Revenue attributed to each ad dollar
Return on ad spend is attributed revenue divided by advertising spend. A result of 4.5× means each advertising dollar is associated with $4.50 of revenue. The companion percentage expresses spend as a share of attributed revenue. Neither figure subtracts product cost or overhead.
A 4.50× campaign is not $42,000 of profit
With $12,000 of media spend and $54,000 of attributed revenue, ROAS is 4.50×. Advertising consumes 22.22% of that revenue, leaving $42,000 before cost of goods, fulfillment, agency fees, returns, and tax. Revenue after ad spend is not profit.
Set one attribution rule before comparing
Use one attribution rule across comparisons. Last-click reporting, view-through reporting, and platform-reported conversions can give different revenue to the same campaign. Refunds, cancelled orders, and repeat purchases also need a stated treatment or a high ROAS can be a reporting artifact.
Revenue and acquired customers are different outcomes
This page is not a customer-acquisition-cost calculation. ROAS treats revenue as the outcome; CAC counts acquired customers and can incorporate sales labor as well as marketing. A 4.5× ROAS can still lose money where gross margin is thin or where the campaign only captures buyers who would have purchased anyway.
Derive the break-even advertising threshold
Set a break-even ROAS from contribution margin. If a sale retains 35 cents of every revenue dollar before advertising, the business needs roughly 2.86× revenue per ad dollar just to cover that spend. The required threshold rises when fulfillment or agency costs are included.
Test whether the channel added demand
Channel-level ROAS needs a denominator that includes the money required to buy the exposure. Platform media spend is sometimes only part of the cost; creative, affiliate commission, feed management, and agency fees can matter. Separate branded search from prospecting because branded demand often has a very different incremental value. A campaign can report excellent attributed revenue by closing existing demand while adding few new buyers. Holdout tests or geographic comparisons are better evidence when incrementality is the decision.
Keep the attribution window visible in every report. Seven-day and thirty-day revenue can describe the same campaign differently, so an unexplained window change can make a trend look better or worse than it is. Preserve the window when comparing channel history.
A campaign report should retain the currency, date range, conversion definition, and refund cut-off used in the numerator. Changing any one of those can alter ROAS without a change in buying behaviour. That record makes a later budget decision auditable rather than cosmetic.
Separate prospecting from retargeting before moving budget. Retargeting can claim revenue from people already near purchase, while prospecting must be judged by a different evidence trail.
Break-even ROAS must use the contribution that remains after every cost included in the decision. With a 35% contribution margin, each $54,000 of attributed revenue supplies $18,900 before advertising. The $12,000 media spend leaves $6,900 after those two items, so the observed 4.50x exceeds the 2.86x threshold computed as one divided by 0.35. If fulfillment, payment fees, and agency work reduce contribution to 25%, the threshold becomes 4.00x; the same campaign is much less comfortable. A headline ratio has no meaning until that denominator is named.
Attribution is a reporting rule, not proof that an impression caused a sale. A last-click report can credit a search ad for an order after email, brand recognition, and direct visits created most of the demand. A view-through setting can add credit without a click. Choose the click window, view window, conversion event, currency, refund cut-off, and treatment of repeat buyers before comparing campaigns. Keeping those definitions constant prevents an apparent improvement that was caused only by a changed dashboard setting.
Budget decisions should separate an efficiency measure from an incrementality question. Prospecting campaigns try to create or reach new demand; retargeting often reaches people already near purchase. Both can report revenue divided by spend, yet reducing one and increasing the other can change future demand in ways the immediate ratio does not reveal. Where the decision is large, compare a holdout group, a geographic test, or another credible counterfactual. The tool calculates attributed revenue; it cannot establish what would have happened without the advertising.
A reporting table should preserve the numerator as carefully as the spend. State whether the $54,000 is gross order value, net sales after refunds, first-purchase revenue, or revenue including repeat orders, and keep the conversion window beside it. A campaign can appear to improve when the window expands or when refunds have not yet matured. For channel allocation, rerun the same definition after enough time for cancellation and return behaviour to appear. The calculator then remains an auditable ratio instead of a moving platform figure whose favourable change cannot be explained.
Check the spend export for credits, make-goods, taxes, and currency conversion before dividing. A platform total may be an authorization rather than settled media cost, while revenue may be recorded after a return period. Using one settled date range for both sides does not prove causation, but it removes a basic reconciliation error. Keep campaign identifiers with the calculation so a change in targeting, bid strategy, or tracking can be separated from a genuine performance movement.