How to audit a Performance Max campaign without the asset group theater.

The audits that miss the point.

Every week I look at a Performance Max campaign someone else set up, and every week the audit document in front of me covers the same five things. Asset group composition. Audience signals. Creative variety. Conversion goal selection. Maybe a note about ad extensions.

None of those are the actual problem.

That's not to say they don't matter. They do. But they're the levers the platform UI lets you touch, and the UI is designed to make the campaign feel optimizable. A prospect can move sliders, add headlines, reshuffle audiences, and feel like progress is happening. In most accounts I've audited, those levers explain maybe 10% of the variance in results. The other 90% lives outside the campaign.

Here's what actually determines whether a Performance Max campaign works.

One: whether PMax is the right campaign type at all.

Performance Max works best when:

• The account has a clean, high-quality product feed (for Shopping)
• The advertiser has strong first-party conversion data
• The goal is scale on proven demand
• The budget is large enough for the algorithm to learn (at least a few hundred dollars a week)

Performance Max fails badly when:

• The goal is demand generation with thin conversion data
• The conversion signal is weak (form views counted as conversions, soft goals, etc.)
• The budget is too small for the algorithm to learn
• Brand search is running through PMax instead of a dedicated Brand campaign

The first thing I check when auditing a PMax campaign isn't the asset group. It's whether PMax should even exist in this account. For a lead-gen business with 15 form submissions a month, a cost-per-acquisition target of $180, and a $3,000 weekly budget, Performance Max is probably not the answer. It's going to burn through budget trying to learn from too little signal.

If PMax is the wrong campaign type, asset groups don't matter. Kill it and restructure.

Two: what the campaign is actually counting as a conversion.

The second thing I look at isn't in the campaign settings. It's in the Google Ads conversion actions panel, and in Tag Manager.

Performance Max optimizes aggressively toward whatever you tell it to. If you told it that "form view" is a conversion, it will find people who view forms. Those people are mostly not buyers. They're people with time to browse.

I've seen PMax accounts with $40 CPAs that were completely unprofitable, because the conversion being optimized was "pageview of contact page." The algorithm did exactly what it was asked. The problem was upstream.

Before auditing anything about how PMax is spending, audit what it's spending toward. Log into Google Ads, check the conversion actions counted as primary, and ask:

• Is this conversion action valuable to the business?
• Is there a stronger signal available (SQLs from the CRM, actual purchases from the cart)?
• Is the conversion being double-counted across actions?

The answer is almost always "we could be optimizing for something better." Enhanced conversions and offline conversion imports are usually the fix, not new asset groups.

Three: whether brand search is inside or outside.

This is the single most common PMax mistake I see. Brand search traffic, people searching "your company name" or "your product login," is some of the highest-converting traffic in the account. Historically it ran through a dedicated Brand campaign at a low cost per click.

If you let Performance Max capture that brand search, two things happen. Your PMax campaign looks great, because it's eating the cheapest, highest-converting clicks in the account. And your cost per click on brand terms goes up, often by 2-3x, because PMax bids more aggressively than a Brand campaign would.

The fix is a brand exclusion list applied to the PMax campaign. This used to require contacting your Google rep; as of this year it's available in the UI under campaign-level exclusions. If your PMax campaign doesn't have a brand exclusion list and there's no separate Brand campaign keeping that traffic out, the campaign's reported ROAS is inflated.

Four: whether it's cannibalizing other campaigns.

Related but distinct problem. Performance Max competes with your other search, shopping, and display campaigns for the same impressions. It usually wins, because it bids more aggressively and has fewer constraints.

What looks like "PMax performance" is often "PMax taking credit for conversions that would have gone to the Brand campaign, the Shopping campaign, or the retargeting campaign." The incremental lift is smaller than the reported lift.

The test: pause PMax for a week and watch the rest of the account. If Search Brand campaign spend rises by an amount roughly equal to the PMax spend that disappeared, and total conversions are flat, PMax wasn't adding much. It was just shifting attribution.

Running this test is uncomfortable. It requires the account to hold still for a week with a chunk of budget turned off. Most agencies won't suggest it because they're afraid of the week of bad numbers. That reluctance is exactly why it's worth doing.

Five: the asset group, last.

By the time you've checked the above, the asset group optimization conversation is mostly cosmetic. If PMax is the right campaign type, the conversions are good ones, brand is excluded, and the campaign isn't just stealing credit, then yes, you can tweak asset groups. Consolidate if you have too many; more than three or four per campaign is usually too fragmented for the algorithm. Supply more creative variety. Provide audience signals as warm starts, not targeting constraints. Use asset group labels to separate performance by product category.

But these changes might produce a 5-15% improvement in efficiency. The upstream fixes, if needed, produce 30-50% improvements or more. Start there.

What to do when the audit points upstream.

The awkward part of auditing Performance Max is that the findings often aren't about the campaign. They're about conversion tracking, attribution strategy, or the decision to use PMax in the first place.

When that's the conclusion, the conversation has to go back to the business owner, not stay inside the agency. "Your conversion setup is counting the wrong thing" is not an optimization. It's a strategic reset. It usually requires the agency and the business to agree on what a real conversion is, update tracking accordingly, and then relaunch or restructure.

A written audit that documents these findings, with an honest prioritization, is usually worth more than six weeks of asset group rotation. For the business, it's the foundation for every decision after. For the agency, it's a clean reason to recommend work beyond the optimization task. For everyone, it beats pretending the black box is transparent just because Google drew a slider on the screen.