Performance Max is built to be a black box. Google spreads one budget across Search, Display, YouTube, Gmail and Discover, picks the placements, mixes the creative, and shows you a headline number. That’s the pitch — and it’s also why most PMax waste is invisible: there are no search terms to read, no placements to exclude by hand, no keyword bids to fix. The usual audit has nothing to grab.
But there is one lever the format does not hide. A PMax campaign is made of asset groups — bundles of headlines, images and a theme — and each asset group reports its own cost and its own conversions. When an asset group spends real money and returns zero conversions while a sibling group in the same campaign converts fine, that’s not the algorithm learning. That is budget going to a bundle of creative the auction can’t place profitably — and because it’s pooled under one campaign, the healthy group’s results paper over it.
How to see it — three clicks
Google doesn’t surface this by default, but the data is one column away:
Open the Performance Max campaign → click Asset groupsin the left menu → make sure the Conversions and Cost columns are shown, and set the date range to the last 30 days. Now read it as a table: any asset group with meaningful cost and a conversions figure of 0 (or a cost-per- conversion many times the campaign average) is the one draining the budget. Sort by cost, descending, and the worst offender is at the top.
Does every asset group with real cost also have conversions?
Campaign → Asset groups → show Cost + Conversions, last 30 days
Zero conversions on an asset group that has spent real money is the leak. One healthy sibling group hides it at the campaign level.
How does each group's cost-per-conversion compare to the campaign's?
A group running at several times the campaign average is buying the same outcome at a premium — a softer version of the same problem.
How long has the dead group been live?
Under two weeks, it may still be in learning. Past that, spending with nothing to show is a decision you’re making by not making one.
Is it the only asset group, or one of several?
If it’s the only group, the fix is better assets and a clearer audience signal, not pausing. If there are others carrying the campaign, pausing the dead one just moves its budget to them.
0 of 4 checked
What to do about it
If the dead asset group is one of several, the cleanest move is to pause it. That doesn’t cut your budget — it hands that spend to the asset groups that are actually converting, which is where the campaign wanted it anyway. If it’s your only asset group, pausing kills the campaign, so the work is instead to strengthen it: more and better headlines and images (thin creative is the usual cause), and a sharper audience signal so Google has a warmer starting point than “everyone.”
Either way, the discipline is the same one that works everywhere in paid ads: put the numbers on a time axis, find the line that spends and never returns, and act on it before the month closes.
This is one of the checks Admiral runs on a connected Performance Max account — it prices each asset group from your own history, flags one that spends with nothing to show, and offers to pause it in one click, recorded and undoable. It’s the same “find the leak, approve the fix” audit that runs on Search, Demand Gen, Display, Video, Shopping and Meta, pointed at the one lever Performance Max leaves visible.
Run the free audit