Performance Max Is
Spending Your Budget.
But Is It Actually Working?

The PMax Illusion: Why the Dashboard Can Look Great While Performance Suffers

Performance Max was designed to find conversions across every Google surface simultaneously. The algorithm optimises budget allocation across Search, Shopping, YouTube, Display, Gmail, Discover, and Maps based on predicted conversion probability. The promise is compelling: one campaign, the full reach of Google's network, and AI handling the complexity of cross-channel budget decisions.

The problem is that this architecture creates several structural incentives that work against the advertiser. PMax's algorithm naturally gravitates toward the easiest conversions available - and the easiest conversions are almost always branded search queries, where people already know you and were going to convert anyway, and Display placements, where inventory is cheap and attribution is loose. A PMax campaign can spend $10,000 per month, report 400 conversions and a strong ROAS, and still be generating close to zero net-new revenue - simply by absorbing traffic that your Search campaigns, organic search, and direct traffic would have captured without the additional spend.

The five diagnostic checks below give you a systematic framework for evaluating your PMax campaign honestly. Work through each one before drawing any conclusions about whether to scale, restructure, or reduce your PMax investment.

Where Is Your Budget Actually Going Across Google's Channels?

Until April 2025 this question was almost completely unanswerable. PMax operated as a genuine black box - one budget number in, aggregate conversion numbers out, with no visibility into which of Google's seven inventory surfaces received which portion of spend. That changed with Google's rollout of Channel Performance Reporting, and in April 2026 Google added a Channel Performance Timeline that shows week-over-week trends by channel. If you have not checked this report in your PMax campaign, it is the single most important thing to do before any other diagnostic step.

How to Access Channel Performance Data

In Google Ads, navigate to your Performance Max campaign. Click Insights and Reports in the left navigation, then select Channel Performance. You will see a breakdown of spend, conversions, and conversion value across Search, Shopping, YouTube, Display, Gmail, Discover, and Maps. Switch to the Timeline view (available from April 2026) to see how channel distribution has shifted over time. A single snapshot can be misleading - a 30 to 90 day trend reveals whether the algorithm has drifted toward lower-quality channels as the campaign has matured.

What Healthy vs Unhealthy Channel Distribution Looks Like

If your Google Ads account structure shows Display absorbing significant budget at poor conversion rates, the most immediate fix is adding placement exclusions. Google allows up to 65,000 placement exclusions per account. Start with known junk categories: mobile gaming apps, parked domains, low-quality content sites, and any placement category where your brand should not appear.

Is PMax Cannibalising Your Search Campaigns and Stealing Brand Traffic?

This is the most financially consequential PMax problem and the least visible one. Cannibalization occurs when PMax absorbs traffic - particularly branded search queries - that your existing Search campaigns or organic results would have captured anyway, at a lower cost and with better conversion quality. The result on your dashboard: PMax conversion count rises, Search conversion count falls, total revenue stays flat. You are paying more to deliver the same results.

The Optmyzr study of 503 accounts is the most comprehensive independent analysis of this problem. It found keyword overlap between Search and PMax in 91.45% of accounts. Among those accounts, the overlap was not limited to broad match keywords. It occurred across all match types, including exact match. Even accounts with exact match branded keywords in dedicated Search campaigns saw PMax winning impressions for those queries. This happens because campaign priority rules have exceptions - if a Search campaign is ineligible for an auction due to budget constraints, geographic settings, or scheduling, PMax fills the gap. And once PMax claims a branded query, it typically claims the conversion credit too.

