Performance Max brand cannibalization is one of the most common and least diagnosed problems in ecommerce Google Ads accounts. It happens when Performance Max campaigns claim credit for conversions that branded search or organic channels would have captured anyway, inflating reported ROAS while actual revenue growth stalls or declines. This article follows a representative ecommerce brand through that exact pattern: strong ROAS on paper, shrinking revenue in reality, and the three structural fixes that reversed the trend. The punchline is counterintuitive. Reported ROAS dropped after the fix. Revenue went up. The business is now scaling again for the first time in months.
If your ecommerce Google Ads ROAS looks good but revenue is flat, this story will sound familiar. And it will show you exactly what to look for and what to change.
The Situation: A Growing Ecommerce Brand With A Stalling Google Ads Account
Business Profile And Ad Spend Level
The brand in question is a mid-market ecommerce business selling consumer goods direct-to-consumer, running roughly $60K per month in Google Ads spend. They had been advertising on Google for over two years, with a mature account that included Search, Shopping, and Performance Max campaigns. The business had grown steadily year over year, with Google Ads as the primary paid acquisition channel.
They had an in-house marketer managing the account day-to-day, supported by a freelance PPC consultant doing monthly reviews. The team was competent. The account was not neglected or poorly built. That is what made the problem so hard to spot.
What The Account Looked Like On The Surface
At a glance, the account looked healthy. Performance Max was the largest campaign by budget and reported the highest ROAS in the account, consistently north of 6x. The Shopping campaigns were performing steadily. Branded Search was running but had been scaled back over the previous six months as the team shifted budget toward PMax, which appeared to be doing the heavy lifting.
The Google Ads dashboard told a story of efficiency. Every standard metric pointed in the right direction.
The Number That Hid The Real Problem: Strong ROAS, Shrinking Revenue
The problem was not inside Google Ads. It was in the P&L. Top-line revenue from new customers had been declining for three consecutive months even as ad spend stayed flat and reported ROAS held steady. The business was spending the same, reporting strong returns, but growing less.
This is the signature pattern of Performance Max attribution issues in ecommerce. The numbers inside the ad platform look fine. The numbers inside the bank account do not.
The Problem: Efficient Ads That Were Not Driving Business Growth
How Performance Max Was Dominating Branded Search
The root cause was PMax brand cannibalization. When the team looked at the actual search terms driving conversions inside Performance Max, a disproportionate share were branded queries. People searching for the company name, the company name plus a product, the company name plus "discount." These were existing customers and people who already knew the brand, and PMax was serving them ads and taking credit for the sale.
This matters because branded search traffic converts at high rates regardless of whether an ad is served. These people were going to buy anyway. Performance Max was simply standing in front of the register and claiming it drove the transaction.
The Attribution Illusion: Claiming Credit For Organic Conversions
The problem went deeper than branded search overlap. Performance Max uses a broad attribution model that credits the campaign for conversions across multiple touchpoints. In this account, PMax was claiming credit for conversions where the last meaningful interaction was an organic search visit or a direct navigation, but PMax had served an impression somewhere in the user's journey, often a Display or YouTube impression the user never meaningfully engaged with.
The conversion count inside Google Ads was real. The incremental business value of those conversions was not.
Why Every Optimization Was Making The Numbers Look Better And The Business Worse
Here is where the pattern becomes self-reinforcing. The in-house marketer and freelance consultant saw PMax delivering a 6x+ ROAS. Smart bid strategies pull budget toward what converts cheaply. So the algorithm, doing exactly what it was told, kept finding more branded and existing-customer traffic. Every "optimization" pushed PMax further toward the easy wins and further away from actual prospecting.
Budget that could have gone toward finding new customers was being spent on people who would have converted anyway. The reported ROAS kept looking good. The business kept shrinking.
