Incremental ROAS calculator.
Meta reports its orders. Google reports its own. The email platform reports its own. None of them knows what the others counted, and none of them knows what would have happened anyway. This calculator puts those numbers next to the shop's own and shows what is left.
Reported ROAS
8.13×
Incremental ROAS
4.06×
Incremental revenue
€65,000
Share of revenue claimed
89.7 %
What the channels are claiming
The channels claim 89.7 % of everything the shop made. Whatever is left is organic, direct and returning customers — plus everything the channels would have got anyway.
Incremental ROAS is reported ROAS multiplied by the share that was genuinely caused by ads. The share is the only number here you cannot read off a dashboard — it has to be measured with a holdout.
01Why the numbers rarely add up
Every channel counts every order it touched. If a customer saw a Facebook ad, clicked a Google ad and opened an email before buying, that single order appears in three reports. Add them together and you can get a number larger than the shop's total revenue, without anyone having made an arithmetic error.
The second problem is the heavier one. An attributed order is an order that came after the ad, not necessarily because of it. Customers who would have bought anyway are counted. Returning customers are counted. People searching for your brand name are counted. None of it is fraud — it is just not what you paid for.
02How it is calculated
Reported ROAS is attributed revenue divided by spend. Incremental ROAS is that same number multiplied by the share the ads genuinely caused. If half the attributed revenue is incremental, your ROAS halves.
That share is the only number in the calculation you cannot look up. Everything else sits in a dashboard. Which makes it the only one worth spending money to measure.
03How to measure the share instead of guessing it
Meta already has the answer built in. Under Experiments in Ads Manager you can run a Conversion Lift test, where part of your audience is deliberately held back from seeing ads. The difference between the two groups is the lift — what the ads caused, rather than what they merely stood next to. Meta recommends a minimum budget of 5,000 dollars and at least 28 days to reach statistical significance, and a decent data source underneath, typically the Conversions API.
The most telling part sits in Meta's own help text: results from a lift test are not meant to be compared with the numbers in Ads Manager. The two use different methods — one measures against a withheld control group, the other against an attribution window. So there is no mystery in them disagreeing. They are not measuring the same thing.
If you want this as continuous optimisation rather than as a study, Incremental Attribution is set at the ad set level. It uses the same randomised control group logic, but lets delivery optimise toward incremental lift instead of toward attributed conversions.
04What it costs to know
A lift test can raise your CPM while it runs. That is not a fault but a consequence of the method: part of the audience is held out of the auction, and the impressions that remain are bought in a smaller market. That is the price of a number that holds.
In return, it is the only way to find out whether that campaign showing ROAS 8 returns eight or two. The difference between those two answers is the entire budget.
Questions
What is a realistic incremental share?
It depends entirely on channel, campaign type and how well known the brand is. Retargeting and brand search sit lowest, because they largely reach people who were already on their way. Cold prospecting sits highest. But a guess is a guess — the point of this calculator is to show how much the answer moves, not to supply it.
Can the channels report more revenue than the shop actually made?
Yes, and it happens often. Each channel counts every order it touched, and none of them deducts for the others. If the total exceeds the shop's actual revenue, nobody miscalculated — those numbers were never meant to be added together.
Is incremental ROAS the same as POAS?
No. POAS is about what is left after cost of goods and other costs — whether the revenue is profitable. Incremental ROAS is about whether the revenue was caused by the ads at all. The two questions are independent, and a campaign can fail both.
Do you need a Meta representative to run a lift test?
No. Conversion Lift can be set up yourself under Experiments in Ads Manager, provided the account meets the minimum spend and conversion requirements. If you do have a Meta representative, they can help interpret the result, but it is not a prerequisite for getting started.
Read on
Incrementality testing in practice: geo-lift and holdout
Attribution can't tell you whether a sale would have happened anyway. Here's how to run a geo-lift or holdout test and measure the real added effect.
What is ROAS?
ROAS (Return on Ad Spend) is revenue per ad euro. We explain how ROAS is calculated, why the number deceives, and what a good ROAS actually is.
What is POAS?
POAS (Profit on Ad Spend) measures the profit each ad euro returns — not revenue. How it differs from ROAS, and why it's the number that decides how you scale.