ROAS is the most-cited metric in paid media and one of the most misunderstood. Both halves of the fraction are narrower than they appear.
The numerator is attributed gross revenue — before cost of goods, shipping, returns, and discounts. The denominator is media spend only, excluding management fees, creative production, and tooling.
It remains genuinely useful for one job: comparing two ad sets inside the same account over the same window, where the attribution model and cost exclusions are identical for both. It is close to useless for deciding whether an entire channel is profitable.
Formula
ROAS = Attributed Revenue ÷ Ad Spend
Break-even ROAS = 1 ÷ contribution margin percentage. At 40% contribution margin, break-even is 2.5x — a 2.0x ROAS is losing money.
Why ROAS matters
Most budget decisions are made from ROAS, so its blind spots become the business's blind spots. Understanding exactly what it excludes is what lets you keep using it for the narrow job it's good at without letting it drive strategy.
Why summing platform ROAS breaks
A customer sees an Instagram ad Monday, searches your brand Wednesday, clicks a Google ad, and buys. Meta claims the conversion under view-through. Google claims it as last click. Your email tool claims it too. Each is honest within its own model, and adding them up overstates reality. In accounts we audit, summed platform-claimed revenue typically runs 130–160% of actual revenue.
Benchmarks
- Break-even at 40% contribution margin
- 2.5x
- Break-even at 30% contribution margin
- 3.3x
- Typical platform overstatement
- 30–60%
Ranges drawn from Digital Squad client accounts and published industry data. Treat them as orientation, not targets — your category may differ substantially.
Common mistakes
Adding ROAS across channels
Attribution models overlap, so the sum double-counts. Compare summed platform revenue to your actual revenue in Shopify or Stripe — the gap is the size of the error you're budgeting against.
Treating a high ROAS as success
A very high ROAS usually means you're only buying people who would have converted anyway — typically brand search and remarketing. It's often a sign you're under-investing in incremental demand, not a sign of health.
Ignoring the attribution window
A 7-day-click ROAS and a 28-day-click-1-day-view ROAS are not comparable numbers. Changing the window changes the metric without changing performance at all.
Optimizing to it without a margin model
Without knowing break-even ROAS, a target is arbitrary. Teams routinely set a 3x target with no idea whether 3x is profitable for their products.
Where we work on this