Performance Max Campaign Diagnostics: How to Audit What Google Won't Show You

Your client's Performance Max campaign has been running for six weeks. The spend is real. The conversions look thin. The client is asking questions you can't confidently answer because Google won't show you what's actually happening inside the campaign. Welcome to the PMax diagnostic problem — the one every agency running these campaigns eventually hits.

Performance Max is Google's most powerful campaign type and simultaneously its most opaque. You hand Google your assets, your audience signals, your budget, and your conversion goals — and then you wait while an algorithm decides where your client's money goes across Search, Display, YouTube, Gmail, Discover, and Maps. When it works, it looks like magic. When it doesn't, you're left diagnosing a black box with almost no visibility into what went wrong.

This post is the diagnostic framework we use when agencies come to us saying "Performance Max isn't working." It's the same process whether PMax conversions are low, cost per acquisition is climbing, or the campaign just feels like it's burning money on Display placements nobody asked for.

Why Performance Max Is a Black Box (And Why That Matters for Diagnosis)

With Standard Shopping or Search campaigns, you control the levers. You pick the keywords. You set the bids. You choose the placements. When something breaks, you can trace the problem back to a specific decision you made.

PMax flips that model. Google controls targeting, placement, creative assembly, and bidding — all of it. You provide inputs (assets, audiences, goals) and Google's algorithm decides everything else. That means when Performance Max isn't working, you can't just open the campaign and find the broken thing. The "broken thing" might be a targeting decision Google made that you can't see, a placement you'd never approve, or a creative combination that makes no sense.

Here's what Google controls inside PMax that you cannot directly audit:

• Which search queries trigger your ads — you get a partial search terms report, but Google redacts a significant portion under "other search terms."
• Which placements your Display and YouTube ads appear on — the Placements report exists but it's incomplete and often dominated by junk inventory.
• How your creative assets are being combined — Google assembles headlines, descriptions, and images dynamically. You see performance ratings but not the actual combinations being served.
• How your budget is allocated across channels — you cannot control or even reliably see how much goes to Search vs. Display vs. YouTube vs. Gmail vs. Discover.
• Which audience segments are actually converting — your audience signals are "suggestions," not targeting. Google can and will go beyond them.

This is why PMax diagnostics are fundamentally different from diagnosing any other campaign type. You can't just look inside the campaign. You have to triangulate using data from multiple sources — Google Ads, GA4, and Search Console — to build a picture of what's actually happening.

What Limited Data You CAN See — And How to Read It

Google has slowly improved PMax reporting, but it's still far less transparent than Search or Shopping. Here's what you actually have access to and what it tells you:

Asset group performance ratings:

Each asset (headline, description, image, video) gets a rating: Best, Good, Low, or "Learning." These ratings are relative within the asset group, not absolute. A "Best" headline in a weak asset group might still be underperforming. Use these to identify which assets Google is favoring and which it's ignoring — but don't treat them as gospel. An asset rated "Low" might just be undertested, not bad.

Search terms report (partial):

You'll see some of the search queries triggering your PMax ads, but Google hides a large chunk under "other search terms." Look at what you can see. If the visible terms are irrelevant — informational queries, competitor names you didn't target, completely off-topic terms — the hidden ones are likely worse. This is your canary in the coal mine.

Insights tab:

The Insights tab in Google Ads shows search categories, audience segments, and some asset combination data. It's high-level and often delayed, but it's one of the few places you can see which audience segments Google is actually serving your ads to. If you set up audience signals for "in-market for commercial HVAC" and the Insights tab shows most conversions coming from "general interest in home improvement," that's a signal your audience signals aren't doing what you think.

Placement reports:

Check where your Display and Video ads are showing up. If you see a wall of obscure mobile apps and parked domains eating budget, that's a problem. You can add placement exclusions, but it's whack-a-mole — new junk placements appear constantly. The placement report is more useful as a diagnostic signal than a management tool.

The PMax Diagnostic Framework: Asset Groups, Search Themes, and Audience Signals

When we diagnose a Performance Max campaign, we look at three layers in order. Each layer can independently cause the campaign to underperform.

Layer 1: Asset group structure and performance

Asset groups are where most PMax problems start. The most common mistakes we see:

• Too few asset groups: One asset group trying to cover everything. Google needs thematic clarity. If your single asset group contains headlines about "emergency plumbing" and "bathroom remodeling," Google can't optimize for either audience effectively.
• Too many asset groups: Splitting so thin that no single group gets enough data to optimize. If each group is getting 5 clicks a week, the algorithm can't learn.
• Weak asset diversity: Three headlines that say the same thing in different words. Two images that are basically identical. Google needs variety to test combinations. Give it the same thing five ways and you've given it nothing to optimize.
• Missing asset types: No video assets means Google auto-generates videos from your images — and they almost always look terrible. Provide real videos, even simple ones, or Google will create something you'd never approve.

Layer 2: Search themes

Search themes were Google's answer to "agencies want some keyword-level control in PMax." They're not keywords — they're signals that tell Google what types of searches are relevant to your asset group. But they matter more than most people realize.

