Incrementality Lift Calculator
Most reported conversions would have happened anyway. Incrementality testing finds the ones your spend actually caused by comparing a treated group against a held-out control. Enter both groups to get true lift and incremental ROAS.
Incremental lift = (treatment conversion rate − control conversion rate) ÷ control conversion rate × 100%. It comes from a controlled test — a geo or audience holdout — where a treatment group sees the campaign and a matched control does not. The gap is the effect your spend caused, not merely the conversions it was present for. This tool returns the lift percentage, the count of incremental conversions, and the incremental ROAS, which is the honest efficiency number once you remove conversions that would have happened anyway.
Incrementality Lift Calculator inputs and result
How to use this calculator
- Run a controlled testSplit your audience into a treatment group that sees the campaign and a matched control group that does not. A geo holdout (whole markets on vs. off) is the most robust; audience holdouts work where geo is impractical.
- Enter both groups’ sizes and conversionsPut in the size and conversions for treatment and control. The tool computes each group’s conversion rate so the comparison is fair even if the groups differ in size.
- Read the incremental liftLift is the percentage by which the treatment rate beats the control rate. This is the effect the spend caused, not the total conversions it was credited with.
- Add spend and value for incremental ROASEnter campaign spend and value per conversion to get incremental ROAS — almost always lower than last-click ROAS, and far more honest.
- Check significance before actingA small lift on small groups can be noise. Confirm the test had enough scale and duration to detect a real effect before you scale or cut.
RGM Expert Says
Incrementality is the question every other metric dances around: of the conversions you reported, how many would have happened without the spend? Attribution cannot answer it — it only redistributes credit among touchpoints it can see. A holdout test answers it directly by creating a counterfactual: a matched group that did not see the campaign. The difference between the two groups is the only number that tells you whether the money did anything. We treat it as the closest thing to ground truth in media measurement.
The result that lands hardest with clients is the gap between last-click ROAS and incremental ROAS. A retargeting campaign showing a glorious 8:1 last-click ROAS routinely collapses to 1.5:1 or worse on a holdout, because most of those buyers were already going to convert — the ads just got in front of an existing purchase. That is not a reason to panic; it is the reason to test. Once you know the incremental number, you can reallocate from spend that merely takes credit to spend that actually creates demand.
Two cautions we never skip. First, match your groups and your windows: the control has to be genuinely comparable, and the measurement window has to be long enough to catch delayed conversions. Second, respect statistical significance. A 3% lift on two groups of a few thousand each can easily be noise; size the test to detect the effect you care about. Google and Meta’s conversion-lift products run this design natively, and geo holdouts give you a vendor-independent version. The math here is simple; the rigor is in the test design.
How it works
Incrementality compares the conversion rate of a treated group against a matched, held-out control. The rate difference is the lift; applying the control rate to the treatment audience reveals how many conversions were genuinely incremental.
- Treatment rate — conversions ÷ group size in the exposed group.
- Control rate — conversions ÷ group size in the held-out group.
- Incremental conversions — treatment conversions beyond what the control rate predicts.
- Incremental ROAS — incremental revenue ÷ spend; the honest efficiency figure.
Lift requires a valid controlled test (geo or audience holdout). Platform tools (Google conversion lift, Meta conversion lift) automate this. See RGM’s guide to incrementality test design; always check statistical significance.
Why incrementality beats attribution
Attribution and incrementality answer different questions, and confusing them costs real money. Attribution asks which touchpoint to credit for a conversion; incrementality asks whether the conversion would have happened at all without the spend. The first is a bookkeeping exercise among touchpoints you can observe; the second is a causal question that only a controlled experiment can settle. When the two disagree — and on retargeting and branded search they almost always do — incrementality is the one that reflects reality.
The practical payoff is reallocation. A holdout test routinely reveals that a channel with a spectacular last-click ROAS is largely harvesting demand that already existed, while an upper-funnel channel with a mediocre last-click number is quietly creating it. Acting on incremental ROAS instead of attributed ROAS shifts budget from credit-takers to demand-creators — the single highest-leverage move many media plans can make, and one you can only justify with a test.
Incrementality is also the antidote to signal loss. As cookies and identifiers fade, click-path attribution gets noisier and less trustworthy, but a geo holdout does not depend on tracking individuals at all — it compares whole markets. That makes incrementality testing more durable, not less, as privacy changes erode the data attribution relies on. Pair it with marketing-mix modeling for the strategic view, and use this calculator for the read-out of any individual test.
Attribution vs incrementality
They are not interchangeable. This is when each is the right tool and why incrementality is the causal answer.
| Question | Best method | What it gives you |
|---|---|---|
| Which touchpoint to credit? | Attribution | A bookkeeping split of observed conversions |
| Did this spend cause sales? | Incrementality (holdout) | The causal lift over a control |
| How do channels combine? | Marketing-mix modeling | Channel contributions at the top level |
| Is the lift real or noise? | Significance testing | Confidence the effect is not chance |
What measurement experts say
Conversion lift compares people who were eligible to see your ads with a randomized control who were not, isolating the conversions your ads actually drove.
Attribution tells you who got the click; only an incrementality test tells you which spend actually caused the sale.