Performance Max Search Term Insights Are Not Queries

Performance Max search term insights group many searches into one category and show results for the whole group. A category tells you where your spend is going. It doesn't tell you which searches inside that category are wasting money. A category with a good average can still hide a few searches that never convert. A category with a bad average can contain searches that are about to convert.
Use insights to decide where to look. Then open the actual terms behind the category: Performance Max reports them in campaign_search_term_view, just like Search and Shopping. Choose your negatives from those terms, mostly as phrase or exact match. Never turn a category name into a broad negative.
What search term insights show, and what they hide
Search term insights put your Performance Max searches into categories and subcategories, then report results for each group. They answer one question: what kind of demand is this campaign reaching? They can't answer the question you need for negatives: which exact searches cost money and never converted?
That difference matters more than it seems. A category is a summary that Google builds. Google decides which searches go together, picks the label, and shows you a total. The searches inside can mean quite different things. "Plumbing services" might hold "emergency plumber near me", "plumber open sunday", "plumber salary" and "how to become a plumber". The first two are customers. The last two are job seekers who will never book a call. In the category view, they all add up to one row.
The category view is designed to show conversions and search volume. It isn't designed to show the cost of each query. That works well for planning, such as spotting a product line you hadn't thought about or seeing that brand searches make up a big share of results. It works badly for cleaning up wasted spend, because waste happens one query at a time.
Insights also can't show what Google hides for privacy reasons. Some searches never appear as individual terms, in any report. If your category totals don't match what you can trace to individual terms, read search terms hidden by privacy and what to do before you assume the gap is your fault.
Why a category is not a query
A category average can make waste look healthy, and it can make good traffic look like waste. Both problems come from the same place. You are judging a group, but a negative keyword blocks a specific string of words.
A healthy category with waste inside
Say a plumbing company's Performance Max campaign has a "plumbing services" category with 40 conversions last month. When you pull the terms behind it, you see this:
| Search term | Cost | Conversions |
|---|---|---|
| emergency plumber near me | $1,800 | 34 |
| plumber near me open now | $700 | 6 |
| plumber salary | $300 | 0 |
| plumbing apprenticeship | $200 | 0 |
| Total | $3,000 | 40 |
At the category level, the cost per conversion is $75 ($3,000 ÷ 40). That's a figure most service businesses would be happy with. Nobody looking only at the insights page would flag it. But $500 of that $3,000 went to people looking for a job or a training course. That's one dollar in every six. The category is fine. Two of its queries are not.
If you work only from categories, you leave that $500 alone every month because the group looks good. Smart Bidding then keeps learning from those clicks as if they were normal traffic. We explain why that matters in how junk queries skew your bids.
A weak category with buyers inside
Now take the opposite case. The same advertiser sees a "water heaters" category with zero conversions on $600. It's tempting to block it. The terms behind it are:
| Search term | Cost | Conversions |
|---|---|---|
| tankless water heater install | $250 | 0 |
| water heater leaking | $200 | 0 |
| water heater reviews | $150 | 0 |
| Total | $600 | 0 |
Only one of these is plainly research: "water heater reviews". "Water heater leaking" is someone with a problem right now, which is exactly what a plumber wants. "Tankless water heater install" is a job with a high ticket price. Zero conversions on $250 is a small sample for that kind of job. It isn't proof. The right move is one phrase negative on "reviews" (or "water heater reviews"). Blocking the category is wrong.
These two tables are the whole argument. The category answers "how is this group doing?" A negative keyword needs an answer to "should this exact search ever trigger my ad again?" The two answers often point in opposite directions.
How to read a category to guess the queries inside it
Insights are still useful. Read them as a map of where to dig, not as a list of what to block. Four habits turn a category into a good guess about the queries inside it.
1. Read the label as the most generic version of the query
Google tends to label a category with the shared topic and remove the modifiers. The modifiers are usually where the intent is. "Plumbing services" doesn't tell you whether the word next to it was "near me", "salary", "course" or "cost". When you see a broad label, list the modifiers that would change the intent in your business:
- Job and career words: salary, jobs, hiring, apprenticeship, course, certification
- Do-it-yourself words: how to, diy, tutorial, fix yourself
- Research words: reviews, vs, best, reddit
- Free or cheap words: free, used, second hand, cheap, template
- Wrong product or wrong place: a competitor's product type, a city you don't serve
Every broad category probably contains some of these. The insights page won't tell you which ones. The term report will.