The Three Cannibalization Patterns to Check For

• ✓ Branded query cannibalization PMax absorbs branded search traffic because branded queries typically have high conversion rates and low CPCs - they look like easy wins to the algorithm. Your branded Search campaign impressions and clicks decline after PMax launch. Total branded conversions stay similar but cost rises because PMax is less efficient than a dedicated branded Search campaign for these queries
• ✓ Non-brand Search overlap PMax triggers for the same non-brand queries as your Search campaigns. The Optmyzr study found this in 56.29% of Search campaigns. When PMax wins the auction for a query that your Search campaign's exact match keyword was targeting, Search loses the impression and any resulting conversion is credited to PMax regardless of which campaign type would have performed better
• ✓ Organic traffic displacement PMax's URL expansion feature can serve ads for queries where you already rank organically in position 1-3. A user who would have clicked your organic result for free now clicks your PMax ad at a cost. Your organic CTR declines, paid traffic for those queries rises, and total traffic appears stable while efficiency falls. Check your Google Search Console organic CTR trends against PMax launch date

How to Diagnose Cannibalization in Your Account

1. Compare Account-Level Metrics Before and After PMax Launch

   Pull total account conversions, total account spend, and total revenue for the 90 days before PMax launched and the 90 days after. If conversions rose but revenue did not rise proportionally, or if cost rose while revenue stayed flat, cannibalization is the most likely cause. This top-down view is more reliable than campaign-level comparison because PMax will always look better in campaign-level reporting.

2. Check Your Search Campaign Impression Share Trends

   Look at your Search campaign impression share over the period since PMax launched. If impression share is declining without a corresponding decline in search volume or bid competitiveness, PMax is absorbing traffic that Search would have captured. This is the most direct signal of cannibalization in real-time account data.

3. Review PMax Search Terms for Brand Overlap

   Go to your PMax campaign Insights tab and look at the Search Terms Insights report. If your brand name, your domain, or close variants of your branded keywords appear as queries driving PMax conversions, PMax is absorbing branded traffic. Every branded conversion claimed by PMax is a conversion your cheaper, more controlled branded Search campaign would have delivered at lower cost.

4. Run the Pause Test

   Pause your PMax campaign for 2 to 4 weeks and monitor what happens to Search campaign performance and total revenue. If Search impressions, clicks, and conversions rise significantly and total revenue stays the same, PMax was cannibalizing. If total revenue declines materially, PMax was delivering genuine incremental conversions beyond what Search captured alone. This is the most direct test available.

Are PMax Conversions Truly Incremental or Just Attribution Credit?

This is the question at the heart of the entire PMax debate - and it is the one that Google's native reporting cannot answer. Attribution models assign credit for conversions to campaigns based on touchpoints. Incrementality testing measures something fundamentally different: whether the ad spend actually caused the conversion to happen, or whether the customer would have converted anyway through another channel, a direct visit, or organic search.

The gap between attributed performance and incremental performance is consistently larger than most advertisers expect. A PMax campaign reporting a 4x ROAS in Ads Manager might deliver a 2x incremental ROAS when properly tested - meaning half the revenue PMax is claiming credit for would have been generated without PMax running at all. This is not fraud; it is the natural behavior of a last-touch and data-driven attribution system operating in an account where PMax overlaps with high-intent Search traffic and organic results.

Three Ways to Test PMax Incrementality

• ✓ Google's Built-In PMax Experiment Tool (recommended first step) Go to Campaigns, then Experiments in Google Ads. Google's native PMax experiment creates a traffic holdout - a percentage of users who would have seen your PMax ads instead see no PMax ads. After 4 to 6 weeks, Google reports the incremental conversion lift. This is the lowest-friction incrementality test available and requires no third-party tooling. The limitation is that Google controls the statistical methodology, which means results should be verified against your own business data
• ✓ Geo-based holdout test Select comparable geographic markets - similar cities or regions in the USA, UK, UAE, Australia or wherever you are running campaigns. Pause PMax in the holdout markets while keeping it active in the test markets. After 30 days, compare total conversion volume and revenue between markets. This approach is more reliable than the platform's own experiment because it measures actual business outcomes rather than Google's attribution model
• ✓ The revenue growth test The simplest version of incrementality testing requires no tools: compare total account revenue and total account conversions, not campaign-level figures. If PMax spend rose 50% and total account revenue rose 50%, PMax is likely incremental. If PMax spend rose 50% and total account revenue rose 5%, PMax is largely recycling existing demand. Use your CRM or ecommerce platform data, not Ads Manager conversion counts, as the source of truth

What Quality Are the Assets Your PMax Is Actually Serving?

