Performance Max vs Search: Which Wastes More?

Every comparison of these two campaign types ends up in the same argument: control versus automation, incrementality, whether Google is quietly harvesting your brand traffic and billing you for it. Those are real questions. They are also unresolvable from outside your account, which is why the pages ranking for this query all end with "it depends on your goals."

So here is a narrower question, one you can actually answer with data you already have. Per dollar spent, which campaign type loses more money to searches that were never going to convert — and which of those losses can you actually stop?

Search wastes more per click; Performance Max wastes more per dollar you can't see

A broad match Search campaign is the fastest way to buy irrelevant traffic in Google Ads. Point it at "commercial plumbing contractor" and within a month you have paid for "plumber salary", "plumbing courses near me", "how to fix a leaking tap youtube", and at least one query containing a competitor's name. That is not a bug in your setup. That is broad match doing what it is built to do, and it is worth asking whether it still earns its keep.

The important part: you will see all of it. Every one of those terms appears in the search terms report with impressions, clicks, cost and conversions attached. The waste is loud. It is annoying. It is also completely actionable — you read the report, you add negatives, the term stops costing you money, and next week the report is shorter.

Performance Max does not work like that. It spends across Search, Shopping, Display, YouTube, Gmail, Discover and Maps from one budget, and it reports search terms only for the search-like part of that. The rest — the Display and Video impressions, the Gmail placements — produces no search term data at all, because there was no search. So the first structural difference is not about quality of traffic. It is about how much of the spend is even describable as a search term in the first place.

In the accounts we audit, the share of a Performance Max budget that ends up attached to a visible, named search term is usually a minority of the total. We are not going to give you a percentage, because it swings wildly with feed quality, asset group setup and how much of the budget Google decides belongs on YouTube this week. But the direction is consistent: in a Search campaign, close to all the spend has a query next to it. In Performance Max, a lot of it does not.

That matters for waste hunting because you can only negate what you can name.

What each campaign type actually shows you

Search shows you every search term above a privacy threshold, with cost, clicks and conversions. Performance Max shows you fewer terms, grouped, without cost per term — and both reports hide low-volume queries entirely.

Let's be precise about the gap, because "PMax has no search term report" is a three-year-old complaint that is no longer true.

Both campaign types report into campaign_search_term_view. Search, Shopping and Performance Max all appear there. Display, Video and App campaigns report nothing — there are no search terms to report. So if you run a Display campaign, waste analysis by search term is not partially hard, it is impossible, and nothing in this article applies to it.

What Search gives you per term:

• impressions, clicks, cost, conversions, conversion value
• the keyword that triggered it
• the match type of that keyword, via match source
• the campaign and ad group

What Performance Max gives you per term is thinner. You get terms grouped into categories, you get the search category label, and you get far less of the per-term financial detail. We have written the full inventory of what is and is not in there — see what the PMax search terms report actually shows — but the summary for waste purposes is that you can often tell that a bad query happened and rarely tell how much it cost.

And both reports apply a volume threshold. Terms searched by only a handful of people never appear, in either campaign type. This is the part that skews every "which wastes more" estimate ever published, including any you would make yourself: the search terms report hides things, and it hides more from PMax than from Search. The long tail of tiny, weird, expensive one-off queries is invisible in both, but in Search the tail is a smaller fraction of total spend because your keywords constrain the range of what can match.

A worked example of the visibility gap

Say you spend $10,000 a month, split $5,000 to Search and $5,000 to Performance Max.

On the Search side, you open the search terms report for the month. It accounts for, say, $4,600 of the $5,000 — the rest is the sub-threshold tail. You sort by cost, descending. You find $600 sitting on terms with zero conversions and no plausible intent: informational queries, job searches, DIY how-tos. You block them. Next month, most of that $600 goes to terms that can convert or stays in your pocket.

On the Performance Max side, you open the equivalent report. It names terms, but the cost column is not there per term in the way you want it. You can see that "free" and "how to" and a competitor's brand name appear in your category list. You cannot see that they cost $600. You can infer, from conversions and from the shape of the categories, that something is wrong. You cannot put a figure on it without guessing — and guessing is exactly what produces the inflated wasted-spend numbers you see in sales decks. If you want to do this properly, calculate wasted spend honestly and accept a range instead of a figure.

So: same $600 of waste, two very different cleanup experiences. Search waste is a task. PMax waste is an investigation.

Where Performance Max hides its waste

The three usual hiding places are brand traffic Google bills you for, near-brand and competitor queries pulled in by your feed and landing pages, and the non-search inventory that produces no search terms at all.

Brand first, because it is the biggest single line item in most PMax waste audits. If your brand is at all searchable, Performance Max will happily serve against it and take credit for conversions that would have happened anyway. Google's answer is brand exclusions, which help, and which do not cover everything — misspellings, brand-plus-modifier combinations and near-brand competitor forms slip past them regularly.

Second, feed-driven and page-driven matching. Performance Max reads your product feed, your final URLs and your page content, and matches queries from all of it. That means it can bid on product attributes you never intended to advertise, and on categories that happen to appear in your content. A Search campaign cannot do this — if the keyword is not in the account, it cannot match, broad match excesses aside. The full list of places this money goes is in where Performance Max wasted spend hides.

Third, and this is the honest limitation: you cannot negate your way out of Display and YouTube placements. Negative keywords do nothing there. There is no keyword. If your PMax campaign is spending 40% of its budget on video views, negative keywords address the other 60% and no more. Anyone selling you a negative keyword tool as a complete fix for Performance Max is overselling, and that includes us — which is why we say the tool cleans the search-term-visible portion of the spend, not the campaign.

