Performance Max is one campaign spanning Search, Shopping, YouTube, Display and Gmail, with very little visibility into where the money went. That opacity is why it gets blamed for everything — and why the real problems hide so well inside it.
Problem one: it bids on your own brand
PMax is allowed to serve against searches for your brand name. Those convert at an enormous rate, because the person was already coming to you. Their revenue lands in the PMax report, the ROAS looks spectacular, the campaign earns more budget — which it spends on more of the traffic you already owned.
Total revenue stays flat. The report stays glowing. Nobody can explain the gap.
Not "what is PMax's ROAS". It is "what would have happened anyway". Carve brand into its own Search campaign so you can see it separately, then judge PMax on what is left.
Problem two: the feed is doing the targeting
For the shopping half of PMax, your product data decides which searches you are eligible for. Supplier titles, missing GTINs and vague product types mean you show up for the wrong things — and PMax will happily spend on them, because it was told to find conversions, not to be sensible about which.
Before blaming the campaign, check how many of your products Google has actually served, and how many sold anything in the last 30 days. Those are almost never the same number. More on feeds.
Problem three: it inherits every tracking error you have
Automated bidding with bad conversion data is not automation, it is amplification. If add-to-cart is a primary conversion, PMax will find you a great deal of add-to-cart.
The three fixes, in order
- Separate brand. A dedicated brand Search campaign, and exclude brand from PMax where the account allows it.
- Split asset groups by product economics rather than one group holding the whole catalogue, so your best-margin products are not subsidising dead weight.
- Clean the conversion action — purchase only, with real revenue attached — before you touch a target.
When PMax genuinely is the right answer
A healthy feed, enough conversion volume for the bidding to learn, real creative assets, and brand already carved out. In that setup it reaches surfaces you cannot easily buy otherwise, and it is often excellent.
What it is not is a substitute for account structure, and it is a poor first campaign for a store that has never run Google. Our full PMax page has the readiness checklist.