Incremental ROAS: the only return number a CFO will actually trust
Paid Media

Incremental ROAS: the only return number a CFO will actually trust

Reported ROAS answers what the platform observed. Incremental ROAS answers what the spend actually caused. Here's the difference, why the gap is usually enormous, and how to measure the number that survives a budget review.

Every platform in your account reports a ROAS, and every one of them is answering the wrong question. Reported ROAS tells you what the platform observed, how much attributed revenue it can associate with the ads it served. The question that actually governs a budget is different: what did the spend cause? That number is incremental ROAS, and the gap between the two is where most media budgets quietly leak.

Two numbers that sound the same and aren’t

Reported ROAS is attributed revenue over spend. If a platform can link a conversion to an impression or a click it served, it counts the revenue. Simple, fast, and structurally generous, because it credits the ad for conversions that were going to happen regardless.

Incremental ROAS is caused revenue over spend. It counts only the conversions that would not have occurred without the ad. Everything the campaign merely intercepted, the retargeted cart that was coming back anyway, the branded searcher who already knew your name, the loyal customer mid-repurchase, gets stripped out. What’s left is the revenue the spend actually produced.

Same shape, opposite meaning. One measures observation; the other measures impact.

Why the gap is usually enormous

The gap between reported and incremental ROAS is the demand that already existed. And on the channels that report the best ROAS, that’s most of the number.

Think about where reported ROAS looks strongest: branded search, retargeting, bottom-of-funnel social. Those are exactly the placements that advertise to people already in motion. High reported ROAS on those channels isn’t a sign of high impact, it’s a sign of high pre-existing intent in the audience. The ad showed up at the finish line and claimed the race. Run the incrementality test and a big share of that revenue turns out to have been arriving with or without the spend.

Reported ROAS is highest exactly where incrementality is lowest, the bottom of the funnel, where the ad intercepts demand it didn’t create. The prettiest number in the account is often the least incremental.

How you actually measure it

You can’t compute incrementality from a dashboard, because the dashboard only sees the exposed group. You need a comparison, a set of people or markets that didn’t see the ad, and the difference between them is the lift.

Three practical ways to build that comparison:

Geographic holdout. Run the channel in a set of matched markets and hold it dark in another set. Compare total outcomes. This is the workhorse for channels with no click to measure, like CTV and out-of-home.

Audience suppression. Withhold ads from a random slice of otherwise-eligible users and compare their conversion rate to the exposed group. This is how you test retargeting and branded search honestly.

Platform lift study. The ad platforms offer their own randomized lift tests. Useful, with the caveat that you’re asking the platform to grade itself, worth triangulating against your own geo tests.

Whichever you use, the logic is the same: measure the difference between exposed and unexposed, attribute that difference to the spend, divide. The comparison group is the entire ballgame.

Translate it into the number that matters

iROAS by itself still isn’t the decision. There’s no universal “good” iROAS, because it has to clear your margin, not a benchmark. The move is to convert incremental revenue into contribution: at your contribution margin, does the demand this spend actually created exceed what it cost?

That reframing changes which campaigns look like winners. A channel with a modest iROAS on high-margin products can beat one with a flashier iROAS on thin-margin goods. Revenue-based ROAS targets hide this; incrementality plus margin exposes it. This is the same discipline the Measurement lane applies across the whole plan, and it’s why branded search deserves its own holdout before you trust its reported return.

The takeaway

Reported ROAS is fine for steering day to day. It is not fine as the basis for a budget decision, because it systematically over-credits the spend that intercepts existing demand and under-credits the spend that creates it. Build the comparison group, measure the lift, translate it to contribution, and put that number in front of finance. It’ll be lower than the dashboard’s. It’ll also be the first return figure in the room that survives the follow-up question.

Frequently asked questions

What is incremental ROAS?

Incremental ROAS (iROAS) is revenue that would not have happened without the spend, divided by the spend. It isolates the causal contribution of a campaign, the sales it actually created, as opposed to reported ROAS, which counts every conversion the platform could associate with an ad it served, including the ones that would have happened anyway. iROAS is usually well below reported ROAS, and it’s the figure that holds up in a budget conversation because it answers what changed, not what was observed.

How is incremental ROAS different from reported ROAS?

Reported ROAS is attributed revenue over spend, it credits the ad for any conversion it can link to an impression or click. Incremental ROAS is caused revenue over spend, only the conversions that wouldn’t have occurred without the ad. The difference is the demand that was already going to convert: existing intent the campaign intercepted and took credit for. On high-intent channels like branded search and retargeting, that gap is often most of the reported number.

How do you measure incremental ROAS?

With an experiment that creates an unexposed comparison group: a geographic holdout (run the channel in some matched markets, hold it dark in others), an audience suppression (withhold ads from a random slice of eligible users), or a platform lift study. You measure the difference in conversions between exposed and unexposed groups, attribute that difference to the spend, and divide. The comparison is the whole point, without a group that didn’t see the ad, you can’t separate cause from coincidence.

Why is incremental ROAS usually lower than reported ROAS?

Because reported ROAS includes demand that already existed. Platforms credit ads for conversions from people who were already going to buy, the retargeted cart, the branded searcher, the loyal customer, and those conversions inflate the reported return without the ad having caused them. Strip them out and only the genuinely created revenue remains, which is why iROAS lands lower, sometimes dramatically, especially on bottom-of-funnel channels.

What's a good incremental ROAS?

The honest answer is that it depends on your margin, not on a universal benchmark. The useful move is to translate iROAS into contribution: does the incremental revenue, at your contribution margin, exceed the spend that produced it? A channel with a modest iROAS on high-margin products can be a better investment than one with a flashier iROAS on thin-margin goods. The number to beat isn’t a target ROAS, it’s profitable incrementality.

Disagree with any of this? That's the useful conversation. LinkedIn  ·  Email  ·  More writing