Ad reviews often hit the same conflict: the platform shows acceptable ROAS, but finance says the business is not making money. The reason is that platform ROAS and break-even ROAS answer different questions. Platform ROAS is conversion value divided by ad spend inside the ad system. Break-even ROAS is the business threshold after COGS, fulfillment, payment fees, discounts, refunds, and support are considered.
Google Ads Target ROAS uses reported conversion value to optimize bids toward an average return target. That is useful, but it does not automatically know your real profit, refunds, inventory pressure, support cost, or cash flow. Unless your value signal is mature, platform ROAS is not profit ROAS. Budget decisions need both numbers, read separately.
The two ROAS numbers answer different questions
Platform ROAS asks: inside this attribution window and ad system, how much recorded conversion value did each dollar of ad spend produce? It is useful for account learning, traffic quality, creative or keyword comparison, and bid-input stability. Break-even ROAS asks: how much order revenue must each dollar of spend generate so the business does not lose money after real order costs?
Both numbers matter, but they are not substitutes. High platform ROAS may come from remarketing, brand search, returning customers, or generous attribution. Low platform ROAS may come from cold acquisition, product exploration, or early creative testing. First identify the job of the campaign, then compare it with the right business threshold.
Platform ROAS inputs and blind spots
Platform ROAS depends on event firing, conversion value, currency, attribution window, deduplication, consent state, and modeling. Google Ads Target ROAS requires conversion values, and different campaign types have conversion-volume and value-history guidance. Google Ads can predict potential conversion value only from the signals you provide.
Blind spots include refunds, return shipping, payment fees, fulfillment cost, stockouts, support cost, discount stacking, new versus returning customer value, and channel overlap. If purchase value is gross order revenue and refunds or product costs are not reflected, platform ROAS can look healthier than profit. If purchase events duplicate or values are wrong, ROAS becomes directly misleading.
Break-even ROAS formula and example
The basic formula is break-even ROAS = 1 / contribution margin rate. Contribution margin should use order revenue minus variable order costs. Suppose price is $100, COGS $35, fulfillment and packaging $10, payment and platform fees $4, shipping subsidy $8, discount $5, and expected refund loss $3. Contribution profit is $35, contribution margin is 35%, and break-even ROAS is about 286%.
If the platform reports 250% ROAS, the campaign may show orders but still sit below first-order break-even. If it reports 400%, do not automatically scale. Check order quality, refund rate, inventory, new-customer share, and cash recovery. Break-even is the floor for not losing money on the first order, not proof that unlimited scaling is safe.
When reported ROAS can still lose money
High ROAS can still lose money for several reasons. Conversion value may include tax, shipping, or revenue that will later be refunded. Orders may be concentrated in low-margin or over-discounted SKUs. Platform attribution may credit remarketing or existing demand without creating incremental sales. Refunds, disputes, or support issues may cluster in that channel. Inventory or fulfillment constraints may create downstream cost.
A useful budget meeting should not report only one ROAS number. Include platform ROAS, backend revenue, contribution margin, refund rate, AOV, new-customer share, inventory cover, and acceptable CPA. Then the team can decide whether the issue is ads, pricing, page trust, tracking, or order quality.
Use the ROAS calculator without over-trusting it
A ROAS calculator should first clarify the break-even line, then support the budget action. Enter price, contribution margin, refunds, fulfillment, payment fees, and ad spend. Calculate break-even ROAS and acceptable CPA, then compare platform ROAS against those lines. If platform ROAS is above the healthy line, inspect order quality and sample size. If it is below break-even, do not only edit ads; inspect pricing, discount, page, and purchase tracking.
The calculator is not the final judge. It cannot decide attribution pollution, brand-search share, sample quality, learning-period noise, or whether LTV is proven. Its value is making the profit boundary visible so the team stops using one platform-reported number for every budget decision.
ROAS formula card
| Metric | Formula | Answers | Does not answer |
|---|---|---|---|
| Platform ROAS | reported conversion value / ad spend | Return the ad system can see | True profit, refunds, inventory, and cash pressure |
| Break-even ROAS | 1 / contribution margin rate | Revenue needed to avoid first-order loss | Whether long-term scaling is attractive |
| Profit ROAS | contribution profit / ad spend | Whether ads leave operating profit | Creative and traffic leading indicators |
| Acceptable CPA | contribution profit or allowed loss | Maximum ad cost per order | Whether attribution is accurate |
Turn this into a repeatable operating loop
Do not treat this article as a one-time reading task. Turn the decisions around The two ROAS numbers answer different questions / Platform ROAS inputs and blind spots / Break-even ROAS formula and example into a small operating loop that your team can run before a launch, after a platform change, or when performance data starts to look inconsistent. The practical output should be a dated note, a checklist status, and a short owner comment, not a vague memory that someone "looked at it." That habit gives future reviews something concrete to compare against.
The table on ROAS formula card starts with Platform ROAS / Break-even ROAS / Profit ROAS. Use those rows as the minimum evidence set. If one row cannot be verified, mark the page, campaign, feed, event, or policy as not ready and write down the exact missing proof. This protects the team from a common ecommerce failure mode: a visible metric moves, everyone reacts, but no one knows whether the store, tracking, content, or offer was actually in a valid state.
After you apply the checklist, connect the result to the linked Ecomwith tool, tutorial, or answer page. The blog should help you make the first decision; the next route should help you calculate, audit, document, or repair the issue. That is also what makes the page useful for search and AI discovery: it states the operating question, shows the evidence, and then points to the next page where the reader can act with more context.