Performance Max Placement Exclusions, Honestly

Can you exclude placements in Performance Max? Not inside the campaign

No. Performance Max has no campaign-level placement exclusion setting. You cannot open a PMax campaign, go to Content, and add a mobile app or a YouTube channel to a campaign-specific exclusion list the way you can in a Display or Video campaign.

What you do have is three controls, all of which sit above or beside the campaign:

• Account-level placement exclusions. One list, applied to everything in the account. Add gamesforkids.example there and it is excluded from your PMax campaigns, your Display campaigns, your Video campaigns, all of them.
• Content suitability settings. Inventory type (expanded, standard, limited), plus the content-type and sensitive-content category exclusions. These are blunt category switches, not a list of URLs.
• The campaign's own inventory setting. PMax lets you pick inventory type per campaign, which is the one placement-adjacent control that is genuinely campaign-scoped.

That is the whole toolkit. Everything else you read about "excluding placements in PMax" is either describing the account-level list, describing a script that writes to the account-level list, or describing Display and Video campaigns and quietly hoping you do not notice the difference.

The practical consequence is worth stating plainly: if a game app is wasting money in your lead-gen PMax campaign but converting fine in a separate Display remarketing campaign, you cannot have it both ways. The account-level list is all or nothing.

What the PMax placement report actually gives you

You get impressions, and almost nothing else. The placement report for Performance Max — found under the campaign's Insights or in the account-level Content > Where ads showed view — lists the websites, apps and YouTube channels your ads appeared on, with impression counts.

It does not give you cost per placement. It does not give you conversions per placement. It does not give you clicks per placement in a form you can trust for ROI decisions. So the usual analytical move — sort by spend, find the placements that took $400 and returned nothing, exclude those — is not available. You are sorting by impressions and guessing at the money behind them.

This is the part the "placement exclusion" posts tend to skip. They show you a filtered list of mobile game apps and tell you to exclude them. Fine. But you are excluding based on a name that looks like junk, not on evidence that it cost you anything. A placement with 90,000 impressions might have cost $3. A placement with 2,000 impressions on a high-intent site might have cost $200. The report will not tell you which is which.

Two honest conclusions follow. First, placement exclusions in PMax are a hygiene exercise, not an optimization with a measurable payback. Second, when you later look at how to calculate wasted ad spend honestly, you will not be able to attribute a dollar figure to the placements you excluded. Anyone who hands you one has invented it.

Where the report is genuinely useful

Brand safety. If your ads are showing on a channel you would be embarrassed to screenshot, the impression-only report is enough to act. You do not need cost data to decide you do not want to be there.

It is also useful as a diagnostic. A PMax campaign that is supposed to be driving ecommerce sales but shows 80% of its impressions on mobile game apps is telling you something about the campaign's asset mix, its audience signals, and probably its conversion tracking. That is a reason to restructure, not just to exclude.

The account-level exclusion list and its real limits

The account-level placement exclusion list is where every PMax placement block lands. It accepts websites, YouTube channels, YouTube videos, mobile apps and mobile app categories. It applies retroactively from the moment you add an entry — it does not clean up past spend, obviously, but it takes effect without needing the campaign to restart learning.

Three limits matter.

It is account-wide, with no scoping. Already said, worth repeating because it changes how aggressive you should be. If you manage one account with one PMax campaign, exclude freely. If you manage an account with PMax, Display retargeting and a YouTube campaign all running, every exclusion is a decision for all three. This is the same tension that shows up in managing negative keywords across client accounts: a control that is convenient when you have one campaign becomes a blunt instrument when you have twelve.

App exclusions are a game of whack-a-mole. There are millions of apps in the inventory. Excluding 200 of them removes 200 of them. The usual answer is to exclude the whole mobileappcategory::69500 bucket — all apps — but that is a real trade-off: some accounts, particularly ecommerce with strong remarketing, do get conversions from in-app inventory. Excluding all apps is a decision with a cost, not a free win. Test it with a two-week before/after on cost per conversion before you make it permanent.

YouTube channel exclusions do not catch everything. You exclude a channel; the same content gets uploaded to a second channel. Video exclusions are even narrower. The brand-safety case for doing it anyway is strong. The efficiency case is weak.

The exclusion that actually moves money in PMax

Negative keywords. Unlike placements, PMax accepts up to 10,000 campaign-level negative keywords, and the search side of PMax is where the spend concentrates in most accounts we audit.

The distinction matters because the two problems feel similar and are not. A junk placement serves an impression you probably did not pay much for. A junk search term serves a click you definitely paid for — often at your highest CPC, because PMax inherits your Shopping and Search inventory. When we audit accounts, the search terms side is almost always where the recoverable waste sits, and we see it in a wide range: in some accounts a few percent of spend, in others a quarter of it, depending on how broad the product feed and the asset group signals are.

A few things to know about PMax negatives specifically:

• They are campaign-level only. Performance Max does not accept shared negative keyword lists. If you run six PMax campaigns, the same term has to be added six times, once per campaign. This is the single most annoying structural fact about PMax negatives and it is covered in detail in negative keywords in Performance Max and the 10,000 cap.
• They coexist with brand exclusions, which are a separate and much weaker mechanism — see Performance Max brand exclusions and what they miss.
• Search themes are positive signals, not negatives. People conflate the two constantly; search themes vs negative keywords in Performance Max explains why adding a theme never removes a query.