The Moment The Team Realized Something Structural Was Wrong
The breakthrough came during a quarterly business review when the CEO compared Google Ads reported revenue against actual new customer revenue from their ecommerce platform. The gap was enormous. Google Ads reported roughly $360K in attributed revenue for the quarter. Platform data showed new customer revenue had actually declined quarter over quarter by a meaningful amount. The math simply did not add up.
That is when they realized the problem was not tactical. It was structural. And no amount of bid adjustments or audience tweaks inside the existing campaign architecture would fix it.
The Diagnosis: What A Proper Audit Revealed
Cannibalization Of Brand Traffic By PMax Asset Groups
A proper audit started with the search terms report for Performance Max, which requires pulling the insights report since PMax does not give you a traditional search terms view. The data showed that an outsized percentage of PMax search conversions came from branded or near-branded terms. The branded Search campaign, which had historically converted at high volume, had seen its impression share drop significantly as PMax absorbed that traffic.
This is textbook brand term cannibalization. Two campaigns bidding on the same high-intent branded queries, with PMax winning the auction and claiming full credit.
Conversion Counting That Inflated Reported ROAS
The audit also revealed a conversion tracking problem that compounded the attribution issue. The account was counting multiple conversion actions, some of which were not actual purchase events, as primary conversions. Page views of the thank-you page, newsletter signups, and add-to-cart events were all feeding into the ROAS calculation alongside actual transactions. This inflated the conversion count and made every campaign, but especially PMax, look better than it was.
A Campaign Structure That Rewarded The Algorithm For Easy Wins
The campaign structure itself was the enabler. Performance Max was running a single asset group with broad targeting and no brand exclusions. The algorithm had full freedom to find the cheapest conversions, and the cheapest conversions were always branded queries from existing customers. There was no structural separation between prospecting and remarketing, between branded and non-branded, between new and returning customers.
This is one of the patterns outlined in detail in this structural fixes case study. When the campaign architecture rewards the algorithm for easy wins, it will take them every time.
Budget Allocation Across Channels That Made No Logical Sense
Finally, the budget allocation across campaign types reflected the distorted reporting. PMax was getting over 60% of total ad spend because it reported the highest ROAS. Standard Shopping and non-branded Search, the actual prospecting engines, were running on reduced budgets because their reported ROAS was lower. The account had been progressively defunded of everything that drives new customer acquisition and overfunded on the campaign taking credit for existing demand.
The Fix: Three Structural Changes That Changed The Outcome
Fix One: Brand Campaign Isolation And PMax Brand Exclusions
The first and most impactful change was implementing brand exclusions in Performance Max. Google now allows advertisers to add brand exclusion lists to PMax campaigns, preventing them from serving on branded search queries. The team applied this, then ensured the dedicated branded Search campaign was running with sufficient budget and bids to capture all branded traffic.
This single change immediately separated branded from non-branded performance in the reporting. For the first time, the team could see what PMax was actually doing on non-branded prospecting, and the numbers were far less impressive than the blended view had suggested.
Fix Two: Conversion Action Restructuring To Reflect Real Purchase Events
The second fix was cleaning up conversion tracking. The team removed all non-purchase conversion actions from the "primary" conversion category and set them to "secondary" so they could still be observed but would not feed into automated bidding or ROAS calculations. Only actual completed purchases with revenue values remained as primary conversions.
This forced the bidding algorithms to optimize toward real transactions rather than inflated conversion counts. It also meant that reported ROAS would now reflect actual return on ad spend, not a composite of transactions and micro-conversions.
Fix Three: Campaign Budget Reallocation Toward Prospecting
With clean data now flowing, the team reallocated budget. PMax received a reduced share of total spend, constrained to non-branded prospecting. Standard Shopping was scaled up with a focus on non-branded product queries. Non-branded Search campaigns targeting high-intent commercial keywords received increased budget.
The logic was straightforward. Branded traffic would convert through the dedicated brand campaign at a fraction of the cost PMax had been charging for the same clicks. The freed-up budget went toward actually finding new customers.
