workflow to prepare the evidence and the comparison for review. Connector access alone does not establish a reliable order-to-campaign join. Confirm the available fields and definitions before presenting the output as an observed commercial result.
A ratio that changes as orders mature
Illustrative, consistently defined values: €10,000 of attributed revenue and €2,000 of ad spend give 5.0× reported ROAS. If €2,000 of relevant refunds are matched to those same orders, the adjusted numerator is €8,000 and the ratio is 4.0×. The example assumes the starting revenue did not already subtract those refunds. Neither ratio includes all business costs or proves incremental sales.
Use this at work
- Define revenue and exclude double-counted adjustments.
- Match refunds to the relevant orders where evidence allows.
- Label the cutoff and maturity of the cohort.
- Review contribution alongside the revenue-to-spend ratio.
Reference material
Platform guidance checked 3 October 2026. Examples and working checklists are Faccelerate editorial illustrations.