Performance Max: why your ROAS is inflated and how to reclaim your brand budget
Why Performance Max ROAS can lie to you
Performance Max can inflate your ROAS because, by default, it also bids on queries containing your brand name — and branded traffic converts unusually cheaply and highly, since these are people who were already looking for you. The campaign claims those conversions and reports a spectacular return, when in reality it only intercepted sales you would have made with no ads at all. It's the same trap we covered in our piece on Google Ads and Meta Ads budgets: a high ROAS in the dashboard doesn't mean high profit in the bank.
The gap between 6.0x and 2.5x isn't cosmetic — it decides how much you can actually afford to pay to acquire a new customer. If you scale budget believing in a 6x return while you really have 2.5x, you burn money at exactly the rate you top the campaign up.
How many campaigns really cannibalize Search? The hard data
Performance Max overlapping with Search campaigns is the rule, not the exception. In a 2025 Optmyzr study of more than 500 ad accounts, 91% of accounts had exact keyword overlap between PMax and Search, and 97% of PMax campaigns overlapped with search campaigns. In other words: if you run PMax and Search at the same time, they compete for the same queries with near-certainty.
| What was measured | Result |
|---|---|
| Accounts with exact keyword overlap PMax ↔ Search | 91.4% |
| Accounts with search-term overlap | 97.3% |
| PMax campaigns overlapping with Search | 97.4% |
| Cases where Search had the better conversion rate | 18.9% (vs 6.2% for PMax) |
That last row matters most. On queries where both campaign types overlapped, classic Search more often won on conversion rate — in 18.9% of cases, versus 6.2% for PMax. So the automation doesn't just take the traffic; it often handles it worse than the dedicated campaign it just starved of budget.
Where the inflated ROAS comes from
An inflated PMax ROAS comes from three overlapping mechanisms: capturing branded traffic, last-click attribution, and a lack of search-term transparency. Each one alone shifts credit toward the automation; together they can double the reported return relative to the real one.
- Branded traffic. PMax happily shows on brand-name queries because they have the highest conversion rate in the whole account. These are the cheapest conversions that exist — and the automation claims them first.
- Fuzzy match. Even with brand exclusions on, PMax still triggers ads for misspellings and name variants (e.g. "zets" instead of "zest") that the exclusion list doesn't catch automatically.
- Opaque search terms. Until recently PMax didn't show which queries it served at all. Even today the report is trimmed — it's hard to prove how many conversions are real acquisition versus intercepted traffic.
How to check whether your PMax is burning brand budget
To check how much brand budget your Performance Max is burning, compare branded conversions before you launched PMax with the period after — if PMax "delivered" conversions but the total number of brand-name transactions didn't rise, that's a sign the campaign only took over existing traffic instead of adding new demand.
- Open the search-terms report (Insights → Search terms) and measure what share of clicks and conversions comes from queries containing your brand name.
- Compare the PMax conversion rate with your dedicated branded Search campaign. If they're similar, PMax is probably living off branded traffic.
- Check the trend of branded conversions over time: after PMax launched, did their total really grow, or did they just "flow" from Search into PMax?
- Recalculate ROAS after excluding revenue from branded queries. That's your real return on new-customer acquisition — and that number, not the dashboard figure, should drive budget decisions.
To see real profit rather than revenue attributed to the campaign, connect this data to a dashboard that links ad spend with sales and margin. Only then will you see whether that "6x ROAS" from PMax actually leaves anything in the bank.
Brand exclusions — how to set them and why they aren't enough
Brand exclusions in Performance Max let you cut the campaign off from queries containing your name, but they aren't a complete fix — Google itself warns that with broad matching PMax can still show on variants and misspellings of your brand name. Treat them as a first filter, not a lock.
- Enable the brand exclusion list at the PMax campaign level (Settings → Brand exclusions) and add your company name and key product names.
- Add the most common misspellings and name variants as separate entries — fuzzy matching won't catch them automatically.
- Keep a separate, dedicated branded Search campaign so you consciously control brand traffic instead of handing it to the automation at a discount.
- After every change, return to the search-terms report and verify the exclusions actually work, rather than just looking enabled.
When Performance Max actually earns its budget
Performance Max truly earns its budget when you measure it on new-customer acquisition, not total revenue. Set the "new customer acquisition" goal, cut off branded traffic with exclusions, and judge the campaign on the real cost of winning a customer you didn't have before. Then PMax's automation across the display network, YouTube and Discover can deliver results that manual campaigns won't reach.
For an e-commerce store with a broad catalogue and a well-maintained product feed, PMax is often the most effective tool for scaling prospecting. The problem almost never sits in the tool — it sits in measuring it against a number it inflates for itself. Fix the measurement and the "PMax vs Search" fight becomes a simple division of roles: Search guards branded traffic, PMax hunts for new customers.
At Zest we treat Performance Max like any other campaign: first we separate branded traffic from real acquisition, and only then judge the return. If you want to be sure your budget works for new customers rather than the automation's vanity metrics, see how we run Google Ads campaigns, or explore how our digital marketing agency measures real return.
How to structure the account so PMax doesn't eat your brand
To stop Performance Max from eating your branded campaign, split the roles at the account-structure level: a separate Search campaign for the brand only, PMax with a new-customer goal and exclusions enabled, and a hard rule that you judge each layer on a different metric. This structure works for most e-commerce accounts because it doesn't leave the automation to decide who gets credit for the easiest conversions.
- A dedicated branded Search campaign (exact and phrase match on the brand name) with its own protected budget — that's what should handle branded traffic, not PMax.
- Performance Max with the "new customer acquisition" goal and brand exclusions enabled — measured solely on customers you didn't have before.
- Search campaigns on generic (non-branded) high-intent queries that you deliberately don't want to hand to the automation.
- One shared dashboard where you see the cost and revenue of each layer separately — otherwise you'll again see a single averaged, inflated ROAS.
FAQ: Performance Max and ROAS
Q.Does Performance Max always cannibalize Search campaigns?
It almost always overlaps with them — in the Optmyzr study, 97% of PMax campaigns overlapped with Search. Whether that's real cannibalization depends on whether PMax takes traffic you'd have had anyway (branded) or adds new demand. Without brand exclusions and search-term analysis, it usually takes.
Q.Should you turn Performance Max off?
No — the problem isn't the tool, it's how you measure it. PMax scales new-customer acquisition well. You just need to cut off branded traffic with exclusions and judge the campaign on real acquisition rather than total revenue.
Q.How do you enable brand exclusions in PMax?
In your Performance Max campaign settings, choose "Brand exclusions" and add a brand list containing your company and product names. Remember that with broad matching the campaign can still show on misspellings — add those manually as separate entries.
Q.How can you tell your PMax ROAS is inflated?
Compare the PMax conversion rate with your branded Search campaign and check the share of branded queries in the search-terms report. If most conversions come from brand-name queries, the reported ROAS is inflated relative to real acquisition.
Q.Do brand exclusions solve the problem completely?
No. Google confirms that with broad matching PMax can still trigger ads on variants and misspellings of your brand name. Exclusions need to be topped up with a misspelling list and paired with a separate, deliberately managed branded campaign.
Runs Google Ads and Meta campaigns measured on real acquisition, not the dashboard's inflated ROAS. Focused on separating branded traffic from net-new sales.
Related articles
Let's talk growth
Send a brief or drop your contact — we'll reply within 24h.
Get a quote