How to calculate ROAS — and why the number can still lie to you
ROAS is the metric most businesses live and die by, and it's genuinely simple to calculate. It's also one of the easiest numbers to be misled by, because the formula is honest while the inputs usually aren't. Let's do both halves: how to calculate it correctly, and how to read it without getting fooled.
The formula
ROAS = Revenue from ads ÷ Ad spend. That's it. Spend $1,000, make $4,000 from those ads, your ROAS is 4 (often written 4x or 400%). Every dollar spent returned four.
A quick cousin you'll hear alongside it: what counts as a "good" ROAS depends entirely on your margins. A business with 80% margins can thrive at 2x. A business with thin margins might need 5x just to break even after the cost of the product itself. There's no universal "good" number — there's only good for your economics. Anyone quoting a magic ROAS target without knowing your margins is guessing.
Where the number starts lying: the revenue input
The formula is clean. The word "revenue" in it is not. Because the revenue figure almost always comes from the ad platform itself — and the platform counts revenue generously, using its own conversions, its own attribution window, and its own habit of claiming sales it merely touched rather than caused.
So when Meta reports a 4x ROAS, it's dividing Meta's self-counted revenue by your spend. That revenue includes sales that would have happened anyway, sales another channel actually closed, and the same customer counted more than once. The division is correct. The numerator is inflated. A correct formula on an inflated input gives you a confident, wrong answer.
Platform ROAS vs. blended ROAS
Platform ROAS is what each dashboard reports for itself — Meta's ROAS, Google's ROAS — each using its own flattering count. Blended ROAS is your total real revenue divided by your total ad spend across everything. It can't double-count, because it starts from the actual money you made.
The two diverge hard. Add up every platform's claimed ROAS and you'll "prove" you made more than your store actually took in — the same double-counting that makes your dashboards disagree in the first place. Blended ROAS is the honest ceiling: it can't claim revenue that doesn't exist. The tradeoff is that it can't tell you which channel did the work — only that, together, they returned this much. Most healthy measurement uses blended as the truth anchor and platform numbers as directional hints, never the reverse.
ROAS vs. CAC: two questions, not one
ROAS and CAC (customer acquisition cost) get used interchangeably, but they answer different questions. ROAS asks "how much revenue did this spend return?" CAC asks "how much did it cost to acquire a customer?"
The gap between them is where a lot of businesses fool themselves. A campaign can post a strong ROAS while quietly acquiring almost no new customers — because it's being credited with revenue from people who already knew you and would have bought regardless. High ROAS, high hidden CAC for actual new customers. If you only watch ROAS, you can pour budget into a channel that's efficiently re-selling to people you already had.
The limit ROAS can't cross
Here's the honest ceiling on the whole metric: ROAS, even calculated perfectly on blended numbers, tells you the ratio of revenue to spend. It does not tell you what would have happened without the spend. A channel showing 3x might be genuinely creating that revenue — or mostly harvesting demand that already existed and would have converted anyway.
Separating "revenue that happened alongside the ads" from "revenue the ads actually caused" is a different measurement entirely, called incrementality — and it's the question ROAS was never built to answer. (We cover that in its own piece.) For now, the useful discipline is simpler: calculate ROAS on blended, real revenue; read it as directional; and never let a platform's self-graded number set your budget on its own. For the wider picture of why these numbers conflict across your tools, see our guide to why your marketing tools never report the same revenue.
The short version
- Calculate it right: revenue from ads ÷ ad spend. "Good" depends on your margins, not a universal target.
- Distrust the numerator: platform-reported revenue is inflated by design.
- Anchor on blended: total real revenue ÷ total spend can't double-count.
- Watch CAC alongside it: strong ROAS can hide weak new-customer acquisition.
Every one of those problems traces back to the same root: the revenue in your ROAS was counted by tools that each see only their slice of the customer. Resolve the customer to one identity first — which is what Chapter does — and ROAS finally gets a numerator you can trust.