The Results: What Changed After The Structural Reset
ROAS Dropped On Paper, Revenue Went Up In Reality
In the first month after the structural changes, reported ROAS across the account declined. This was expected. The inflated numbers from branded cannibalization and overcounted conversions were gone. What replaced them was a lower but honest ROAS figure that reflected actual non-branded prospecting performance.
Meanwhile, total revenue from new customers, measured in the ecommerce platform rather than Google Ads, increased in the second month and continued climbing. The business was spending the same amount but acquiring customers it was not reaching before.
New Customer Acquisition Rate Before And After
The most meaningful metric shift was in new customer acquisition. Before the fix, the percentage of total orders coming from first-time buyers had been declining quarter over quarter. After the structural reset, that trend reversed. The business was reaching people who had never heard of them before, rather than paying to re-acquire people who were already searching for the brand by name.
How The Business Now Thinks About Google Ads Performance Measurement
The team now evaluates Google Ads using a layered measurement approach. In-platform ROAS is monitored but not treated as a business metric. Instead, they track blended customer acquisition cost, new-versus-returning customer revenue split, and incrementality by comparing periods where campaigns are paused against periods where they run. The reported ROAS number is lower than it was during the cannibalization period. The business is growing faster than it has in over a year.
What This Account Teaches Every Ecommerce Advertiser
Why Reported ROAS Is Not A Business Metric
Reported ROAS inside Google Ads measures what the algorithm claims credit for. It does not measure what the ad spend actually caused. When Performance Max cannibalizes branded search, reported ROAS goes up because the campaign is claiming credit for high-converting branded traffic. Business growth goes down because no new demand is being generated. Every ecommerce advertiser needs to separate in-platform metrics from business outcomes, and the moment those two diverge is the moment you have a structural problem.
The Structural Audit Every Ecommerce Account Needs
This is not a one-off story. Performance Max brand cannibalization is widespread across ecommerce accounts, especially those that have scaled PMax without implementing brand exclusions. If you are running PMax alongside branded Search and you have not explicitly excluded your brand from PMax, some degree of cannibalization is almost certainly happening. The audit is straightforward: compare PMax search term insights against your branded Search campaign, reconcile Google Ads reported revenue with platform revenue, and check which conversion actions are set as primary.
If your in-house team does not have the bandwidth or the structural experience to run this audit, this is exactly where groas changes the equation. The proprietary engine trained on over $500 billion in profitable ad spend flags these cannibalization patterns automatically, and depending on the product, a senior strategist either advises your team on how to fix it or fixes it directly.
When To Stop Optimizing And Start Rebuilding
The lesson from this account is that some problems cannot be optimized away. The in-house marketer and freelance consultant were both competent. They were making smart tactical moves within a broken structure. The bids were fine. The ads were fine. The audiences were fine. The architecture was wrong, and no amount of tactical optimization compensates for a structural flaw.
For businesses running Google Ads in-house or through a freelancer, this is the gap that shows up over time. Tactical execution reaches a ceiling, and structural problems compound quietly until revenue growth stalls. This is the scenario groas is built for. In a Done With You engagement, a senior strategist works alongside your team to catch and fix structural issues while you stay in control of execution. In a Done For You engagement, groas owns the entire account end-to-end, and problems like PMax cannibalization get caught before they cost you months of growth.
The proprietary engine does not just optimize bids. It monitors structural health across campaign architecture, conversion tracking, budget allocation, and attribution. The senior strategist layer ensures that when something is structurally wrong, it gets diagnosed and rebuilt, not just tweaked.
Month-to-month, no long-term contracts, $0 onboarding. If your ecommerce account has strong ROAS but stalling revenue, the structural audit is the place to start.
For in-house teams who want to keep driving but need the engine and a strategist, get started with Done With You. For businesses that want Google Ads fully handled, apply for Done For You and groas figures out the right plan on the call.
Frequently Asked Questions
What Is Performance Max Brand Cannibalization?
