Since its launch in 2021, Performance Max reported one number per campaign and left you to guess where the money went.
That changed in 2025, after years of "black box" complaints.
Google Ads now has a channel performance report that splits a PMax campaign into Search, YouTube, Display, Discover, Gmail, Maps and search partners, with cost and conversions against each.
The first thing most advertisers do with that report is admire the pie chart. The more useful thing is to answer a question that has nagged UK accounts since PMax launched: how much of the "performance" is just branded search: people searching for your own company name, who would have clicked your brand ad or your organic listing anyway?
This article walks through that check, step by step, and then the fix.
What Performance Max reporting now shows by channel
The channel performance report breaks a Performance Max campaign down by the Google surface the ad ran on. Google announced the report on 30 April 2025, showed it at Google Marketing Live on 21 May and started rolling the beta out on 30 May.
Google's own help page lists four parts: a performance summary at campaign level (actual CPA or ROAS against target, conversions, cost), a channels-to-goals chart, a time series you can filter to one channel, and a channel distribution table.
The table is the part that matters. It carries impressions, clicks, interactions, conversions, conversion value, cost, results and results value per channel, and you can download it.
Two things to know before you open it. The data only runs from 6 June 2025, so there is no history before that. And "Shopping" is not a channel: Google treats product ads as a format called ads using product data that can appear across Search, YouTube, Maps and search partners, so you segment by format as well as by channel.
Google is also plain about the limit of the report. In its words, "while you can't directly control budget allocation per channel, you can influence it". You cannot set a bid or a budget for the Search channel. You can only exclude things.
Why the Search row is the one to worry about
The Search row is where brand cannibalisation lives, because that is the only channel where someone can type your company name. Google's definition of the Search channel covers google.com, the Google Search app and search partner sites. A brand search is the cheapest, highest-converting click in any account, and PMax has every incentive to buy it.
Google says its rules prevent this. The keyword matching help page states that "exact match keywords that are identical to the search term are prioritized" over Performance Max. The same page then lists the exceptions: the Search campaign is limited by budget, the keyword has low search volume, the ad group's ads or landing page are disapproved, or targeting requirements are not met.
Phrase and broad keywords get no such protection; Google says search themes carry the same priority as phrase and broad match.
So the rule only helps if your brand campaign has the exact query as an exact match keyword, with budget to spare, every hour of the day. Misspellings, "yourbrand reviews", "yourbrand near me" and the dozens of variants people actually type fall through.
The published data suggests the overlap is routine. Optmyzr analysed 503 accounts in February 2025 and found 91.45% had PMax serving on queries that matched a Search keyword, including exact match ones. Adalysis looked at 3,300 non-retail PMax campaigns and found 45% of PMax search terms also appeared in the account's Search campaigns, with Search converting better 84% of the time.
Overlap is nearly universal, but performance differences are usually minor.
Frederick Vallaeys, Optmyzr, Is PMax cannibalizing Search?
Vallaeys' point is that the overlap does not always cost you. The problem is that when it does, it hides inside a healthy-looking blended CPA. That is what the steps below are for.
The step-by-step check
The check takes about an hour with the channel report, the brand campaign's impression share columns and the PMax search terms report open side by side. Do the steps in this order.
Find the PMax launch date. Open the campaign's change history and note the day it went live, or the day its budget last jumped. Everything else is a before-and-after comparison around that date.
Pull brand campaign impression share for the two periods. On the Campaigns page add the competitive metrics columns Google describes on its impression share page: search impression share, search exact match IS, search lost IS (rank) and search lost IS (budget). Compare 30 days before launch with 30 days after. A drop in exact match IS with no budget cap and no bid change is the first warning sign. (Google does not give impression share for PMax itself, so the brand campaign is your only window.)
Compare brand campaign clicks and CPC across the same two periods. When a cannibalised brand campaign loses auctions it tends to lose the cheap ones, so clicks fall and average CPC rises at the same time.
Open the channel distribution table and read the Search row. Note cost, conversions and CPA for Search against the other channels. A Search CPA far below everything else, in an account where brand is the only cheap traffic, usually means PMax is buying your name.
Open the PMax search terms report. In the Search terms report there is a view called search terms and landing pages for Performance Max, with data from March 2023. Filter the terms for your brand name, misspellings and domain. Then open the search terms insights card at campaign level and look at the categories Google has built automatically. A category containing your brand name with a high share of the campaign's conversions settles the question. The insights card shows clicks, impressions and conversions but not cost, so pair it with the channel table for spend.