PMax's performance is directly constrained by the quality of what you put into it. The algorithm's job is to assemble the best possible ad from your asset library for each placement and user. If your asset library is thin, generic, or misaligned with your conversion goals, the algorithm has nothing good to work with - and it will default to lower-quality placements where poor creative is less damaging to click rates.

This is one of the most consistently underinvested areas in PMax accounts. Many advertisers launch PMax with repurposed creative from other campaigns - stock photos, generic headlines, and no video - and then conclude PMax does not work when the issue is entirely the assets they provided. Google's own internal data shows video-enabled PMax campaigns outperform image-only campaigns by 25 to 40% on average. The algorithm cannot manufacture performance from weak inputs.

Asset Quality Audit

• ✓ Check Asset Strength rating in each Asset Group In Google Ads, navigate to your PMax campaign, then Asset Groups, then View Details. Google rates each asset group as Poor, Good, or Excellent based on the completeness and variety of assets provided. Any asset group rated Poor is actively limiting campaign performance. The rating considers headline variety, image variety, video presence, and description coverage
• ✓ Review individual asset performance ratings Within Asset Details, Google now rates individual assets as Low, Good, or Best. Any asset rated Low after sufficient impressions should be replaced. If multiple headline assets are rated Low, your ad copy is not resonating with the audience that asset group is reaching - rewrite with a focus on specific benefits and clear CTAs rather than generic brand messaging
• ✓ Confirm genuine video assets are present for YouTube inventory If your Channel Performance report shows YouTube receiving budget and you cannot confirm dedicated video assets were uploaded, Google is serving auto-generated video. Check Asset Groups and look for any video assets. Auto-generated videos are identifiable by their title format. If present, either upload proper 15-30 second video assets or add YouTube as a placement exclusion
• ✓ Verify audience signals are using your actual customer data PMax's audience signals tell the algorithm which users to prioritise when learning. Generic in-market audiences are weak signals. Your Customer Match list of existing buyers is the strongest possible signal. If you have not uploaded your customer email list as a Customer Match audience in your PMax asset group signals, the algorithm is learning from general interest patterns rather than from the actual behaviour of people who buy from you
• ✓ Check URL exclusions are blocking non-converting pages PMax's Final URL Expansion can direct traffic to any page on your site, including blog posts, career pages, and informational content. Go to Campaign Settings and add URL exclusions for any page that does not have a clear conversion path. Without exclusions, PMax may spend budget driving traffic to pages that never convert - and the algorithm learns from those visits as if they are valid signals

Is the Conversion Signal PMax Is Learning From Clean and Accurate?

PMax is only as good as the conversion data you feed it. This is not a platitude - it is a mechanical constraint. The algorithm makes every budget allocation, channel selection, and bid decision based on conversion signal quality. If that signal is inaccurate, the consequences compound over time: PMax optimises harder toward the wrong conversions, channel distribution drifts toward placements that generate the most falsely attributed events, and the reported ROAS diverges further from actual business results with every passing week.

Unlike Search campaigns, where conversion tracking errors affect bidding but do not redirect entire budget pools across multiple channels, PMax conversion signal quality determines channel mix, audience targeting, and geographic distribution simultaneously. A conversion tracking error in a PMax account has more widespread consequences than the same error in a standalone Search campaign.