Where Search wastes money that Performance Max does not

Search has one failure mode Performance Max largely avoids: it can be left running on a keyword set that nobody audits, matching broadly forever, with no automated correction.

Performance Max is bad at telling you what it bought. But it is genuinely good at the narrow thing it does — pattern-matching to conversions — and it will drift away from a query pattern that never converts, on its own, given enough volume. Not reliably, not quickly, and not if your conversion signal is thin. But it does happen.

A broad match Search keyword does not learn. It will match "plumber salary" this month and next month and in two years, until a human adds a negative. The bidding algorithm will bid less on it over time if Smart Bidding has enough data, but it will keep entering the auction. Cheap clicks that never convert are still clicks you paid for.

This is why the "Search gives you control" argument cuts both ways. Control is only an advantage if you exercise it. An account where nobody has looked at the search terms report in six months has a Search waste problem that is worse, in absolute dollars, than its PMax problem — because nothing in Search self-corrects and nothing in Search is embarrassed about matching badly. How often you check search terms matters more for Search than for PMax, for exactly this reason.

There is a second Search-specific leak: why Google shows you irrelevant search terms in the first place is partly close variants. Exact match "emergency plumber london" will match things you would not recognise as the same query. Phrase match will do worse. And you cannot turn close variants off. So even a tightly built exact-match account bleeds a little, permanently.

The cleanup tools are not equal, and the gap is smaller than people think

Search accepts negative keywords at account, campaign and ad group level, plus shared lists. Performance Max accepts up to 10,000 campaign-level negatives and does not accept shared lists. That is the whole difference — and it is less severe than the internet suggests.

Here is what that means in practice.

For a Search account, you build one or two shared negative lists — a "never relevant" list with job, salary, DIY, free and course terms, a competitor list if you have decided to block competitor searches rather than keep them — and apply them to every campaign. Add a term once, it applies everywhere. Standard Shopping works the same way: negatives and shared lists, both supported, though Shopping negatives behave differently in how they interact with product matching.

For Performance Max, no shared lists. Every negative has to be added at campaign level, on each campaign, one campaign at a time. If you have five PMax campaigns and 400 negatives, that is 2,000 entries to maintain, and when you remove a term you remove it five times. The 10,000 cap sounds generous until you are managing it by hand — the practical implications of the 10,000 cap bite long before you hit the number.

The cap itself is rarely the binding constraint. The maintenance burden is. This is a plumbing problem, not a limits problem, and it is the single strongest argument for automating the job rather than doing it in the interface.

Two more things to get right in both campaign types:

Match types. Broad, phrase and exact are your three options for negatives, and negative match types do not work like positive ones. Broad negatives in particular behave unpredictably across match types — a broad negative can block more than you expect and, in some configurations, less. Phrase and exact negatives are more predictable. In Performance Max, where you cannot see the full consequence of a block, predictability is worth more than reach.

Search themes are not negatives. People keep trying to steer PMax away from bad traffic by adjusting search themes. That is not what they do — search themes are inputs, negatives are exclusions, and adding a theme never removes a query.

A decision you can actually make from waste data

Do not pick a campaign type based on which one wastes less. Pick based on which waste you are equipped to manage, then budget the labour honestly.

Three practical positions, depending on what your account looks like.

If you have one person and twenty campaigns. Search waste will be your bigger number, because nobody is reading the reports. Fix the process before you fix the campaign mix. An hour a month on search terms across your Search campaigns will recover more than any structural change. Start with the one-hour audit checklist.

If you already clean Search weekly and your PMax spend is growing. Your marginal waste is now in Performance Max, and specifically in brand and near-brand queries. Set brand exclusions, add your competitor and non-buyer negatives at campaign level on every PMax campaign, and accept that the Display and Video portion of the budget is beyond the reach of negative keywords. If that portion is large and unproductive, the answer is a campaign structure change, not a negative keyword.

If you are spending under $2,000 a month. Neither report has enough volume to be useful weekly. Both will be dominated by sub-threshold terms you cannot see. Check monthly, block obvious patterns, and do not over-engineer. Adding 3,000 negatives to a $2,000 account is how you get negatives that hurt performance.

One thing that will not help in any of these cases: optimization score. It does not know which of your search terms were pointless, and it will happily recommend removing negatives that are doing their job. What it measures and what it ignores are two different lists.

Honest accounting: what automating this does and does not fix

Automation solves the maintenance problem in both campaign types. It does not solve the visibility problem in Performance Max, because no tool can see data Google does not publish.

PPCleaner reads campaign_search_term_view every night for Search, Shopping and Performance Max, judges which terms were never going to convert, and adds them as negatives — campaign-level for PMax, because that is the only option, and one entry per campaign, which is the part you would otherwise do by hand. It never blocks a term that converted in the last 90 days, and every block can be undone in one click for 90 days. It does not touch budgets, bids, ads or keywords.

What it does not do, and cannot:

• It cannot see PMax search terms below Google's reporting threshold. Neither can you.
• It cannot attribute cost to individual PMax search terms with the precision available in Search, because that data is not in the API at that grain.
• It cannot do anything about Display, YouTube, Gmail or Discover spend inside a Performance Max campaign. No search term, no negative keyword, no fix.
• It cannot tell you whether Performance Max is incremental. That is a holdout test, not a search term question.

So the answer to "which wastes more" is uncomfortable but useful: Search wastes more dollars in accounts nobody watches, and Performance Max wastes more dollars you will never be able to itemise. The first problem is a labour problem with a known solution. The second is a transparency problem you can only partially work around.

If you are choosing where to spend your next hour, spend it on whichever campaign type you last looked at longest ago. If you want the general method rather than the comparison, how to find wasted ad spend in Google Ads walks the whole report, and when nothing is converting at all there is an order to check things in that starts before search terms.