Is a high ROAS always good? When the number misleads

"Is a high ROAS good or bad?" sounds like a trick question and is not. A high number answers a narrow question — how much came back per dollar spent on that campaign — and the three things it hides cost real money.

What a high number might be telling you

That the campaign works. It happens, and it is the case everyone assumes.

That the campaign is too small. A low budget picks off the easiest audience: people who already know the brand, already visited the site, already intended to buy. ROAS looks great and revenue does not grow.

That attribution is generous. Remarketing and brand search harvest intent created elsewhere. Credit goes to the campaign that closed, not the one that opened.

That the margin does not fit. A ROAS of 4 on a product with a 20% margin does not pay the operation, however pretty it looks.

Only the first case is good. The other three are the same number saying different things.

Break-even, which resolves half the question

Before judging any ROAS, calculate yours.

Break-even ROAS = 1 ÷ contribution margin.

A 40% margin gives break-even at 2.5. A 70% margin gives 1.43. A 20% margin gives 5.

That changes the whole conversation: the same ROAS of 3 is excellent in one operation and ruinous in another. Anyone comparing their ROAS to a competitor's is comparing different cost structures.

The full calculation is in contribution margin.

The scenario that confuses most: ROAS fell, revenue doubled

It happens in every operation that scales, and the wrong reaction is nearly automatic.

Spend rises, ROAS falls from 8 to 5, someone asks to go back to how it was. Except revenue doubled and 5 is still above break-even.

In that case the operation improved: it generates more absolute profit at a lower margin per sale. Going back means trading growth for a prettier number in the report.

A falling ROAS is only a problem in two situations: when it crosses break-even, or when it happens without revenue growth. The second is a genuine efficiency decline.

Telling those two apart requires reading ROAS and revenue together, always.

The signal that a high ROAS is too small

Three checks that give it away:

Low spend relative to the potential. If the campaign spends a fraction of what the market supports, the high ROAS comes from skimming.

Low frequency and flat reach. You are not touching new people.

Most revenue coming from remarketing. You are harvesting, not creating.

In all three the test is direct: raise budget and watch. If ROAS falls a little and revenue grows, there was demand sitting idle. If it collapses, the ceiling was close.

The criteria for when raising is worth it are in when to scale a campaign.

The platform's ROAS is not the business's ROAS

Two corrections almost nobody makes:

It uses revenue, not margin. US$ 10,000 of revenue with US$ 7,000 of variable cost is not US$ 10,000 of return.

It credits within the platform's window. A sale that would have happened anyway enters the numerator if it occurred inside the window and touched an ad.

The second point is why adding Meta's and Google's ROAS and comparing to real revenue frequently exceeds 100%. Both platforms credit the same sale. The mechanism is in attribution window.

A numeric example

Two campaigns from the same store, same 40% contribution margin, break-even at 2.5:

Campaign A has twice the ROAS and generates a third of the result. If the decision is made on the ROAS column, budget moves to A — and the operation shrinks.

What A is probably doing: picking off the easiest audience, with remarketing or brand in the mix. What it cannot do is grow: putting US$ 28,000 into it would drop ROAS to near B's, because the easy audience runs out.

The correct reading is the last row, not the third.

The metric ROAS does not replace

ROAS looks at the campaign. It does not know how much of your operation depends on advertising.

It is entirely possible to hold ROAS steady while the operation becomes more dependent on media every month, because non-ad sales are not keeping pace. The number that shows that is in ROAS or TACoS.

Reading both together is what separates "the campaign pays" from "the business is healthy".

When a low ROAS is correct

A top-of-funnel campaign has a low ROAS by nature: its job is introducing the brand, not closing sales.

Demanding ROAS from it and switching it off is the most costly mistake in a mature account. Three weeks later the bottom of the funnel worsens too, because nobody is entering — and the relationship between the two does not appear in any platform report.

The stage-by-stage reading is in funnel stage metrics and the full design in sales funnel.

How to use the number without being fooled

1. Calculate break-even with the real margin, from finance.
2. Set the target above it, according to the objective: profit or growth.
3. Read ROAS and revenue together, always, never one alone.
4. Split by funnel stage, and do not demand return from an introduction campaign.
5. Revisit the target when costs change. Shipping, tax or a different mix moves break-even.

Step 5 is the most forgotten: most operations chase a target set on a cost structure that no longer exists.

And a warning about step 1: the margin has to come from finance, not from a marketing estimate. The gap between the margin people believe they have and the real one is commonly ten to fifteen points, and inside that gap sits the difference between a profitable operation and one growing while losing money.

The summary

A high ROAS is good when it comes with growing revenue and sits above your break-even. It is a warning when it comes with a small campaign, dominant remarketing or a margin that does not fit.

And it is never the only thing to look at. The document that puts ROAS, revenue and stage side by side is on the paid media reporting page.