What is incrementality in marketing?
Incrementality measures the sales a campaign caused, not just the sales it claimed. Here are the lift and iROAS formulas, how Google and Meta run lift tests, a worked example, and when incrementality matters most.
Incrementality is the share of conversions or revenue that happened only because of a campaign: the sales you would not have made without it. It is measured by comparing a test group that could see the ads with a control group that could not. The gap between the two is the incremental lift.
How do you measure incrementality?
Split people or regions at random into a test group and a holdout, run the campaign to the test group only, then compare results over the same period.
| Metric | Formula |
|---|---|
| Incremental conversions | Test conversions − control conversions (equal-sized groups) |
| Lift | Incremental conversions ÷ control conversions |
| Incremental ROAS (iROAS) | Incremental revenue ÷ ad spend |
Google's Conversion Lift splits an audience into people who see your ads and a control group who don't, and says "the difference in conversions between these 2 groups will tell you the lift." Meta's March 2026 measurement update calls incrementality tests like its Conversion Lift "the gold standard in measurement."
Incrementality example
An illustrative UK retailer runs a four-week test with two groups of 100,000 people. The test group makes 1,300 purchases, the control group 1,000. That is 300 incremental purchases, a 30% lift. At £80 per order, incremental revenue is £24,000 on £12,000 of spend: an iROAS of 2.0, even if the platform reports a ROAS of 4.0.
When incrementality matters
- Retargeting and brand search: they claim many conversions that would have happened anyway.
- View-through heavy channels: video and social results often lean on view-through conversions.
- Before scaling: a channel with high reported ROAS but low lift is a poor place for the next budget increase.
Tests need enough volume and a clean holdout. Results apply to the period and audience tested.
Related terms
See marketing mix modeling, data-driven attribution and marketing ROI.
How Grower helps
Grower's attribution engine de-duplicates conversions into one count and lets you switch between last click, first click, linear, time decay, position-based and custom models. Where models disagree sharply on a channel, that channel is a good candidate for a lift test.
Frequently asked questions
What is the difference between incrementality and attribution?
Attribution divides credit for conversions that already happened among the touchpoints on the path. Incrementality asks how many of those conversions the campaign caused, using a control group that never saw it. Attribution is continuous and cheap to run. Incrementality needs a designed test, but it is the stronger evidence of cause.
How do you run an incrementality test?
Use a platform lift study, such as Google's Conversion Lift or Meta's Conversion Lift, or design a geo test that pauses ads in matched regions. Google offers user-based and geography-based versions, though it says Conversion Lift isn't available for all accounts. Fix the test length and success metric before you start, and avoid changing budgets mid-test.
What is a good incremental ROAS?
It depends on your margins. An iROAS above your break-even ROAS means the campaign made money it would not otherwise have made. Below break-even, the channel loses money at the margin, whatever the platform's reported ROAS says. Work out your break-even point first, then judge each test against it.
Work out the return a campaign needs to clear with the marketing ROI calculator, or book a Grower demo to compare models on one de-duplicated count.