Performance Max brand cannibalization occurs when a PMax campaign serves ads on branded search queries that a dedicated brand Search campaign or organic results would have captured anyway. PMax claims credit for these high-converting branded transactions, inflating reported ROAS while contributing little or no incremental revenue. The core issue is that the algorithm gravitates toward the easiest conversions, which are almost always branded queries from people who already know your business. Without brand exclusions applied to PMax, this cannibalization happens by default in most ecommerce accounts.
How Do I Know If Performance Max Is Cannibalizing My Branded Search?
Check three things. First, pull the PMax search term insights report and look at the proportion of conversions coming from branded or near-branded queries. Second, compare your dedicated branded Search campaign's impression share over the past six months. If it has declined while PMax spend increased, overlap is likely. Third, reconcile Google Ads reported revenue against your ecommerce platform's actual new customer revenue. If Google Ads reports strong returns but platform-level new customer revenue is flat or declining, PMax is likely claiming credit for conversions it did not cause.
Why Does My Google Ads ROAS Look Good But Revenue Is Flat?
This is the signature symptom of PMax attribution issues combined with brand cannibalization. Performance Max optimizes toward the cheapest conversions, which are typically branded queries from existing customers. The reported ROAS stays high because those conversions are easy and inexpensive. But because the budget is being spent on people who would have converted anyway, no new demand is generated. Revenue stagnates or declines even as in-platform metrics look healthy. The fix requires structural changes to campaign architecture, not bid or audience adjustments.
Should I Turn Off Performance Max Completely?
Not necessarily. Performance Max can be a productive prospecting channel when properly constrained. The key is applying brand exclusions so PMax cannot bid on branded queries, ensuring only real purchase events are set as primary conversions, and allocating budget so PMax is not consuming spend that should go to non-branded Search and Shopping. The campaign type is not the problem. The lack of structural guardrails is.
How Do I Exclude Brand Terms From Performance Max?
Google allows you to add a brand exclusion list directly within PMax campaign settings. Navigate to your PMax campaign, go to settings, find the brand exclusions section, and add your brand name and common variations. This prevents PMax from serving ads on searches that include your brand terms. You should also ensure your dedicated branded Search campaign has sufficient budget to capture all branded traffic at a lower cost per click than PMax was paying.
What Conversion Actions Should Be Set As Primary In Google Ads?
For ecommerce accounts, only actual completed purchase events with accurate revenue values should be set as primary conversion actions. Micro-conversions like add-to-cart, newsletter signups, or thank-you page views should be set as secondary or observation-only. When non-purchase events are set as primary, they inflate ROAS calculations and cause automated bidding strategies to optimize toward low-value actions rather than real transactions.
How Does groas Detect PMax Cannibalization Before It Hurts Revenue?
The groas proprietary engine, trained on over $500 billion in profitable ad spend, continuously monitors structural health across campaign architecture, conversion tracking, and attribution patterns. It flags cannibalization signals like branded query overlap, declining new customer ratios, and divergence between in-platform ROAS and actual revenue. In a Done With You engagement, a senior strategist advises your team on structural fixes. In a Done For You engagement, groas catches and resolves these issues directly before they cost you months of lost growth.
How Long Does It Take To See Results After Fixing PMax Cannibalization?
Expect reported ROAS to decline in the first few weeks as inflated branded attribution is removed from the numbers. This is a sign the fix is working, not a regression. Actual new customer revenue typically begins increasing within the second month as budget reallocates toward real prospecting. Full stabilization of the new campaign structure usually takes six to eight weeks, after which you have honest reporting and a growth trajectory that matches reality.
Can groas Help If I Already Have An In-House Google Ads Team?
Yes. groas Done With You is built specifically for this scenario. Your in-house team stays in the driver's seat while the proprietary engine handles heavy execution underneath. A senior strategist provides a weekly report on what was done plus a strategy call every other week, catching structural problems like PMax cannibalization that in-house teams often lack the bandwidth or cross-account pattern recognition to diagnose. Month-to-month, no long-term contract, $0 onboarding. Get started if your team is ready to scale with better structural support.
.png)
