Performance Max Branded Search Spend and How to Measure It

Performance Max reports one CPA for the whole campaign. That number blends two very different kinds of traffic. One is people who typed your brand name and were going to buy anyway. The other is new demand that PMax actually found. If you don't split them, you can't tell whether PMax is good at prospecting or just good at standing between your customers and your checkout.

Most advice on this topic stops at "add a brand exclusion". That fixes the symptom without telling you what the brand traffic was worth. It also misses what PMax does with the budget it no longer spends on brand. The method below gives you a number you can defend in a client meeting or a budget review.

Why brand credit makes Performance Max look better than it is

Brand searches convert cheaply and often. When PMax serves on them, those cheap conversions pull the campaign's average CPA down and hide how expensive the non-brand part really is.

Take an illustrative month. A PMax campaign spends $20,000 and reports 400 conversions, a $50 CPA. That looks fine against a $60 target. Now split it:

Without brand, PMax spent $16,000 for 200 conversions. That is an $80 CPA, a third above target. The campaign isn't hitting the target. Your brand is.

This is different from the question of whether PMax is taking queries from your non-brand Search campaigns. That overlap is covered in Performance Max Cannibalizing Search? Compare the Terms. Here the question is narrower: how much of PMax's reported result is brand demand, and what would that demand have cost you elsewhere?

There's a second cost on top of the inflated result. Brand clicks inside PMax are often more expensive than the same clicks in a dedicated brand Search campaign. PMax bids toward a conversion goal across the whole campaign, not toward the cheapest price for a navigational query. So you pay twice: once in money, and once in a misleading CPA that may persuade you to raise the PMax budget.

Step 1: Label every PMax search term

You can't measure brand share from the campaign totals. You need the search terms, with cost, clicks and conversions per term, sorted into buckets you define yourself.

Pull them from the PMax search terms report in the interface, or from campaign_search_term_view through the API. That view covers Search, Shopping and Performance Max. If you haven't worked with it before, What the PMax Search Terms Report Actually Shows explains what is and isn't in it. Use at least 30 days, and 60 to 90 if conversion volume is low.

Google's Insights tab also groups PMax search terms into categories. Those categories are grouped by theme, not by brand versus non-brand. A theme like "trail running shoes" can mix "acme trail shoes" and "best trail shoes". Use them to sanity-check your labels, not as the labels.

Sort each term into one of six buckets. Using a fictional shoe brand, Acme:

1. Pure brand / navigational. "acme", "acme shoes", "acme.com", "acme login", "acme order status". These people are looking for you. Organic results or a $0.50 brand click would have caught almost all of them.
2. Brand plus product. "acme trail 4", "acme running shoes women", "acme trail 4 size 10". Still your demand, but with buying intent on a specific item. For retailers these queries often show Shopping ads inside PMax, which matters later.
3. Brand plus non-buying intent. "acme careers", "acme returns", "acme phone number", "acme warranty claim". These are brand searches that rarely convert at all. They aren't a brand-credit problem. They're plain waste.
4. Competitor brand. "nike pegasus", "brooks ghost". Not your demand. Judge them like generic terms.
5. Generic non-brand. "trail running shoes", "waterproof running shoes wide". This is what you hired PMax to find.
6. No search term. The gap between the campaign total and the sum of the search term rows. It includes Display, YouTube, Gmail and Discover inventory, which report no search terms, and terms hidden by privacy thresholds. See Google Ads Search Terms Hidden by Privacy, and What to Do for how big that hidden share can get.

Labeling tips that save time:

• Build a short list of brand roots, including misspellings: "acme", "acmee", "acm shoes", "acme-shoes". Close variants mean Google will match brand queries you never wrote down. Close Variants in Google Ads and Why Exact Isn't Exact covers why.
• Include product line names that only you sell. If "Trail 4" is uniquely yours, "trail 4 review" is a brand search even without "acme" in it.
• Don't count a generic term as brand just because a converter later visited your homepage. Label the query, not the user.