Add it up. Take the brand-term conversions inside PMax and ask what the brand campaign would have paid for them. The difference is what the overlap costs you, and it is usually larger than it looks because the brand campaign's own CPC has gone up in the meantime.
A worked example
The numbers below are illustrative. They are built to show the shape of the problem and are not taken from a client account. They assume a UK services business with a brand search campaign and a new PMax campaign launched on the first of the month.
| Metric | 30 days before PMax | 30 days after PMax | What it suggests |
|---|---|---|---|
| Brand campaign exact match IS | 94% | 71% | Losing auctions on your own name |
| Brand lost IS (budget) | 0% | 0% | Not a budget problem |
| Brand lost IS (rank) | 6% | 29% | Something outbid you: that something is you |
| Brand campaign clicks | 1,180 | 790 | A third of brand clicks moved |
| Brand campaign average CPC | £0.42 | £0.61 | The cheap clicks went first |
| PMax Search channel CPA | n/a | £9 | Suspiciously low against other channels |
| PMax YouTube channel CPA | n/a | £84 | Closer to true cold-traffic cost |
| PMax search terms containing brand | n/a | 61% of Search conversions | Confirms it |
Read in isolation, the PMax campaign in that table reports a blended CPA of maybe £30 and looks like a success, which is the same way a last-click attribution report flatters whichever click came last. Read against the brand campaign, around 400 clicks a month have moved from a £0.42 campaign into one whose Search channel is paying more for the same person. The brand campaign's CPC has risen 45% because it is now bidding against its own account.
Jonathan Kagan at Amsive published a test in January 2024 where a client paused PMax for a period. Brand search impressions rose 67%, brand clicks rose 92% and brand search CPC rose by more than 60% at the same time, which he put down to PMax having been bidding on the same queries.
His conclusion was that PMax was competing with paid search, and that he could live with it only if overall search results improved.
The fix: brand exclusions first, negatives second
The fix is a brand exclusion list on the PMax campaign, backed by a negative keywords list for anything the brand list misses. Google built the exclusion for exactly this.
Its PMax FAQ says that with a brand list applied, "your Performance Max campaign will avoid serving for branded queries you indicate on Search and Shopping inventory". The same page says the exclusion also catches clear misspellings and foreign scripts, which plain negative keywords do not.
To apply it, follow Google's brand exclusion steps: open the PMax campaign's settings, expand additional settings, click brand exclusions, pick or create a brand list, and save. The same page has a bulk route for several campaigns at once.
Retailers get a checkbox to keep Shopping ads showing on excluded brand searches, which is sensible if your product listings are the thing people want when they search your name.
Three details from Google's brand settings page trip people up:
- Brand lists live at account level and are applied per campaign, so build one list and reuse it.
- If your brand has variants with no shared word (a trading name and a product name, say), each needs its own entry, or the list will not catch the other.
- Google warns that "brand settings limit the reach of a campaign and performance may decrease". Expect the PMax CPA to rise after you exclude brand. That rise is the honest number.
Then add negatives for whatever is left. Google now allows campaign-level negative keywords on PMax (up to 10,000) and account-level negatives (up to 1,000). The catch is in Google's own wording: "negative keywords do not match to close variants". So list the misspellings you saw in step five as well as the brand name.
Optmyzr's study found PMax still served on fuzzy brand matches with exclusions switched on. Its recommendation was to add those misspellings as exact match keywords in the brand Search campaign so the priority rule kicks in.
Finally, feed the brand campaign. The exact match priority only holds when the brand campaign can serve, so a shared budget or a daily cap that runs out at 4pm hands the evening's brand searches to PMax.
What to expect after the change
Expect the PMax campaign to look worse and the account to look better. Brand campaign impression share should climb back towards where it was, its CPC should ease, and PMax's Search channel CPA should move up towards its other channels. Total conversions across the account should hold roughly steady, because the brand searchers were always going to convert somewhere.
If total conversions fall sharply instead, the PMax campaign was doing less than its report claimed, and that is useful to know before the next budget review. Google's channel report cannot tell you that on its own. It can only show you where the conversions sat, which is why the brand campaign's before-and-after is the control.
We run brand search as its own exact match campaign on every account we manage, with brand exclusions on PMax from day one, because untangling it later costs a month of clean data. It is the kind of hygiene a UK Google Ads agency should set up before it reports a single number.
The decision rule we use: if a PMax Search channel CPA is less than half the account's non-brand Search CPA, assume brand leakage until the search terms report proves otherwise.
If you would like us to run this check on your account, our paid media team can do it from read-only access, or get in touch and we will talk through what the channel report is showing you.