2. Compare the category with what you actually sell
A category close to your offer, like "emergency plumbing" for a plumber, usually has mostly good queries with a few bad ones. A category that is only loosely related, like "plumbing tools" for a service business, usually has mostly bad queries with a few good ones. That ratio tells you how much effort to put in. It still doesn't tell you what to block.
3. Look at volume next to conversions
A category with a lot of search volume and few conversions is where most waste piles up. A category with low volume and zero conversions is often just a small sample. Look at both, then open the high-volume, low-conversion ones first.
4. Find the asset group behind it
Categories often trace back to one asset group with loose assets, such as a final URL that lets Google send traffic to any page, or headlines that match a wider topic than you meant. If one category keeps showing up with odd traffic, check whether one asset group is pulling it in. We cover that pattern in when one PMax asset group pulls irrelevant traffic. Sometimes fixing the asset group is better than adding any negative.
The wrong negatives category-level thinking produces
Turning categories into negatives creates three typical mistakes. Each one costs more than the waste it was meant to remove.
Mistake 1: the category name as a broad negative
You see a weak "water heaters" category and add water heaters as a broad negative. Broad negatives behave unpredictably across match types, so you can't easily predict what they block. A broad negative on a two-word category label can catch "water heater repair", "water heater leaking" and "water heater installation". Those are the queries a plumber wants most. The category had $600 of zero-conversion spend. The negative might now block searches that would have produced next month's best jobs, and no report will tell you, because blocked searches don't appear anywhere.
Mistake 2: blocking the category, missing the term
You see a healthy category and leave it alone, like the $3,000 "plumbing services" row above. The $500 of job-seeker traffic keeps running. This is the quieter mistake. Nothing breaks, but the waste never stops, and month after month it adds up to a lot. Small leaks like this are the subject of low volume terms in the search terms report add up.
Mistake 3: negatives that conflict with your Search campaigns
Advertisers who run Search next to Performance Max sometimes block a whole category in PMax to "push it to Search". That's a reasonable goal. But adding category labels as negatives in PMax rarely does it cleanly. Some of the blocked queries weren't covered by any Search keyword, so the traffic simply disappears. Before you block for this reason, compare which campaign actually serves each term. Performance Max cannibalizing Search? Compare the terms shows how. If brand searches are the concern, the method in measuring Performance Max branded search spend is more reliable than guessing from a category.
All three mistakes share one cause. You made a query-level decision from group-level data.
From category to the actual terms
Performance Max search terms are reported at query level, so you don't have to guess. Search, Shopping and Performance Max all report search terms in campaign_search_term_view, the same data source behind the search terms report. Display, Video and App campaigns report none. In Performance Max, that means the terms you see come from search placements. Clicks on display or video placements inside the same campaign won't show up as search terms.
Here's a workflow that keeps categories in their proper place:
- Start in insights. List the categories with high volume and weak conversions, plus any category whose label is only loosely related to what you sell.
- Open the terms. For each category on your list, pull the individual terms with their cost, clicks and conversions over a period long enough to judge. For most accounts that's at least 30 days, and longer for expensive, slow-converting services.
- Judge each term on intent first, numbers second. "plumber salary" is a job search whatever the click count. "tankless water heater install" is a buyer even at zero conversions. Our method for doing this without guessing is in add negative keywords from search terms without guessing.
- Choose the narrowest match type that does the job. Phrase match on the modifier ("salary", "apprenticeship", "reviews") usually blocks the whole family of bad queries without touching good ones. Exact match suits one-off oddities. Avoid broad. If you want ready-made patterns, see phrase match negative keyword examples you can copy safely.
- Check against what converted. Before you add a negative, make sure it wouldn't block any term that converted recently. A phrase negative on "install" would have been a disaster in the water heater example.