Write down the total spend, clicks and conversions for buckets 1 and 2 combined. That is your brand bucket. Keep bucket 3 separate. You'll treat it differently.

Step 2: Price the brand bucket at what brand Search pays

Now multiply the brand clicks PMax bought by the CPC your brand Search campaign pays for the same queries. The gap between that figure and what PMax spent is your overpayment, and you can measure it without running any test.

In the example, the brand bucket was $4,000 for 2,000 clicks, a $2.00 average CPC. Suppose your brand Search campaign pays $0.50 a click for "acme" and its variants. The same 2,000 clicks through brand Search would have cost:

2,000 × $0.50 = $1,000

PMax spent $4,000. The overpayment is $3,000 a month, before asking whether those clicks needed buying at all.

Two things to check before you trust that number:

Is your brand Search campaign actually eligible? Google states that when a Search keyword is identical to the query, the Search campaign takes priority over PMax. So if PMax is buying "acme" in volume, ask why. Common reasons: the brand campaign is limited by budget, the variant isn't in its keyword list, or the brand keyword is paused in some ad groups. Check the brand campaign's impression share lost to budget before blaming PMax.

Are the CPCs comparable? Pure navigational terms ("acme login") should be compared with the brand campaign's CPC on the same terms, not the brand campaign's average. Brand-plus-product terms that serve Shopping ads in PMax can't be priced against text-ad brand clicks one for one. Price them separately, or leave them out of this step if you have no Shopping campaign to compare against.

This step answers "how much extra am I paying for brand clicks?". It doesn't answer "how many of those conversions would I have got anyway?". For that you need the test.

Step 3: Run a four-week before/after test

Exclude brand from PMax for four weeks, then compare the whole account's conversions with the four weeks before. Measure the total, not PMax alone. PMax's own numbers will fall by design, and that fall tells you nothing on its own.

Set it up

• Pick the exclusion method. Either a brand exclusion list at campaign level, or campaign-level negative keywords. PMax accepts up to 10,000 campaign-level negatives but does not accept shared negative keyword lists, so you can't just attach your existing brand list. Use exact and phrase match for brand negatives. Broad negatives behave unpredictably across match types. Performance Max Brand Exclusions and What They Miss compares the two approaches.
• Make the brand Search campaign ready to absorb the traffic. Raise its budget so it isn't capped. Add every brand variant from your Step 1 list as a keyword. If it loses impression share to budget during the test, the test is broken.
• Fix the comparison windows. Four weeks before, four weeks after, same weekdays. Avoid promotions, launches and holidays in either window.
• Leave conversion lag room. Wait at least one full conversion window after the test ends before reading the results, or the "after" period will look worse than it was.
• Record a demand control. Note total brand search impressions across all campaigns, plus the brand query trend in Google Trends. If brand demand rose or fell 20% between windows, your result is mostly about demand, not PMax.

After applying the exclusion, check the PMax search terms again within a week. Some brand variants usually slip through. Add them as negatives.

Read the numbers

Continuing the example:

Here is how to read it:

• PMax lost its 200 brand conversions but gained 40 non-brand ones, so it fell from 400 to 240.
• Brand Search gained 150 conversions for $750 extra. That's a $5 CPA on recaptured demand.
• 150 of the 200 brand conversions came back through Search. That is 75% of PMax's brand credit that was demand you would have captured cheaply anyway.
• The other 50 didn't come back as paid conversions. Some of those people probably clicked your organic listing. Check organic brand conversions in your analytics for the same windows before you decide they were lost.
• Total conversions went from 700 to 690. A difference of 10 over four weeks is likely within normal variation.

Note the spend line. PMax still spent $20,000. It didn't save the $4,000 it used to spend on brand. It moved that money into non-brand traffic and got 40 more conversions for it. That's a marginal CPA of $4,000 ÷ 40 = $100.

That number is what the test is really for. You now know:

• PMax's real non-brand CPA at the old spend level: $80.
• What the last $4,000 of PMax budget buys when brand is removed: $100 per conversion.
• What brand demand costs through Search: $5 per conversion.