If your search themes are too broad ("plumbing"), Google will cast a wide net and you'll show up for informational queries that never convert. If they're too narrow ("24 hour emergency gas line repair in downtown Phoenix"), you'll starve the algorithm of volume.

Cross-reference your search themes with the actual search terms report. Are the visible search terms aligned with your themes? Or is Google interpreting your themes loosely and showing ads for tangentially related queries? This mismatch is one of the most common PMax performance killers we diagnose.

Layer 3: Audience signals

Audience signals in PMax are suggestions, not hard targeting. You can add custom segments, in-market audiences, demographics, and customer lists — but Google treats all of these as starting points, not boundaries. The algorithm will go beyond your signals if it thinks it can find conversions elsewhere.

This means you need to diagnose whether your audience signals are actually influencing the campaign or being effectively ignored. Check the Insights tab for audience segment breakdowns. If Google is finding most of its conversions outside your specified signals, you have two possibilities: either Google found a better audience than you suggested (good), or it's chasing cheap conversions from low-quality traffic (bad). The only way to tell is looking at conversion quality downstream.

Key insight: The biggest PMax diagnostic mistake agencies make is looking at each of these layers in isolation. A well-structured asset group with strong search themes will still underperform if the audience signals are pulling in the wrong traffic. All three layers have to work together.

Is PMax Cannibalizing Your Search Campaigns?

This is the question that keeps agency owners up at night. You're running PMax alongside standard Search campaigns, and suddenly Search performance drops. Conversions shift to PMax. CPAs look fine on paper. But something feels off.

Here's the reality: PMax will cannibalize branded search traffic unless you explicitly exclude brand terms. This is by design. PMax prioritizes easy conversions to hit its target CPA, and branded queries are the easiest conversions available. If your PMax campaign is claiming credit for conversions that your branded Search campaign would have captured anyway, your actual incremental ROAS is much lower than what Google Ads reports.

How to check for cannibalization:

• Compare Search campaign metrics before and after PMax launch. Pull impression share, click volume, and conversion data for your Search campaigns from the 90 days before PMax went live vs. the 90 days after. If Search impressions and conversions dropped proportionally to what PMax picked up, that's cannibalization.
• Check PMax search terms for brand queries. If a large percentage of PMax search terms are your client's brand name or close variations, PMax is claiming easy branded traffic as its own conversions.
• Look at total account-level conversions. If PMax added 50 conversions but total account conversions only increased by 10, the other 40 were cannibalized from existing campaigns. This is the simplest and most telling test.
• Use brand exclusions. Google now allows brand exclusions in PMax. Apply them and watch what happens to both PMax and Search performance over the next 2-3 weeks. If PMax performance tanks when you exclude brand terms, it was heavily reliant on branded traffic.

When PMax "Isn't Working" vs. When It's Working But You Can't See It

Not every PMax performance concern is an actual problem. Sometimes the campaign is doing exactly what it should, but the reporting doesn't show it clearly. Before you restructure everything, rule out these scenarios:

PMax might be working if:

• Conversions are happening on a delay. PMax often targets upper-funnel users through Display and YouTube who convert days or weeks later. If you're looking at last 7 days and the conversion window is 30 days, you're missing the picture. Check the conversion lag report in Google Ads.
• Assisted conversions are high. PMax might be driving the first or middle touch in a multi-step conversion path. Go to GA4 > Advertising > Conversion paths and look for PMax as an assisting channel. It might be doing heavy lifting that last-click attribution doesn't capture.
• View-through conversions are excluded from your reporting. PMax runs video and display ads that generate view-through conversions. If your reporting only counts click-through, you're understating PMax's contribution.
• The campaign is still in learning. PMax needs approximately 50 conversions in 30 days per asset group to fully optimize. If volume is below that, the algorithm is still testing and performance will be inconsistent. Patience is warranted — but not infinite patience.

PMax is genuinely not working if:

• Spend is high, conversions are near zero, and it's been more than 4-6 weeks.
• The search terms you can see are wildly irrelevant to the business.
• Placements are dominated by mobile app inventory and parked domains.
• GA4 shows PMax traffic has a significantly higher bounce rate and lower engagement than other campaigns.
• Account-level conversions haven't increased since PMax launched — it's just redistributing credit.

Using GA4 to Get the Visibility Google Ads Won't Give You

GA4 is your best friend for PMax diagnostics because it shows you what happens after the click — something Google Ads reporting barely touches.

What to pull from GA4:

• Traffic quality by campaign. In GA4, go to Acquisition > Traffic Acquisition and filter for your PMax campaign. Compare engagement rate, average engagement time, and events per session against your Search campaigns. If PMax traffic has 20% engagement rate and Search has 65%, PMax is bringing in low-quality traffic regardless of what Google Ads conversion data says.
• Landing page performance for PMax traffic. Use the Pages and Screens report filtered to PMax as the traffic source. Which pages are PMax visitors landing on? Are they the right pages? If your PMax campaign is sending traffic to your homepage instead of service pages, the final URL expansion setting might be overriding your URL inputs.
• Conversion path analysis. Go to Advertising > Conversion paths. Look at where PMax appears in the journey. Is it the first touchpoint? Last? Middle? If PMax is consistently the last touch on paths that started with organic or direct, it's claiming credit for conversions it didn't generate. If it's consistently the first touch, it might be doing legitimate prospecting.
• Audience comparison. Build an exploration in GA4 comparing PMax users vs. Search users. Look at demographics, geography, device type, and on-site behavior. If PMax is pulling in a completely different audience than Search, that's either a feature (new audience discovery) or a bug (wrong audience targeting). Context matters.

