Incrementality in advertising is the share of sales or conversions that would not have happened without the ads. You find it with a test: compare a group that saw the ads with a group that did not. 60% incrementality means that of every 100 euros a campaign claims, 40 would have come in anyway.
Incrementality in advertising is the share of sales or conversions that would not have happened without the ads. You find it with a test: compare a group that saw the ads with a group that did not. 60% incrementality means that of every 100 euros a campaign claims, 40 would have come in anyway.
An ad platform claims every order where the customer touched an ad somewhere along the way. It cannot tell whether the ad changed their mind or whether they simply searched for a store they already knew on the way to checkout. Incrementality answers the one question that matters when you set a budget: what would happen if we switched this campaign off?
That is why you will not find it in Google Ads or Meta Ads Manager. Attributed revenue is the output of attribution, a rule for handing out credit. Incrementality is the output of an experiment.
You need two groups that differ only in whether they saw the ads: a test group and a control (holdout) group. The revenue gap between them is the lift. You then compare the lift with what the campaign claimed and with what it cost.
| Lift | revenue with ads − revenue without ads | Example: test 53,000 EUR − control 43,000 EUR = 10,000 EUR |
| Incrementality | lift ÷ attributed revenue × 100 | Example: 10,000 ÷ 20,000 EUR attributed × 100 = 50% |
| Incremental ROAS | lift ÷ ad cost × 100 | Example: 10,000 ÷ 4,000 EUR cost × 100 = 250% |
In this example the campaign reports a ROAS of 500% but really added only half of that. Budgeting on the 500% means overpaying.
The control group has to be comparable: same period, similar size, similar behavior. Without one you end up comparing against last month, and seasonality leaks into the result. Our CarDetailer case study shows how much: leads from ads grew 155% between January–June and 1 Aug–13 Sep 2026. But in the same windows a year earlier they grew 87% with no changes at all. Adjusted for season and trend, the new website and the new campaign account for roughly +36% (Google Ads).
Switch-off test (holdout over time). Turn one campaign off completely for one to two weeks and watch total store revenue from your back office, not the campaign numbers. It costs nothing and works for any account. The weak spot: seasonality and demand swings feed into the result, so you need a comparable period.
Geo test. Pause ads in some regions and keep them running in the rest. Compare regions that have historically moved together. Both groups go through the same season at the same time, so it cancels out. Open-source tools such as Meta's GeoLift help with the analysis. In a small market there are few regions, so a geo test only picks up a large effect.
Platform lift study. Google Ads and Meta can both split people into a group that sees the ads and a control group that does not. Google Ads offers Conversion Lift based on users or on geography; it is not available to every account, and access goes through a Google representative. Since 2025 Google says a test can run on a budget from 5,000 USD. Meta runs lift studies in its Experiments tool and, since 2025, also offers an incremental attribution setting that estimates incremental conversions with a model.
Enter the result of your test. The calculator returns incrementality, incremental ROAS and the profit from the lift after ad cost.
There is no universal number. What matters is whether incremental ROAS beats your break-even ROAS, which is 100 divided by your margin in percent. A campaign with 40% incrementality can make money if its claimed ROAS is high and the margin is good. A campaign with 100% incrementality can lose money if it is expensive.
The logic is simple: the closer someone is to buying when they see the ad, the more likely they would have bought without it. Expect low incrementality from searches for your own brand and from retargeting people with a full cart. It tends to be higher when you reach people who do not know you yet. Only a test can confirm it, though, not a guess.
If your product margins vary, calculate the lift on margin directly. It is the same idea as POAS, just cleaned of orders that would have come in anyway.
These terms get mixed up in reports. Attribution hands out credit, ROAS turns that credit into a ratio, and incrementality checks whether the credit was earned.
| Term | Question it answers | Where the number comes from | When to use it |
|---|---|---|---|
| Incrementality | What the ads actually added | Test with a control group | Budget, pausing a campaign or channel |
| Attribution | Who gets credit for a conversion | Model and window in the tool | Running campaigns inside a platform |
| ROAS | Attributed revenue per unit of spend | Ad platform | Quick check against the threshold |
| POAS | Margin per unit of spend | Ad platform + margin data | Catalogs with mixed margins |
| Marketing mix model | Combined effect of all channels | Statistical model on long data series | Large budgets across many channels |
Terms you need to calculate and explain incrementality.
Once you pause a campaign, its ROAS drops to zero. That tells you nothing. Watch total revenue and orders in your store back office and compare them with the control group or a comparable period.
On half the budget a campaign buys different auctions than on the full budget, and the difference cannot be scaled up to the whole. For a test, switch the campaign off completely or use a control group that never sees the ads.
A result holds for a specific campaign, period and budget. A January test does not hold for Black Friday, and a brand campaign result does not hold for Performance Max. Repeat the test when budget, catalog or season change significantly.
The share of sales or conversions that would not have happened without the ads. It is measured with a test against a control group, not read from an ad platform. 60% incrementality means 40% of the attributed revenue would have come in anyway.
Switch one campaign off completely for one to two weeks and track total store revenue against a comparable period or against regions where ads keep running. Then compare the lift with what the campaign claimed over the same time.
Attribution decides which channel gets credit for a conversion. Incrementality asks whether the conversion would have happened without the ads at all. Attribution can give credit to a campaign that caused nothing.
The revenue lift from a test divided by ad cost, in percent. Unlike regular ROAS, it only counts revenue the ads actually added. You compare it with your break-even ROAS.
Yes. Google Ads has Conversion Lift based on users or geography, though not every account has access. Meta offers lift studies in Experiments and an incremental attribution setting that estimates incremental conversions with a model.
Send this and you get a 90-day strategic plan for raising the performance of your marketing. Free, within 48 hours of account access.
Jan gets back to you within 48 hours, usually with two questions and a time. Then the 90-day plan starts.