If your target CPA is $60, the case for cutting PMax back to around $16,000 is now clear, and you have the numbers to show for it.

What the result tells you to do

The test usually ends in one of three places. Each points to a different decision.

Brand Search recovers most of the lost brand conversions and the total holds. This is the example above. PMax's brand credit was mostly captured demand. Keep the exclusion. Then set PMax's budget against its real non-brand CPA, not the blended one. If that CPA is above target, cut budget or work on the non-brand side: asset groups, listing groups, search themes and generic waste. CPA Too High in Google Ads? Check Search Terms First is a good place to start on that.

The total drops noticeably and brand Search doesn't recover the difference. First rule out the boring causes: a capped brand budget, missing variants, a demand dip in your control. If those are clean, PMax was reaching brand searchers your Search campaign couldn't. The usual reason in retail is brand-plus-product queries showing Shopping ads. A text ad for "acme trail 4 size 10" competes poorly with a product image and price. Options: remove only pure navigational brand terms from PMax and leave brand-plus-product in, or run a Standard Shopping campaign for brand queries. Standard Shopping accepts negative keywords and shared lists like Search, so you can control it more tightly.

Very little changes either way. PMax wasn't serving much on brand in the first place, or brand Search was already winning those auctions. The Step 2 overpayment figure will confirm it. Leave the setup alone and spend your time elsewhere.

Whatever the outcome, you can reverse it. Remove the exclusion or the negatives and PMax is eligible for brand again. How to Undo Negative Keywords in Google Ads covers doing it cleanly.

Why "just add a brand exclusion" is incomplete

The common advice on this topic, excluding brand from PMax and running a brand Search campaign, is usually right as an end state. It is incomplete as a method, for four reasons.

It doesn't measure anything. You exclude brand, PMax's CPA rises, and someone asks whether PMax broke. Without a before/after on total conversions and a demand control, you can't answer.

It assumes the budget comes back. It doesn't. PMax spends what you give it. Excluding brand turns $4,000 of cheap brand clicks into $4,000 of non-brand traffic at whatever the marginal price is. If you don't adjust the budget afterwards, you've changed what you're buying, not how much you spend.

It ignores the Shopping side. For e-commerce, a lot of brand-plus-product demand shows up as Shopping ads inside PMax. Blanket exclusion can remove you from those results unless something else covers them.

It treats audience signals as the fix. Some advice says to remove customer lists from signals so PMax stops "targeting" existing customers. Signals are hints, not restrictions. Removing them doesn't stop PMax from serving on brand queries. Performance Max Audience Signals Are Not Targeting explains why. The query-level control is the exclusion or the negative keyword, not the signal.

There's also a limit no method gets around. The "no search term" bucket can't be labeled. If 25% of PMax spend sits in Display, YouTube and hidden terms, you can't tell how much of it reached people who already knew your brand. Your measurement covers the search side only. Say so when you report it.

Brand exclusions and negatives also don't always behave as expected inside PMax. If brand terms keep appearing after you've excluded them, Performance Max Negative Keywords Not Working? Here's Why goes through the usual causes.

Where PPCleaner fits, and where it doesn't

PPCleaner won't make the brand decision for you. It never blocks a term that converted in the last 90 days, and converting brand searches are exactly that. Whether to move them from PMax to Search is a budget and attribution decision. It belongs to you, informed by the test above.

What it does handle is bucket 3: brand searches with no buying intent. "acme careers", "acme returns", "acme customer service phone number", "is acme legit reddit". These terms carry your brand name, so they look safe in a quick scan, and they can quietly eat a share of the brand bucket. PPCleaner reads PMax search terms every night, judges which ones were never going to convert, and adds them as campaign-level negatives within PMax's 10,000 limit. It doesn't touch budgets, bids, ads or keywords. Every block can be undone in one click for 90 days.

It also keeps your Step 1 labels honest over time. Once the generic and junk brand terms are cleaned out each night, what's left in PMax's search terms is closer to the real split between brand and new demand, and your next measurement starts from cleaner data.