If you already run a profitable store, the mechanics below take ten minutes. The reason they matter takes longer to see — so start there.
Why branded traffic quietly distorts your Performance Max numbers
Performance Max chases the cheapest conversions in the account, and branded queries are the cheapest of all because the intent is pre-built. Left unchecked, PMax absorbs orders that your organic listings, direct traffic, or brand Search campaign were already winning. Its reported ROAS climbs while your total business orders stay flat.
Practitioner analyses put the leak at roughly eight to fifteen percent of PMax budget flowing to branded queries in unprotected accounts, inflating apparent ROAS by about fifteen to thirty percent (GrowthSpree, ClickTrends). Treat those as field estimates, not laws. The agency smec frames the cost mechanism plainly: you end up paying "$1.50 for a click that you could have bought for $0.20" (smec).
A worked example
Say your PMax campaign spends $4,000 a month and reports $24,000 in conversion value — a clean-looking 6.0x. You pull the search-terms report (PMax exposed search terms starting in 2025, per The Media Image) and find $500 of that spend went to your own brand name, returning $6,000 in value.
Strip the brand out and the honest picture is $3,500 spend → $18,000 value = 5.14x on net-new demand. Worse, most of that $6,000 was coming anyway through organic and direct, so the incremental return on the $500 is close to zero. The 6.0x headline was borrowing from a channel you already owned for free.
This is why the brief matters at the profit layer, not the reporting layer. At a thirty percent contribution margin — common for print-on-demand — your break-even is 1 ÷ 0.30 = 3.33x, so a flattered 6.0x can hide non-brand performance that is dangerously close to the line. The Google Ads economics guide walks the full break-even math if you want the derivation.
How to exclude brand from Performance Max, step by step
The flow has three parts: build a list, apply it, then rebuild the cheap defense you just removed from PMax.
Step 1 — Build a brand list
Go to Tools & Settings → Shared library → Brand lists and create a new list. Add your own brand from Google's indexed brand database; the index already understands variants and misspellings, so you rarely need to enumerate every typo yourself.
If your brand is not in the database, submit it for inclusion. Expect a review window of several weeks before it becomes selectable, so do this early rather than the day you launch a campaign.
Step 2 — Apply the exclusion to each PMax campaign
Open the campaign, go to Settings → Brand exclusions, and apply your list. There is a separate checkbox labeled something like "Allow Shopping ads on searches that mention excluded brands" — leave it unchecked if you want a true brand block, since Shopping ads can otherwise keep serving on branded queries even with the exclusion on.
Brand exclusions apply per campaign, so repeat for every non-brand PMax campaign in the account. If you run several, apply the list in bulk rather than one at a time.
Step 3 — Rebuild a dedicated brand Search campaign
Excluding brand from PMax does not mean abandoning those queries — it means defending them deliberately and cheaply. Stand up a single-keyword or tightly-themed brand Search campaign so you still capture brand intent at a low CPC instead of ceding it to competitors who bid on your name.
This separation is the whole point: non-brand demand lives in PMax where the bidding system can hunt for it, and brand demand lives in a campaign you can read and control. For a deeper treatment of the same split, see the companion piece on excluding brand from Performance Max.
The one number that tells you it worked
Do not judge the exclusion by PMax's ROAS — that number will drop, because you just removed the cheap conversions propping it up. That drop is the system reporting honestly again, not a regression.
Judge it two ways instead. First, PMax's search-terms report should show branded queries fading out over a cycle or two. Second — the honest test — watch total business orders against total ad spend, not per-campaign ROAS; if orders hold while the blended picture gets more truthful, the exclusion did its job.
Give smart bidding one to two conversion cycles to re-solve after the change, since any significant edit triggers recalibration, and tROAS needs at least fifteen conversions with values in the trailing thirty days to bid stably on Search and Shopping (Google Ads Help). If your non-brand volume sits below that floor after the split, that is a volume problem to solve, not a reason to undo the exclusion.
When a brand-only PMax campaign makes sense
The inverse setup — brand inclusions, where a PMax campaign serves only on your branded terms — is a narrower tool. It is occasionally useful for consolidating brand defense with feed-driven Shopping ads, but for most operating stores a dedicated brand Search campaign is cheaper and far more transparent.
Remember that PMax is mostly a Shopping engine under the hood: smec reports that seventy-four to ninety-seven percent of PMax cost goes to feed-based ads (smec). So "brand-only PMax" is really "brand Shopping plus some display," which is rarely what you want for pure brand defense. Running PMax alongside Shopping and Search is already the norm — Optmyzr's study of thousands of campaigns found most advertisers run them together (Optmyzr).
If you are standardizing this across a larger account or multiple brands, the enterprise control story is covered in the SA360 and Performance Max breakdown, and the latest control changes are tracked in the November 2025 Performance Max updates.
Where Victor fits
Brand cannibalization is a bookkeeping problem disguised as a performance win, and it is exactly the kind of thing that hides when you only read ROAS. Victor, the AI employee inside PodVector AI, connects to your Google Ads and Shopify data and computes true per-order profit — so a flattered PMax ROAS can't quietly mask a non-brand segment that is barely clearing break-even.
As a full Google Ads operator, Victor can surface the branded-query leak and propose the fix, and every write action is approval-gated — you approve before anything changes in the account. If you want this watched continuously instead of audited quarterly, put Victor on your store. For stores that would rather hand the whole channel to specialists, the Google Shopping ads agency comparison is the next read.
FAQs
Does excluding brand from Performance Max hurt my total sales?
Not if you rebuild a brand Search campaign to catch those queries. You are moving brand demand from an opaque PMax auction to a cheap, controllable one — the orders still get captured, usually at a lower CPC. What drops is PMax's reported ROAS, because it loses the conversions that were inflating it.
How long until the exclusion takes effect?
The setting applies immediately, but smart bidding needs one to two conversion cycles to recalibrate after any significant change (Google Ads Help). Expect branded search terms to fade over that window rather than vanish the same day. Confirm it by checking the PMax search-terms report after a cycle.
Why does my PMax ROAS look worse after I excluded brand?
Because it was never that good — the brand terms were borrowing cheap conversions from channels you already owned. A lower, brand-free ROAS is your true non-brand acquisition economics. Judge the change on total business orders versus total spend, not on the per-campaign number.
Do I still need the "Allow Shopping ads" checkbox off?
Leave it unchecked if you want a complete brand block. With it checked, Shopping ads can keep serving on branded queries even though the exclusion is applied, which partially defeats the purpose. Turn it off unless you have a specific reason to let branded Shopping through.
Can I exclude competitor brands too?
Yes — brand exclusions work on any brand in Google's index, not just your own, so you can keep PMax off competitor names if you do not want to pay for that traffic. Build it into the same list and apply it per campaign. Re-audit the search-terms report periodically, since the query mix shifts as the system re-solves.