Performance Max: Why Your Campaigns Burn Budget and How to Fix It

Google pushes Performance Max hard as “the campaign of the future.” In practice it is often a black box that spends budget opaquely and cannibalizes your other campaigns.

What PMax does under the hood

Performance Max places ads across every Google inventory at once: Search, Shopping, Display, YouTube, Discover, Gmail, Maps. The algorithm decides where the budget goes on its own. In practice PMax favors brand traffic and retargeting — they convert better, and Google optimizes for the conversion metric rather than for acquiring new users.

Why PMax is often overrated

Campaign analysis shows that 60–80% of PMax conversions come from Search placements, and a sizeable share of those are brand queries you would have won through regular campaigns anyway. CPA on brand traffic is 3–5 times lower than on non-brand, which distorts the overall picture of performance. Without a breakdown by segment you cannot see where the conversions really come from.

How to structure PMax properly

Split campaigns by goal: one PMax for new customers, a separate one for retargeting. Turn off URL expansion — otherwise Google will send traffic to whatever pages on your site it sees fit. Configure listing groups manually, with priorities based on margin. And always add brand exclusions through Customer Match or a dedicated list — otherwise PMax will intercept brand traffic from your regular campaigns.

When PMax works well

PMax delivers real performance once you have at least 300+ conversions a month at the account level — only then does the algorithm get enough data to learn from. The second hard requirement is a quality product feed for eCommerce: without it the system cannot optimize Shopping placements. If both conditions hold and you have excluded brand properly, PMax genuinely can extend your reach at an acceptable cost.

Metrics to keep it in check

The key metrics for monitoring PMax: new customer rate (it should be above 40%, otherwise the campaign is only doing retargeting), Search Impression Share on non-brand queries, and assisted conversions in the attribution report. The best way to check PMax's real value is a pause test of 2–4 weeks: pause the campaign and see whether sales actually drop or PMax was simply cannibalizing your other channels.