If you have an hour to spend on a PMax campaign and you are choosing between building a placement exclusion list and working through the search terms, work through the search terms. The placement list is brand safety. The search terms are money.

Reading the PMax search terms report before you exclude anything

Performance Max search terms appear in campaign_search_term_view, the same table that serves Search and Shopping. Display, Video and App campaigns report no search terms at all — which is itself a clue about where the placement problem lives.

The PMax view is not the same as the Search view. It is grouped, thresholded and incomplete; a meaningful share of spend sits in rows you cannot see, and the categories Google shows you in the Insights page are not search terms at all. What the PMax search terms report actually shows goes through this row by row. The short version: treat what you can see as a sample, not a census, and expect the real term list to be longer and worse than the visible one.

Two practical habits:

Export at the campaign level, not the account level. Because PMax negatives are campaign-scoped, an account-level export will mislead you about which campaign needs which negative.

Check more than once a month. PMax reshuffles its query mix as the feed and assets change. A term that did not exist in March takes $180 in April. How often should you check search terms in Google Ads argues for a cadence rather than a quarterly panic.

Terms worth blocking that look harmless

The obvious junk — free, cheap, diy, how to make — gets blocked early because everybody blocks it. The expensive terms are the ones that look commercial:

• Job and salary searches. project manager salary, warehouse jobs near me. A careers page that ranks well will absorb real clicks from PMax, and those clicks cost full price. Job seekers clicking Google Ads lists the pattern.
• Support and warranty searches. reset my [product] password, warranty claim. Existing customers, paid for twice.
• Competitor names. Sometimes worth keeping, sometimes a steady drain. Competitor searches in Google Ads, block or keep is the decision tree.
• Wrong-product near-misses. The ecommerce classic: you sell dog beds, PMax buys dog bed frame. E-commerce Google Ads search terms that drain budget has more of these.

A placement and exclusion routine that does not take a day

Half an hour a month on placements, a shorter check more often on search terms. In that order of effort, because that is the order of payoff.

Monthly, for placements (20 minutes):

1. Open Content > Where ads showed at the account level. Set the date range to the last 30 days.
2. Sort by impressions, descending. Look at the top 100 rows only. The long tail is not worth your time.
3. Flag anything that is a children's game, a content farm, a made-for-advertising site, or a YouTube channel you would not want on a slide.
4. Add those to the account-level exclusion list. Note the date you did it.
5. If more than half of your top rows are apps, consider the all-apps category exclusion — and note the cost per conversion for the two weeks before, so you can compare.

Weekly or fortnightly, for search terms (15 minutes):

1. Export the search terms for each PMax campaign separately.
2. Sort by cost. Read down until the rows stop costing anything meaningful.
3. Add negatives at the campaign level. Use phrase and exact. Broad negatives behave unpredictably across match types and will block more than you intended — negative keyword match types don't work like positives is the full explanation.
4. Never block a term that converted. Checking a 90-day conversion window before adding a negative is the single rule that prevents most self-inflicted damage.

If step 4 sounds obvious, it is the step people skip when they bulk-paste a list from a competitor's blog post. Negative keywords hurting performance? Six signs describes what it looks like afterwards.

What a script can and cannot do

The placement excluder scripts floating around are real and they work. They read the placement report, apply a filter — usually "exclude any mobile app" or "exclude anything matching this regex" — and write to the account-level exclusion list.

What they cannot do is solve the scoping problem, because the API does not expose a campaign-level placement exclusion for PMax. The script writes where you would write: account level. It saves you clicking, not deciding. And because the placement report has no cost column, a script cannot make a spend-based judgement either. It is a name-matching robot, and it should be treated as one.

Where placement exclusions fit against everything else

Placement exclusions are the fourth or fifth thing to fix in a PMax campaign that is spending without converting, not the first. The order that tends to find money fastest:

1. Conversion tracking. If PMax is optimizing to a bad signal, nothing downstream matters. Google Ads spending money, no conversions? Check this order runs through this.
2. Search terms and negatives. Highest-cost clicks, best control, campaign-scoped.
3. Asset group and feed structure. What you feed PMax determines what it buys. A single asset group covering 400 unrelated products will match widely no matter how many negatives you add.
4. Inventory type and content suitability. One setting change, applied campaign-wide.
5. Placement exclusions. Brand safety and hygiene.

PMax is not uniquely wasteful — Performance Max vs Search: which wastes more found the answer is less obvious than the complaints suggest — but it is uniquely opaque. The controls that exist are coarse and the reporting that would let you use them precisely is withheld. Accepting that saves a lot of time. Where Performance Max wasted spend hides covers the rest of the hiding places.

One last honest note about what PPCleaner does here. We work on search terms: we read what each campaign paid for in campaign_search_term_view, judge which terms were never going to convert, and add them as campaign-level negatives every night — for Search, Shopping and Performance Max alike. We never touch budgets, bids, ads or keywords, and we never block a term that converted in the last 90 days. Every block is reversible in one click for 90 days, which is the point of how to undo negative keywords in Google Ads.

We do not manage placement exclusions. There is no campaign-level control to manage, the report carries no cost data to act on, and we would rather say that than sell you an automation for a decision that comes down to whether a YouTube channel looks respectable. Do that part by hand, once a month, for twenty minutes.