PMax Conversion Signal Checklist

• ✓ Verify you are optimising for a genuine revenue-generating event PMax should optimise toward the event closest to actual revenue: a Purchase, Qualified Lead Form Submission, or confirmed booking. If your conversion action is a page view, session duration, or add-to-cart event used as a proxy, PMax will optimise toward finding people who browse, not people who buy. Update conversion actions in campaign settings to use your highest-quality conversion events
• ✓ Check for duplicate conversion actions inflating counts A common error: multiple GTM tags and hardcoded tracking scripts firing simultaneously for the same conversion event. In Tools and Settings, under Conversions, review all active conversion actions. Any that measure the same event type as another active action should be removed or set to Secondary. Duplicate conversions teach PMax that acquisition is twice as easy as it actually is
• ✓ Confirm Enhanced Conversions are configured correctly Enhanced Conversions send hashed first-party customer data (email, phone) alongside conversion events, significantly improving signal quality in a post-cookie environment. Check your data layer is passing these parameters correctly. A match rate below 40% in your Enhanced Conversions diagnostics means a significant portion of your conversions are reaching Google without user identifiers, reducing signal quality and impairing PMax's audience learning
• ✓ Reconcile Ads Manager conversions against CRM data monthly Compare Google Ads reported PMax conversions against your actual CRM leads or orders for the same 30-day period. A ratio above 1.3:1 means your tracking is over-reporting conversions to PMax, causing the algorithm to believe it is more efficient than it is and encouraging further budget allocation. Fix the tracking before attempting any other optimisation

5 Structural Fixes That Actually Make PMax Work Better

Once the five diagnostic checks are complete, most accounts will have identified at least two or three specific problems. The fixes below correspond directly to the most common findings. Apply them in the order listed - each one improves the signal quality that makes the subsequent fixes more effective.

Fix 1: Add Brand Exclusions and Protect Search With Exact Match

Add your brand name, brand variations, product names, and domain as brand exclusions in PMax campaign settings. This single change stops PMax from claiming credit for branded conversions that your organic results and dedicated branded Search campaigns were already capturing. Run a 30-day comparison after making this change. If total branded conversion volume does not decline, you have confirmed PMax was cannibalizing and you have just improved efficiency at no cost.

For non-brand terms, protect your highest-converting Search keywords by running them as Exact Match in a dedicated Search campaign alongside PMax. Google's auction rules give Exact Match Search campaigns priority over PMax when a user's query exactly matches your keyword. This ensures your most valuable traffic is handled by the campaign type you control, while PMax focuses on expansion into new territory.

Fix 2: Add Placement Exclusions for Low-Quality Display Inventory

Navigate to your Google Ads account settings (not just the PMax campaign settings), then Placements, then Exclusions. This is where account-level placement exclusions live - these apply across all campaigns including PMax. Add exclusions for: mobile gaming app categories, parked domain networks, low-quality content site categories, and any specific placements identified in your Channel Performance report as receiving budget with zero conversions. Google allows up to 65,000 placement exclusions per account.

Fix 3: Upload Your Customer Match List as the Primary Audience Signal

Export your existing customer email list from your CRM. Go to your PMax campaign, click on the Asset Group, then Audience Signals, and add the Customer Match list as your primary audience signal. This is the strongest possible input you can give PMax's learning system - real behavioral data from people who have already bought from you. The algorithm will use this to identify similar users across all channels, significantly improving the quality of prospecting traffic PMax generates compared to starting from generic in-market audiences.

Fix 4: Add URL Exclusions for Non-Converting Pages

PMax's Final URL Expansion can send traffic to any page on your site. Go to Campaign Settings in your PMax campaign and add URL exclusions for blog posts, team pages, careers pages, press pages, and any other content pages that do not have a conversion path. For ecommerce accounts this typically means excluding all category and informational pages and allowing PMax to send traffic only to product pages and collection pages with clear purchase CTAs. This also improves the landing page conversion rate for PMax-driven traffic by ensuring users reach pages built to convert.

Fix 5: Run a PMax Experiment Before Making Major Budget Decisions

Before scaling PMax budget or shutting it down entirely, run Google's built-in PMax Experiment. Go to Campaigns, then Experiments. Create a new experiment comparing your current PMax setup against either a holdout (no PMax) or a modified version with the structural fixes applied. Run the experiment for a minimum of 4 weeks. The experiment report shows true incremental conversion lift - the difference in conversions between the PMax-active group and the holdout. This data, cross-referenced against your CRM revenue figures, gives you the most reliable answer available to the question this entire guide started with: is Performance Max actually working?

Frequently Asked Questions