Why Performance Max Is Stealing Credit From Your Brand Search Campaign

Performance Max's reported return on ad spend is a genuinely impressive number on most dashboards, and that's exactly the problem. A meaningful share of that number, on a lot of accounts, is your own company name - a search that would have converted whether or not an ad ever appeared, because the visitor was already typing your brand into Google looking for you. Performance Max asset groups have no brand awareness built in by default, and they will bid on your own name alongside every other query the algorithm decides is worth targeting, quietly folding traffic you already owned into a campaign that gets credited as if it earned it.

How This Actually Happens

Performance Max is designed to find conversions wherever the algorithm believes they're available, across Search, Display, YouTube, Discover, Gmail, and Maps inventory, using a single set of assets and a single budget. That breadth is the entire selling point of the product, but it comes with a specific blind spot: PMax has no concept of "this is our brand, don't compete with our own organic and branded search presence for it" unless you explicitly tell it not to. Left alone, an asset group will happily serve an ad against your own company name, a query that, in the overwhelming majority of cases, was already going to end in a click and very likely a conversion regardless of whether a paid ad showed up at all.

The effect on reporting is what makes this genuinely deceptive rather than just wasteful. That branded conversion gets attributed to Performance Max, at a cost per acquisition that looks fantastic because branded search converts at a high rate almost by definition, someone searching your company name already knows who you are and likely already intends to buy or sign up. Blend enough of that traffic into a campaign's overall numbers and the campaign's reported ROAS climbs well above what its actual incremental performance, the results you wouldn't have gotten anyway, would show.

Why This Is Easy to Miss for Months

Nobody sets out to let this happen. It's easy to miss because nothing about the dashboard flags it as a problem, it flags it as a win. A campaign posting strong ROAS doesn't invite scrutiny, it invites more budget. The account structure that makes this worse is common: a dedicated brand Search campaign already exists, bidding on the company name at a low cost per click because there's essentially no competition for your own name, and Performance Max is running in parallel with a broad or "maximize conversion value" goal and no exclusions configured. The two campaigns are now competing for the exact same auction, and depending on bid strategy and asset strength, Performance Max can win a meaningful share of that traffic away from the campaign that was already handling it efficiently, then take credit for a conversion the brand campaign would have delivered at a fraction of the reported cost.

The tell is usually visible in the numbers if you go looking for it: a widening gap between a PMax campaign's reported conversion volume and its performance once brand terms are excluded, or a brand Search campaign whose impression share has quietly declined even though its own budget and bids haven't changed. Most accounts don't go looking, because the aggregate numbers across the account look fine.

A Quick Way to Estimate the Overlap Before You Touch Anything

Before changing any campaign settings, it's worth spending twenty minutes estimating how big this problem actually is in your account, so you're not fixing a theoretical issue or, worse, walking into a budget conversation with a guess instead of a number. Pull Performance Max's search terms insights (or listening reports, depending on what's surfaced in your account) and look specifically for your brand name and close variants appearing as a meaningful share of the terms driving reported conversions. Separately, check your dedicated brand Search campaign's impression share trend over the past few months - a slow, unexplained decline in impression share on a campaign whose budget and bids haven't changed is a strong sign something else is winning auctions it used to win outright.

Neither check is a precise incrementality measurement, but together they tell you quickly whether this is a minor rounding error in your account or a genuinely significant chunk of reported performance, which changes how urgently it's worth fixing and how much the resulting conversation with budget stakeholders needs to be handled carefully.

Setting Up Brand Exclusions Properly

Google added account-level brand exclusion controls for Performance Max specifically to address this, and the fix is a configuration change rather than a rebuild:

• Build a brand list in the shared library. Under the account's shared library, create a brand list containing your company name and its common variants and misspellings, then apply that list as an exclusion to every Performance Max campaign running in the account.
• Keep a dedicated brand Search campaign running separately, with its own tight keyword control and budget. This is the campaign that should own branded queries, at the low cost per click brand terms typically command, rather than leaving that traffic to be picked up incidentally by a broader campaign.
• Expect imperfect coverage, not a complete fix. Brand exclusion matching is good but not exact - unusual combinations of a brand term with a generic word, or lateral phrasing, can still slip through the exclusion and get picked up by Performance Max. Treat the exclusion as a major reduction in overlap, not a guarantee of zero.
• Re-run the comparison after the exclusion is live. Compare the campaign's conversion volume and cost per acquisition before and after brand exclusion is applied. The size of that gap is a reasonably honest estimate of how much of the previous reporting was brand cannibalization rather than incremental performance.

What's Actually Worth Excluding, and What Isn't

Brand exclusion is specifically about your own company name, not about narrowing Performance Max's targeting more broadly. Competitor names are a different situation entirely: a query containing a competitor's name is not traffic you already own, and in most cases there's no reason to exclude those terms from a campaign capable of capturing that demand, provided your ad and landing page experience for that traffic is actually appropriate for someone comparing options rather than already decided. The distinction matters because it's easy to over-correct after finding a cannibalization problem and start excluding broadly out of caution, which just as easily removes legitimate opportunity the campaign was correctly finding in the first place. The fix here is narrow and specific: exclude your own brand, leave the rest of the targeting alone.

Don't Stop at Exclusions. Audit Attribution Regularly.

Brand exclusion fixes the auction-level overlap, but it doesn't fix the reporting conversation on its own. A recurring, even just monthly, cross-campaign review of branded impression share, brand Search campaign performance, and Performance Max's non-brand conversion volume is what actually keeps this from drifting back. Campaign structures get rebuilt, new asset groups get added, and a well-intentioned "let's just add our whole product catalog to this one PMax campaign for simplicity" change can quietly reopen the same overlap months later if nobody's checking.

The Uncomfortable Part of This Conversation

Fixing brand cannibalization almost always makes Performance Max's reported numbers look worse, not better, at least on paper. That's the correct outcome, not a sign something broke: the campaign's real incremental contribution was always the smaller number, exclusion just makes the dashboard stop lying about it. Reporting that number honestly to whoever approves the ad budget is a harder conversation than reporting an inflated ROAS, but it's the conversation that actually protects the budget in the long run - a campaign whose real performance is understood correctly gets funded on solid ground, while one propped up by double-counted brand traffic is one bad quarter away from a very confusing budget review.