If you've noticed Performance Max spending on brand terms, you're not imagining it. PMax is Google's all-in-one campaign type: one campaign, one budget, ads across Search, Shopping, YouTube, Display, Discover, Gmail and Maps, with Google's automation choosing where each dollar goes. For a lot of online stores it's now the main way products show up in Shopping results.
Store owners usually spot the problem the same way. A good chunk of PMax's conversions come from people who searched the store's own name. They were already looking for you. PMax shows them a Shopping ad, they click, they buy, and the campaign reports a strong return on ad spend. Then you check total store revenue, and it hasn't moved the way that ROAS suggests.
Why it happens
Nothing sinister. PMax is told to hit a conversion or ROAS goal, and brand searchers are the easiest conversions going: high intent, cheap clicks, quick checkouts. A system optimizing for reported conversions will find them, and keep finding them, unless you tell it not to.
Two things make it worse:
- No brand Search campaign. If you don't run one, PMax gets your brand searches to itself.
- A feed that doesn't help Google tell your products apart. Thin titles and missing identifiers give PMax less to match non-brand searches with. For a store that sells its own label, the store name ends up being the strongest signal around.
First, confirm it's happening
Check before you change anything.
- Search terms inside PMax. Google has expanded search-term reporting for PMax over the past couple of years. Look at the campaign's Insights and, where your account has it, the search terms report. Search for your brand name and its common misspellings.
- Your brand Search campaign, if you have one. If its impressions dropped when PMax launched, PMax may be winning those auctions instead.
- Store revenue against platform revenue. Compare total Shopify, WooCommerce or BigCommerce revenue for a period with what Google Ads claims for the same period. Blended MER (total revenue divided by total ad spend) is a blunt cross-check, but an honest one.
Where these reports sit in the interface has changed several times. If something isn't where you expect it, check Google's current Performance Max help pages.
Campaign-level fixes come first
These usually do more than the feed, so start here.
- Brand exclusions. PMax lets you apply brand lists as exclusions, which stops it serving on searches for those brands across Search and Shopping inventory. Exclude your own brand.
- A dedicated brand Search campaign. Exact and phrase match on your name. Google's documented rule is that when a Search keyword is identical to the search term, the Search campaign is preferred over PMax. Check the current wording, because Google has adjusted it before.
- Negative keywords. Google has been adding negative keyword options to PMax in stages, account-level lists first, then campaign-level negatives. What you can use depends on what has reached your account.
- New customer acquisition. If new buyers are the point, PMax can bid higher for them or bid only for them. It learns who's "new" from your customer lists or conversion tags, so it's only as good as that data.
Expect PMax's reported ROAS to fall once brand is excluded. That's fine. The brand revenue didn't vanish. It moved to the Search campaign, where you can see it and pay less for it.
The feed checklist
This is where the feed earns its place. A clean feed gives PMax better material for matching products to non-brand searches, and gives you the structure to split campaigns by what actually makes you money. Field names below follow Google's product data specification; check the current spec before bulk edits.
| Feed field | What to fix | Why it matters here |
|---|---|---|
title | Lead with what the product is: type, key attribute, size or color. Put your brand later, unless shoppers search brand plus product. | Titles strongly shape which searches a product matches. A title that's mostly your store name invites brand queries. |
gtin | Add correct GTINs where a manufacturer has assigned them. Never make them up. If a product truly has none, set identifier_exists to "no" as the spec allows. | Google asks for GTINs wherever they exist. Missing ones can limit how your products show. |
brand | Use the real product brand, spelled the same way every time. | Required for most products under Google's spec, and keeps grouping clean. |
product_type | Your own category path, e.g. "Kitchen > Knives > Chef's knives". | Gives you tidy listing groups and readable reports. |
google_product_category | The most specific category that fits. | Helps Google understand the product instead of guessing. |
custom_label_0 to custom_label_4 | One label for margin band (high, mid, low), one for best-sellers, one for seasonality. | Lets you split campaigns or listing groups so each margin band gets its own budget and target. |
Split by margin, not just category
Custom labels pay off here. Say you sell kitchen equipment, and some lines make about 50% margin while others make about 15%. (Hypothetical numbers.) One PMax campaign with one ROAS target treats them the same and will spend wherever conversions come easiest, low-margin products included.
Two or three campaigns split by margin label let you:
- Accept a lower ROAS target where margin is high, and demand a higher one where it's thin.
- Move budget toward products that actually leave you profit.
- Read brand and non-brand results with less noise.
Don't split into a dozen. Automation generally needs a decent volume of conversions per campaign, and every split thins that out.
What not to do
- Don't switch PMax off in a panic. Fix the inputs and measure against store revenue first.
- Don't judge the change after a week. Bidding needs time to settle after a structural change.
- Don't stuff other brands' names into your titles. Google's spec asks for titles that accurately describe the product, and keyword-stuffed titles risk disapprovals.
None of this guarantees higher profit. It does show you what your non-brand spend is really doing, which is the number most PMax reports hide.
How we handle it
On e-commerce growth work, we report brand and non-brand separately and check results against your store's own revenue, not just Google's figures. If you want an outside look at your current account first, the ads account audit is $690 for one account, delivered within 5 business days of access.
