The Definition
Platform ROAS is attributed revenue divided by ad spend, as reported by the ad platform. Incremental ROAS is the revenue that would not have existed without the spend, divided by the same spend. The first measures what your ads touched. The second measures what your ads caused.
They are routinely treated as the same number. They are not. The difference is often large enough to change budget decisions. This is the short version of the distinction; the full testing method lives in our guide to whether your paid media spend is truly incremental.
Why the Platform Number Is Usually Higher
Ad platforms report attributed conversions: sales where the platform can show its ad was seen or clicked somewhere on the path. Attribution records contact, not cause. Some of the people your ads touch were already coming:
- Brand searchers who had already chosen you and clicked the ad sitting above your own organic listing.
- Retargeted visitors returning for a basket they already intended to buy.
- Repeat customers inside a remarketing audience who buy on their own schedule.
Every one of these is a real sale and a real attributed conversion. Only a fraction are caused sales. Platform ROAS counts all of them; incremental ROAS counts only the fraction. That is why platform ROAS is usually the ceiling of your ads' contribution, not a complete measurement of it. Because brand and retargeting segments contain more already-coming buyers, they are often where the two numbers diverge most. The same caveat applies to industry benchmark tables, including our own ROAS benchmarks by industry: they describe platform ROAS, so treat them as context rather than targets.
A Worked Example
The following account is invented, and the incrementality assumptions in the table are illustrative only. They exist to show the arithmetic, not to describe any typical account. Your own rates can only come from testing.
A business spends £10,000 a month. The platform reports £50,000 in attributed revenue, a 5.0x platform ROAS. The spend splits three ways:
| Segment | Spend | Attributed revenue | Illustrative assumption (invented) | Incremental revenue |
|---|---|---|---|---|
| Brand search | £2,000 | £25,000 | 3 in 10 conversions caused | £7,500 |
| Retargeting | £3,000 | £15,000 | 4 in 10 conversions caused | £6,000 |
| Cold prospecting | £5,000 | £10,000 | 9 in 10 conversions caused | £9,000 |
| Total | £10,000 | £50,000 (5.0x) | £22,500 (2.25x iROAS) |
Again: the assumption column is invented for illustration. But look at what the arithmetic does. The account's headline is 5.0x. Its illustrative incremental return is 2.25x. The ranking of the segments also flips. Brand search reports 12.5x and contributes the least caused revenue per pound. Prospecting reports 2.0x and contributes the most. A budget decision made on platform ROAS moves money towards brand and retargeting. A decision made on iROAS moves it the other way.
That flip, not the absolute numbers, is the point. The segments that look best in platform reporting often test lower than their reports imply. Cold prospecting can be undervalued because attribution loses buyers who convert later or through another route. Finding where your account actually sits is what a paid media efficiency engagement establishes before any budget gets moved.
Which Metric for Which Decision
Neither number is wrong. They answer different questions, and the mistake is using one where the other belongs.
Use platform ROAS for in-channel steering. Compare creative A against creative B, one audience against another, or one bid strategy against another inside the same campaign type. The attribution bias applies roughly equally to both sides, so the comparison can still help even when the absolute numbers are inflated.
Use incremental ROAS for budget-level decisions. This includes whether a channel deserves to exist, whether to scale it, and whether brand search needs its own budget at all. These decisions compare segments with very different incrementality profiles, which is where platform ROAS can mislead. If your automated campaigns blend brand and prospecting into one reported number, start by auditing what Performance Max is actually doing. And if the budget decision on the table is an increase, audit the waste before you scale it.
One practical consequence: a channel with falling platform ROAS can be improving, and a channel with a stellar platform ROAS can be dead weight. You cannot know from the dashboard alone.
How to Get Your iROAS Without a Data Team
You do not need to buy a measurement platform to start. The full method is a ladder, and the first rungs are free:
- Reconcile: put 90 days of platform-claimed revenue next to blended revenue from your finance system. The gap is your first signal.
- Separate: split brand from non-brand and retargeting from prospecting in every report, and see where the headline ROAS actually concentrates.
- Pause-test: switch off one suspect segment for two to four weeks and watch blended revenue rather than platform metrics.
- Experiment: geo splits and holdouts, when your volume justifies the standard of proof.
Most accounts learn something useful by rung two. Try the arithmetic on your own numbers with the ROI calculator, then test the assumptions properly.