Incrementality Calculator
Incrementality is the revenue that exists only because you advertised. It is the gap between what an exposed audience did and what a matched holdout would have done anyway. That distinction matters because platform-reported ROAS credits every conversion a channel can claim, including sales that would have happened with no ad at all. This free incrementality calculator turns a holdout result into incremental lift, incremental revenue, and iROAS in one step, and it requires no email.
Enter your numbers to see results.
Attribution answers which touchpoint to thank. Incrementality answers whether the spend changed anything. A retargeting campaign can post an 8x platform ROAS while delivering almost no incremental revenue, because it is harvesting buyers who were already on their way to checkout. View-through conversions, branded search, and last-click models all push reported numbers above the honest one. iROAS is the honest number: incremental revenue divided by spend.
Enter the raw revenue or conversions from your exposed and holdout groups along with the size of each group, plus your spend. The calculator scales the holdout to the exposed group's audience size for you, so a 10% holdout compares fairly against the other 90%. Add your platform-reported ROAS and it also shows the attribution haircut, the share of claimed revenue that was never really incremental.
How it works
First the holdout is scaled: multiply the holdout result by the exposed group size divided by the holdout group size, so both groups represent the same number of people. Incremental lift then equals the exposed group result minus the scaled holdout, divided by the scaled holdout, expressed as a percentage. Incremental revenue is the exposed group minus the scaled holdout. iROAS, the incremental return on ad spend, equals incremental revenue divided by spend.
When you also enter platform-reported ROAS, the attribution haircut equals one minus iROAS divided by platform ROAS, as a percentage. It quantifies how much of the platform's claimed return would have happened without the ads. Negative lift is allowed: if the exposed group did not beat the holdout, the spend showed no measurable incrementality.
One honest caveat: these are point estimates. A lift measured on a small or short holdout can be statistical noise in either direction, so size the holdout and the test window with the incrementality test calculator before treating any single readout as proof.
Worked example
Suppose you exposed 1,000,000 people to ads and held out 100,000. The exposed group drove $120,000 in revenue and the holdout drove $8,000, on $25,000 of spend. The holdout scales by 1,000,000 divided by 100,000, so its $8,000 becomes $80,000 at the exposed group's size. Incremental revenue is 120,000 minus 80,000, or $40,000. Lift is 40,000 divided by 80,000, or 50%. iROAS is 40,000 divided by 25,000, or 1.6x.
Now suppose the platform reported a 4.0x ROAS on that same spend. The haircut is 1 minus 1.6 divided by 4, or 60%. In plain terms, 60% of the revenue the platform took credit for would likely have happened anyway, and your true incremental return is 1.6x, not 4x.
What is a good iROAS?
A good iROAS depends on margin and goal, exactly like ordinary ROAS, but the bar is set against your break-even rather than against the inflated platform number. A DTC brand whose blended ROAS sits in the healthy 2.5 to 4x range will often find prospecting iROAS comes in well below that, frequently 1.0 to 2.0x, because incrementality strips out the warm-audience and retargeting revenue that padded the blended figure. Enterprise B2B, which can run profitably at 0.5 to 1.5x blended ROAS, judges incremental campaigns against that same lower bar.
The number that should alarm you is a large attribution haircut. A 50 to 70% haircut on a retargeting or branded-search line is common and tells you most of that reported return was harvesting existing intent. Prospecting and upper-funnel campaigns usually show the highest incremental lift and the smallest haircut, because they reach people who were not already converting. That is the opposite of what last-click attribution rewards, which is why so many budgets are misallocated.
Treat iROAS as the figure you scale on and platform ROAS as the figure you sanity-check. If a channel posts a high platform ROAS but a thin iROAS, the honest move is to shift budget toward the campaigns that actually create demand. Compare your incremental returns and your blended efficiency to QRY's monthly paid media benchmarks to judge whether your true return is healthy for your vertical and margin profile.
See QRY's monthly paid media benchmarks to compare your numbers against the portfolio.
Frequently asked questions
What is a good iROAS?
A good iROAS is judged against your break-even, not against your inflated platform ROAS. DTC prospecting iROAS often lands at 1.0 to 2.0x even when blended ROAS is 2.5 to 4x, because incrementality removes the warm-audience revenue. Enterprise B2B can be profitable on incremental campaigns at 0.5 to 1.5x. The key signal is whether incremental return clears your margin requirement.
How do you calculate incrementality?
Run a holdout: keep a matched group from seeing your ads, then compare results. Scale the holdout to the exposed group's audience size, since holdouts are usually a small slice. Incremental revenue equals the exposed group minus the scaled holdout, lift is that difference divided by the scaled holdout, and iROAS is incremental revenue divided by spend. For example, $120,000 exposed versus a holdout that scales to $80,000, on $25,000 spend, is $40,000 incremental, 50% lift, and a 1.6x iROAS.
Why is my iROAS lower than my platform ROAS?
Platform ROAS credits every conversion a channel can claim, including view-through conversions and sales from people who would have bought anyway. Retargeting and branded search are the worst offenders because they harvest existing intent. iROAS only counts the revenue that exists because you advertised, so it is almost always lower and is the more honest number to scale on.
What is an attribution haircut?
The attribution haircut is the share of platform-reported revenue that was not actually incremental. It equals one minus iROAS divided by platform ROAS. If your platform claims 4x but your iROAS is 1.6x, the haircut is 60%, meaning 60% of the claimed revenue would likely have happened without the ads. Large haircuts usually appear on retargeting and branded-search lines.
What is the difference between incrementality and attribution?
Attribution decides which touchpoint gets credit for a conversion that already happened. Incrementality asks whether the conversion would have happened at all without the ad. A campaign can win on attribution while adding almost no incremental revenue, which is why incrementality is the better basis for budget decisions.
How big should the holdout group be?
Large enough to detect the lift you care about with statistical confidence, and matched to the exposed group on the variables that drive conversion. A 10 to 20% holdout is a common starting point. You do not need to pre-scale anything here: enter the raw results and both group sizes, and the calculator normalizes the comparison. Use an incrementality test calculator to size the holdout and the test duration before you launch.
Related calculators
Is your platform ROAS telling the truth?
See how the QRY portfolio separates incremental return from attribution inflation with our monthly paid media benchmarks.
See the benchmarks