Treating PMax as an incremental channel means measuring the conversions it genuinely added — not the ones it claimed but would have happened anyway through brand, organic, or other channels. Because PMax harvests cheap conversions across all of Google, its reported numbers overstate its true contribution; incrementality testing (holdouts, geo experiments) reveals what it actually drove.
Performance Max almost always looks like your best channel, and that should make you suspicious, not happy. PMax reaches across all of Google and is exceptionally good at finding cheap conversions — including the ones that were going to happen regardless, through brand searches, organic visits, and demand other channels created. It claims them, reports a gorgeous ROAS, and earns more budget on the strength of conversions it didn’t actually cause. To understand the underlying data infrastructure, review our guide on server-side tagging.
The only question worth answering is incrementality: of everything PMax reported, how much would not have happened without it? That number is usually far smaller than the dashboard — and it’s the only one you should be budgeting against.
Claimed vs. incremental conversions
A conversion PMax reports and a conversion PMax caused are different things. The gap between them is where budget gets wasted on a channel taking credit for work it didn’t do. For more on query control, see our approach to negative keywords.
| Claimed (reported) | Incremental (true) | |
|---|---|---|
| Includes brand harvest | Yes | No |
| Includes existing demand | Yes | No |
| Reflects true contribution | No | Yes |
| Right basis for budget | No | Yes |
Why PMax over-claims
PMax’s reach is the root of the problem. Given access to Search, Shopping, brand queries, and remarketing, it naturally scoops up the cheapest, highest-converting traffic — which is exactly the traffic that converts with or without an ad. Unchecked, it inflates its own numbers by harvesting demand other parts of your marketing already created, then presents that as performance.
What inflates PMax’s reported numbers
Relative share of non-incremental claims.
How to measure incrementality
You measure what PMax added by withholding it and watching what changes. A geo holdout (PMax on in some regions, off in comparable others) or a time-based holdout shows how many conversions actually disappear without it — that’s the incremental number. Excluding brand from PMax first removes the most obvious inflation. The goal is a defensible read on contribution, not the platform’s self-report.
So should I stop trusting PMax’s reports?
Don’t stop using PMax — stop taking its self-reported ROAS at face value. It’s a powerful channel for finding genuinely new demand once you’ve measured and subtracted what it would have claimed anyway. Trust the incrementality test, not the dashboard.
Every channel wants credit for the conversion, and PMax is structurally positioned to grab more than its share. Measuring incrementality is how you cut through the self-congratulation and fund PMax for what it actually adds — which is the only honest way to size any channel.
This article is a spoke node connected to our core technical hubs. To explore the broader architecture, visit our primary pillar pages: