Article
Jun 10, 2026
Performance Max Audit Checklist: 21 Checks for the Post-Black-Box Era
Google finally shows you where PMax spends your money. Here's the 21-check audit built on that data, with red-flag thresholds and fixes
Most Performance Max audit checklists you'll find on page one of Google were written between 2022 and early 2025, when PMax was a black box. They audit account structure, bidding strategy, and asset coverage because that's all the API would surface. They don't audit channel splits or full search terms, because those reports didn't exist yet.
They do now. A real performance max audit checklist in June 2026 starts from the transparency data Google shipped over the last 14 months, not the proxies we used before it.
TL;DR
Google now reports PMax spend, clicks, and conversions split across Search, YouTube, Display, Discover, Gmail, and Maps.
Full search-terms reporting means performance max wasted spend and brand cannibalization are now measurable, not guessed.
A healthy 2026 PMax account should track within range of LocaliQ's $5.42 CPC and $66.69 CPL cross-industry baselines.
Red flags that mean rebuild, not tweak: >40% Display/Gmail spend on a lead-gen account, >60% conversions from brand queries, asset group ROAS spread above 3x.
A proper audit is 15–25 hours of analyst work. Anything under 4 hours is a template scan, not an audit.
1. Why pre-2025 PMax audit advice is obsolete
For most of PMax's life, you bought a campaign type and got a delivery report that told you almost nothing about where the money went. Old audits compensated with inference: signal quality, asset coverage scores, conversion volume by hour, account-level trends. Useful, but circumstantial.
That changed when Google rolled out channel-level performance reporting for Performance Max, surfacing impressions, clicks, cost, and conversions across Search, YouTube, Display, Discover, Gmail, and Maps, alongside full search-terms reporting. At Google Marketing Live 2026 on May 20, the company added Ask Advisor spanning Ads, Analytics, and Merchant Center, plus more AI campaign tooling.
The transparency concessions are real. The surface area an auditor has to supervise also got bigger. Both things are true.
The practical effect: if your current audit doesn't open the pmax channel performance report and the full search terms view in the first 20 minutes, it's auditing the wrong thing.
2. Channel-level checks: where the budget actually goes
The first job of a 2026 audit is to find out what kind of campaign you're really running. A PMax campaign labeled "Search-heavy" by your previous agency might be spending 55% of its budget on Display and Gmail placements that convert at a fraction of Search's rate.
In our client work, the channel split we see most often on lead-gen accounts that came to us for an audit looks something like this: 28% Search, 41% Display, 14% YouTube, 11% Gmail, 6% Discover and Maps combined. The Display and Gmail share is doing the volume work that keeps the CPA looking acceptable on cheap, low-intent impressions.
3. Search terms: brand cannibalization vs. genuinely new demand
The full search-terms report is the single biggest change to how a PMax audit gets done. Before it existed, performance max brand cannibalization was a theoretical worry that smart practitioners flagged but couldn't quantify. Now you can.
The math is simple. Export the search terms report. Tag every query as branded, competitor, generic, or long-tail. Sum cost and conversions by bucket. If 60%+ of your PMax conversions come from queries containing your brand name, and you also run brand Search campaigns, PMax is taking credit for traffic your brand campaign would have closed at one-third the CPC.
4. Asset and creative checks
Asset-level reporting now gives you conversion data per asset group, not just "Best / Good / Low" labels. That's a real upgrade.
What to look for:
Check 11. Asset group ROAS spread. If your best group does 4.2x and your worst does 1.1x at similar spend, the worst is dragging the smart bidder's learning. Pause or rebuild.
Check 12. Headline diversity per group. Three headlines that are minor rewrites of each other don't give the system room to test.
Check 13. Video assets present in every asset group. Without them, Google auto-generates video, which is usually worse than a 6-second cutdown of something you already shot.
Check 14. Final URL expansion setting. On for ecommerce with deep catalogs; off for lead-gen with a single high-intent landing page.
5. Feed, audience signal, and landing page checks
These are the inputs the smart bidder treats as ground truth. Bad inputs, bad outputs.
Check 15. Merchant Center feed health, disapprovals, missing GTINs, stale prices. Anything above 3% disapproved is suppressing eligible impressions.
Check 16. Audience signals reflect your current customer, not the customer you had 18 months ago. Refresh customer match lists quarterly.
Check 17. Conversion actions in the account. More than one primary conversion action with overlapping definitions causes double-counting and bad bid signals.
Check 18. Enhanced conversions for leads or for web, configured and validating. The diagnostics tab tells you in one click.
Check 19. Landing page Core Web Vitals on the URLs PMax actually sends traffic to (not just your homepage). Slow pages tank the system's CVR estimate and shift spend to channels that don't need a fast page.
6. Benchmarks: what healthy looks like in 2026
Use external baselines as a sanity check, not a target. LocaliQ's 2026 search advertising benchmarks put the cross-industry average at $5.42 CPC, $66.69 cost per lead, and an 8.18% conversion rate. PMax usually runs hot on impressions and cool on CPC versus pure Search, so a healthy PMax CPC often sits below that $5.42 average while CPL tracks within roughly 20% of the $66.69 figure for comparable industries.
If your account is 3x off the LocaliQ baseline in either direction, something structural is wrong, not something a 5% budget shift will fix.
A note on incrementality: a low CPL doesn't mean PMax caused the conversions. For accounts under $20k/month, the incrementality question matters more than the CPL question. We wrote about that math in this piece on Google Ads incrementality on a small budget.
7. Red flags that mean rebuild, not tweak
Some findings can be fixed with a negative keyword list and a signal refresh. Others mean the campaign was set up for the wrong objective and no amount of optimization will fix it.
Check 20. Brand conversion share above 60% with an active brand Search campaign. Rebuild with brand exclusions and re-baseline.
Check 21. Asset group ROAS spread above 3x sustained over 60 days. The campaign is averaging across two different businesses.
Whether that's worth in-housing or sending out depends on the rest of your media stack, we ran the numbers in PPC agency vs. in-house: the cost math.
FAQ
Q: How often should I run a Performance Max audit checklist?
A full 21-check audit makes sense quarterly for accounts spending over $10k/month, or after any major Google product change like the May 2026 Ask Advisor launch. Lighter monthly checks on channel split, search terms, and asset group ROAS are enough between full audits to catch drift early.
Q: What's the fastest way to detect performance max brand cannibalization?
Export the full search terms report for the last 90 days, tag every query containing your brand name, and sum conversions. If branded queries account for more than 50% of PMax conversions while you also run a brand Search campaign, you have cannibalization. The cost-per-conversion gap between the two campaigns tells you what it's costing you.
Q: Can I use the pmax channel performance report to exclude bad channels?
Not directly. Google added reporting transparency without adding channel-level exclusions for most advertisers. What you can do: shift budget toward campaigns and asset groups that index toward your preferred channels, tighten audience signals, and use account-level placement exclusions. Structural channel control still requires splitting objectives across campaigns.
Q: What counts as performance max wasted spend versus learning budget?
Learning budget is spend on queries or placements during the first 2–3 weeks of a campaign or after a major change. Wasted spend is spend on zero-conversion queries after 60+ days, on channels whose cost share exceeds conversion share by 15+ points, or on duplicate brand queries already covered by Search. The reports now let you separate them.
Q: Are the 2026 LocaliQ benchmarks reliable targets for PMax specifically?
They're reliable as sanity checks, not targets. LocaliQ's $5.42 CPC, $66.69 CPL, and 8.18% conversion rate are cross-industry search advertising averages. PMax blends Search with cheaper inventory, so CPC usually runs lower and CVR varies by channel mix. Use them to flag accounts that are 2–3x off baseline, not to set quarterly KPIs.
This week, pull the channel performance report and the full search terms export for your top-spending PMax campaign, tag the queries, and calculate your brand conversion share. That's the single highest-impact hour you'll spend on the account this quarter.
If the numbers come back ugly and you want a second set of eyes on what to do next, get in touch.
How should a small team prioritize ai max campaign type google ads?
Start with the workflow that already has a baseline: hours, leads, errors, or budget waste.
What should be measured before investing in ai max campaign type google ads?
Measure cycle time, volume, handoffs, error rate, and the current owner.
When should performance max audit stay manual instead of automated?
Keep it manual when judgment, approval, brand nuance, or customer trust is on the line.
How does google ads ai max campaign type change the budget for ai max campaign type google ads?
google ads ai max campaign type usually adds integration, QA, and monitoring work.
What is the first project to launch from this performance max audit playbook?
Launch the narrowest workflow with a visible result.
Related reading
AI Email Personalization: What Lifts Revenue and What's Just Merge Tags
AI for a Landscaping Business: Five Automations for About $345 a Month