Performance Max gets blamed for wasting budget more than any other Google Ads campaign type — and most of the time the blame is fair, just aimed at the wrong cause. PMax isn’t inherently inefficient. It’s a black box that rewards a clean feed and starves everything else. Hand it a messy catalog and no guardrails, and it will happily spend your budget rebuying customers who already searched your brand name. Structure it properly and it becomes one of the most efficient channels in the account.
Why PMax drains budget faster than Search or Shopping ever did
Search and Shopping campaigns fail loudly — a bad keyword or a disapproved product is easy to spot. PMax fails quietly. Google’s automation decides which inventory, which placements, and which audience signals to chase, and the reporting it hands back is thin by design. An account can burn through 30–40% of a day’s budget on branded search terms it would have won for free in a normal Search campaign, and nothing in the standard PMax dashboard flags it.
The fix isn’t turning PMax off. It’s removing the reasons it has to guess.
The feed is still doing most of the work
PMax leans harder on your Merchant Center feed than any other campaign type, because Shopping assets are usually where most of its budget lands. Thin titles, missing GTINs, wrong categories, and stale pricing don’t just cause disapprovals — they push PMax toward whichever products happen to have the cleanest data, regardless of margin or actual demand. If your feed has disapprovals right now, that’s the first thing to fix before touching a single setting inside the campaign; see our full breakdown of Merchant Center disapprovals and fixes for the exact checklist.
Asset groups: the lever most accounts never touch
Most PMax accounts run a single asset group for the whole catalog, which forces Google to average performance across products that have nothing in common. Splitting asset groups by product category, margin tier, or price point lets the algorithm learn a cleaner signal for each one and lets you set realistic ROAS targets per group instead of one blended number that under-serves your best sellers and over-serves your worst.
Guardrails that actually move the needle
- Brand exclusions — a dedicated brand Search campaign with a negative-brand list applied to PMax stops it from rebuying people who already typed your name.
- Account-level negative keywords — broad PMax now supports a limited negative keyword list; use it for irrelevant generic terms burning spend with no purchase intent.
- Audience signals, not audience targeting — feed PMax your best customer lists and site visitors as a signal to learn from, not a restriction; it still finds new buyers outside the list.
- Placement exclusions — cut low-quality display and video inventory in the Display Network and YouTube placements report once there’s enough data to see where it’s wasted.
- A real ROAS floor — set target ROAS per asset group based on actual margin, not a single account-wide number copied from last quarter.
When to split PMax out of a single catalog-wide campaign
If one hero SKU or category drives a disproportionate share of revenue, give it its own PMax campaign with its own budget and ROAS target instead of letting it compete inside a catalog-wide one. The same applies to new launches: a brand-new product with no conversion history will almost always lose budget priority to established sellers inside a shared campaign, which is the opposite of what you want while you’re trying to get it off the ground.
Reporting is the other piece most accounts get wrong. The default PMax dashboard blends everything into one ROAS number, which hides exactly the problems above — a brand-term leak, an underperforming asset group, or a category quietly losing money inside a profitable-looking total. Pulling search term insights, the channel breakdown by asset group, and the placement report weekly, rather than glancing at the headline ROAS, is what actually catches waste before it compounds over a month of spend.
None of this is a one-time setup. PMax keeps re-learning as the feed, seasonality, and competition shift, which is exactly why it needs the same continuous audit treatment as the rest of a Google Ads account — not a quarterly check-in. See how we structure Search, Shopping, and Performance Max with real guardrails for e-commerce brands.