Define the measurement boundary first
Every advertising ratio needs a period, campaign scope, attribution model and revenue definition. Spend from one week cannot be compared with conversions reported over a different attribution window without adjustment. Platform-attributed sales, store revenue and cash collected are different measures. Write the boundary beside the result before comparing channels or periods.
Use consistent treatment for cancellations, taxes, shipping income and returns. Revenue reported immediately after a campaign may later fall because of refunds. For products with material return rates, use net attributed revenue or apply a return allowance so acquisition targets are not based on revenue that will not remain.
Know what each metric answers
ROAS is attributed revenue divided by advertising spend. ACoS is the inverse expression: spend divided by attributed revenue. TACoS divides spend by total marketplace revenue, including organic sales. MER also compares total revenue with marketing spend, but its exact spend and revenue scope should be documented. CAC divides acquisition-related sales and marketing cost by new customers, while CPA may count any chosen conversion rather than a new customer.
CTR, CPC and CPM diagnose delivery before profit. CTR shows how often an impression produces a click, CPC shows cost per click, and CPM shows cost per thousand impressions. They can explain why acquisition cost changed, but none of them establishes profitability without conversion rate, order value, margin and return behavior.
Derive break-even from contribution, not gross revenue
Before advertising, subtract product cost, marketplace fees, payment fees, fulfillment, shipping, packaging, expected returns and other order-variable costs from revenue. The remaining pre-ad contribution is the maximum theoretical advertising spend at break-even. Dividing revenue by that amount gives break-even ROAS; dividing the amount by revenue gives break-even ACoS.
A business target should normally preserve a profit buffer rather than use the full break-even allowance. Reserve the desired contribution or net margin first, then treat only the residual as maximum acquisition cost. This protects the plan from normal volatility in fees, returns and conversion rates.
Worked campaign example
Suppose a product sells for 80 USD. Non-ad variable costs total 52 USD, leaving 28 USD pre-ad contribution. Break-even ACoS is 35% and break-even ROAS is about 2.86. If the seller wants an 8 USD contribution after advertising, maximum ad cost becomes 20 USD, target ACoS becomes 25%, and target ROAS becomes 4.0.
The target is valid only for that product and cost structure. A cheaper product, a higher return rate or a different marketplace fee changes the target. Portfolio campaigns should use a weighted contribution model or be separated by product economics; a single blended target can shift spend toward high-revenue but low-margin products.
Use TACoS and MER as business context
ROAS can improve because a platform attributes more existing demand to ads, while the total business does not grow. TACoS and MER add total revenue context. If attributed ROAS rises but total revenue, new customers and contribution remain flat, inspect brand search, repeat buyers and attribution changes before declaring an improvement.
Conversely, a temporary rise in TACoS may be rational during a launch if it creates incremental demand and later organic sales. Test that hypothesis with cohorts or geographic experiments where possible. Do not assume future organic benefit without evidence, and set a time limit for the investment case.
A repeatable review cadence
Review delivery metrics daily only when spend volume justifies it, but make economic decisions on a window long enough to include normal conversion delay and returns. Compare actual results with the target derived from current product contribution, not with an arbitrary industry benchmark. Record material changes to price, fees, landing page, promotion and attribution so trend breaks can be explained.
- Diagnose with CTR, CPC, conversion rate and average order value.
- Decide with contribution after advertising and new-customer economics.
- Validate growth with total revenue, cohorts and cash, not attribution alone.