Pro tip:Turn off "URL expansion" in your PMax campaign settings if you want to control which pages get traffic. When URL expansion is on, Google will send PMax traffic to any page on the site it thinks will convert — including blog posts, about pages, and other pages that were never meant to be landing pages.

Search Console: Understanding What PMax Is Actually Targeting

Here's a diagnostic move most agencies miss entirely: using Google Search Console alongside Google Ads to understand PMax behavior. Search Console won't show you PMax ad data directly, but it reveals critical context.

How to use Search Console for PMax diagnostics:

• Compare organic query data with PMax search terms. Pull the top queries driving organic traffic from Search Console and compare them against the PMax search terms report. If PMax is bidding on queries where you already rank #1-3 organically, it's paying for traffic you'd get for free. This is a form of cannibalization most agencies never check.
• Identify organic traffic drops that coincide with PMax launch. If organic clicks for certain queries dropped after PMax went live, PMax ads might be pushing organic results further down the page on branded or high-ranking queries, effectively forcing you to pay for your own traffic.
• Use Search Console to fill the "other search terms" gap. When Google Ads hides PMax search terms, Search Console can show you what queries are driving traffic to pages that PMax is also targeting. It's not a perfect overlap, but it gives you directional data about the search landscape PMax is operating in.
• Spot content gaps. If Search Console shows high-impression, low-CTR queries in your niche, those are potential search themes to add to PMax asset groups — or opportunities where PMax could be deployed strategically rather than broadly.

The power move is layering all three data sources. Google Ads tells you what PMax is reporting. GA4 tells you what PMax traffic is actually doing on site. Search Console tells you the broader search context PMax is operating within. No single source gives you the full picture.

When to Restructure vs. When to Let It Learn

This is the judgment call that separates experienced PMax managers from everyone else. Restructuring too early kills the learning phase. Waiting too long burns budget on a broken setup.

Let it learn when:

• The campaign has been live for less than 4 weeks and has fewer than 50 conversions.
• Performance is inconsistent but trending in the right direction week over week.
• The search terms you can see are relevant, even if volume is low.
• GA4 engagement metrics for PMax traffic are reasonable (not dramatically worse than other channels).
• You recently made changes to asset groups, search themes, or audience signals. Every significant change restarts the learning phase.

Restructure when:

• It's been 6+ weeks with sufficient budget and conversions are consistently below target with no improving trend.
• The search terms report is dominated by irrelevant queries that don't match your search themes.
• GA4 shows PMax traffic has dramatically lower engagement than other paid channels.
• Account-level performance is flat or worse since PMax launched — total conversions haven't increased.
• Placements are overwhelmingly low-quality Display and app inventory.
• The campaign structure has fundamental issues — single asset group covering multiple services, no audience signals, missing asset types.

When you do restructure, resist the urge to change everything at once. Restructure the asset groups first. Give it 2-3 weeks. Then adjust search themes. Then audience signals. If you change all three simultaneously, you won't know which change made the difference — and you'll be back to guessing in another six weeks.

Important:Every time you make a significant structural change to a PMax campaign, the learning phase restarts. Budget during the learning phase is essentially R&D spend. Make sure your client understands that before you start rebuilding — or you'll be having a harder conversation in three weeks when performance temporarily dips.

The Real Challenge: Doing This Across Three Platforms Simultaneously

Everything in this post is doable. None of it is quick. A proper PMax diagnostic means pulling data from Google Ads (search terms, asset performance, placements, insights), GA4 (engagement metrics, conversion paths, landing page performance, audience comparisons), and Search Console (organic query overlap, traffic cannibalization, content gaps) — then cross-referencing all of it to build a coherent picture of what's happening.

Most agencies don't do this. Not because they can't, but because it takes 4-6 hours per client and the billable math doesn't work when you're managing 15 accounts. So the PMax campaign keeps running, the client keeps asking questions, and nobody has a clear answer because nobody has done the cross-platform analysis required to generate one.

That's the gap this type of diagnosis fills. Not more dashboards. Not more automated alerts. An actual human analysis that connects the dots between what Google Ads reports, what GA4 reveals, and what Search Console exposes.

Need a PMax Diagnosis?

PMax diagnostics require cross-referencing Google Ads, GA4, and Search Console simultaneously. That's exactly what our $497 Campaign Diagnosis does. We get read-only access to all three platforms, run the full diagnostic framework outlined above, and deliver a report showing exactly where the PMax campaign is breaking down — with specific, prioritized recommendations for what to fix first.

You get the answers. Your client gets the proof. The campaign gets the fix it actually needs instead of another round of guessing.