A Good ROAS Is Whatever Your Margin Says It Is

The question arrives in every new client engagement within the first hour, phrased almost identically: what's a good ROAS? The questioner wants a number, ideally one number, ideally four. The 4x figure has become advertising folklore, repeated across a thousand blog posts until it acquired the texture of fact. Here is the problem with it, and I can show it in two lines of arithmetic. A company with 80 percent gross margins earning a 2x ROAS is making money on every sale. A company with 20 percent margins earning a 4x ROAS, the folklore's "good" number, is losing money on every sale and celebrating. The benchmark is not just imprecise. For a meaningful share of the businesses using it, it points in the wrong direction entirely.

The math deserves to be written down once, plainly, because it governs everything. Your break-even ROAS is one divided by your gross margin. Sixty percent margin: break-even at 1.67. Thirty percent margin: break-even at 3.33. Ten percent margin, common in electronics and grocery: break-even at 10, a number that would get a media buyer a raise at most companies while the finance team quietly bled. Anything above break-even contributes; how far above you need to run depends on what else each order must fund, overhead, shipping subsidies, returns, and what you believe about repeat purchase. A subscription business that keeps customers for years can rationally run below break-even on the first order, which is a decision about payback period, not a ROAS at all. The moment you see the formula, every industry benchmark chart becomes what it always was: a table of other people's margins.

Watch the formula work on a real P&L, because the abstraction hides how violent the differences are. Take two ecommerce brands each spending $50,000 a month at a reported 3.5x ROAS, $175,000 in attributed revenue. Brand A sells supplements at 75 percent gross margin: break-even ROAS 1.33, so at 3.5 they are generating roughly $81,000 of gross profit after ad spend, a thriving program with room to push spend and accept a lower marginal ratio. Brand B sells consumer electronics at 22 percent margin: break-even 4.55, so the identical 3.5x means every attributed order destroyed value, about $11,500 of gross profit against $50,000 of spend. Same platform, same dashboard, same "good" ROAS, and one of these companies is celebrating its own decline. I have sat in both meetings. Only one of them had ever written down the break-even.

Then there is the deeper problem, the one that survives even after the margin math is fixed: the R in ROAS is not what you think it is. Platform-reported ROAS counts attributed revenue, and attribution is a claims process, not an audit. Every platform's model is structurally generous to itself, harvesting credit for customers who searched your name, clicked a retargeting ad on the way to a purchase they had already decided on, or bought after an "engaged view." I have run the reconciliation dozens of times: add up each platform's claimed revenue and the total routinely reaches 150 to 250 percent of what the bank statement recorded. This is the machinery I took apart in the number your CFO actually believes, and it is why sophisticated advertisers increasingly manage to MER, total revenue over total ad spend, or better, to incrementality tests. The dashboards' ROAS can be a useful directional signal within a platform. As a statement of business truth, it is a press release.

ROAS has a third defect that costs growth rather than money: it is a ratio, and ratios are maximized by shrinking. The safest spend in your account, brand search and retargeting, produces spectacular ROAS precisely because it harvests demand that mostly existed already. Cut everything else and your ROAS soars while your revenue line flattens, the exact pattern I described in the metric going up is hiding the one going down. Every incremental dollar of real growth spend, new audiences, new channels, upper funnel, arrives with a worse ratio than the dollar before it. A business run to maximize ROAS will therefore systematically starve its own future, efficiently. The right question is never what ROAS maximizes the ratio; it is what marginal ROAS you can accept on the last dollar while still hitting your growth target, and that number should sit close to break-even, on purpose. If your blended ROAS is far above your break-even, you are probably not efficient. You are probably underinvested.

For businesses with repeat purchase, upgrade the metric once more before you set targets: compute ROAS on contribution margin and judge it against payback period. A coffee subscription acquiring customers at 0.8x first-order ROAS looks like arson on the dashboard and like arbitrage in the cohort table, if the second and third orders arrive on schedule; the discipline it requires is the one I described in the subscriber you buy on discount leaves on schedule, knowing your retention curve well enough to spend against it honestly. First-order ROAS is a cash-flow constraint, not a verdict, and companies that grade acquisition on it alone systematically lose their best long-term customers to competitors who did the cohort math.

So here is the answer the query deserves, in usable form. Compute your break-even: one over gross margin. Set your target above it by whatever your overhead and payback tolerance require; for most healthy ecommerce that lands between 1.3 and 2 times break-even, which for a 50 percent margin business means a target near 3, not because three is magic but because your P&L said so. Measure it with revenue you can reconcile to the bank, not the platform's claims. Then manage the marginal number, not the average, and treat a suspiciously beautiful blended ROAS as a symptom of timidity. A good ROAS is not four. A good ROAS is the worst one your margin can afford at the growth rate you want, and the discipline to know the difference is worth more than any benchmark chart ever published.