Break-even ROAS for an order is approximately 1 divided by the share of revenue left after variable non-ad costs. If 40% of a sale remains for acquisition, the order needs 2.5× revenue per ad-cost unit to cover those variable costs. That is not company-wide break-even: fixed costs, taxes and attribution gaps still matter.
- 012× ROASBelow ad break-even50% of break-even
- 024× ROASAt ad break-even100% of break-even
- 036× ROASAbove ad break-even, before overhead150% of break-even
What ROAS says—and what it does not
ROAS is attributed sales value divided by advertising spend. A 3× ROAS means the platform reports three units of sales for one unit of ad spend. It does not mean three units of profit. The store must still pay the supplier, delivery, refunds, payment processing and overhead. Platform attribution can also differ from your actual order records.
Derive the threshold from contribution
Start with 100 units of revenue. If product, fulfillment, fees and expected returns total 60, the pre-ad contribution margin is 40%. Spending 40 to win that order consumes the remainder. Break-even ROAS = 1 ÷ 0.40 = 2.5×. If a product has only 25% pre-ad contribution, the threshold rises to 4×. These are illustrative arithmetic examples, not performance targets or promises.
Add a safety margin
Your real target should allow for fixed costs and measurement uncertainty. Returns often arrive later than ad reports, and new-customer acquisition may cost more than blended acquisition. Check the number against settled orders over a meaningful period, not the dashboard from a single day. Shopify's analytics guidance explains the funnel view that helps separate traffic from completed orders.
If your ad budget is small, the first question may be whether the test can produce enough observations to learn at all. Use the ad-test budget guide before treating a reported ROAS as truth.
Calculate pre-ad contribution for the exact product or bundle, divide one by that fraction, and compare the result with settled orders—not only attributed sales.