- Add at campaign level. A Performance Max campaign accepts up to 10,000 campaign-level negative keywords. It doesn't accept shared negative keyword lists. Every negative goes directly onto the campaign. If you run several PMax campaigns, you add the same negative to each one. Standard Shopping, by contrast, accepts both negative keywords and shared lists, the same way Search does.
- Go back to insights a few weeks later. This is where categories become useful again. If the "plumbing services" category kept its conversions and lost some spend, your negatives worked. If its conversions dropped, one of your negatives was too wide. Go back to the term level to find it.
The order matters. Insights tell you where to start and give you a way to check the result afterwards. The terms are where you actually decide.
Where the usual advice stops short
Most guides on this topic say the same thing: open search term insights, look for irrelevant categories, add negatives. That advice is not wrong, but it skips the step that decides whether the negatives help or hurt.
It treats a category as if it were a query. Real waste usually doesn't come from a category that is irrelevant as a whole. When a whole category is irrelevant, it tends to be obvious, and it is often a setup problem, such as a final URL that lets Google send traffic to unrelated pages. Ongoing waste usually comes from modifiers inside relevant categories: "salary" inside plumbing, "free" inside software, "used" inside equipment, "diy" inside home services. Insights tend to hide exactly these modifiers, because a category is built by removing them.
The common advice also skips match types. "Add the category as a negative" in practice means someone types the category label into the negative keyword box. If they leave the default or choose broad, the negative reaches further than they think. And it skips the check against converted terms. Without that check, every negative is a bet that nothing good was hiding behind the label.
Finally, the usual advice treats insights as the only view of PMax queries. That made sense when query-level data was harder to get. Today the terms are in the same report data as Search and Shopping. If you only work from categories, you are choosing to work with less detail than you already have.
If you have added negatives this way and Performance Max still seems to show ads on the same queries, the problem may be the negatives themselves, not the traffic. Performance Max negative keywords not working covers the usual causes, and negative keyword conflicts Google Ads won't flag covers the cases where a negative quietly blocks something you wanted.
What PPCleaner does here, and what it can't
PPCleaner works at the term level, not the category level. Every night it reads the search terms your Search, Shopping and Performance Max campaigns paid for from campaign_search_term_view. It judges which terms were never going to convert and adds them as negatives. For Performance Max, they go on as campaign-level negatives, because that's the only kind PMax accepts. In practice, that means "plumber salary" and "plumbing apprenticeship" get blocked, while "tankless water heater install" stays live even though its category looks weak.
Two rules keep this safe. PPCleaner never blocks a term that converted in the last 90 days. That's the check from step 5 above, done every night. And every block can be undone in one click for 90 days, so a negative that turns out to be too wide costs you a click to reverse, not an account audit. PPCleaner never touches budgets, bids, ads or keywords.
Here's what it can't do:
- It can't act on terms Google doesn't report. Searches hidden for privacy reasons aren't in
campaign_search_term_view, so neither PPCleaner nor anyone else can add them as negatives one by one. - It can't clean up display or video placements inside Performance Max. Those placements produce no search terms, so negative keywords aren't the tool for them.
- It can't go past the 10,000-negative limit per PMax campaign. For most accounts that's a lot of room. Very large catalogues with years of negatives can get close, and then choosing which negatives to keep matters.
- It doesn't fix the asset group or final URL that pulls the traffic in. If one asset group keeps attracting the wrong category, negatives treat the symptom. Fixing the asset group is still your job.
Search term insights are still worth opening. Use them to see what kind of demand your campaign reaches and to check whether your cleanup changed anything. Just don't choose a single negative from a category label.
Frequently asked
Can I add a search term insights category as a negative keyword?
No. A category isn't something you can target or exclude. You can type its label as a negative keyword, but it then works as an ordinary keyword. As a broad negative it can block good queries that share the label's words. Add negatives from the individual terms behind the category, mostly as phrase or exact match.
Where do I see the actual search terms for Performance Max?
Performance Max search terms are reported in campaign_search_term_view, the same data source used for Search and Shopping search terms. Searches withheld for privacy reasons and clicks from display or video placements won't appear there as search terms.
Can I use a shared negative keyword list on Performance Max?
No. Performance Max doesn't accept shared negative keyword lists. You add negatives at campaign level, up to 10,000 per campaign. Standard Shopping does accept shared lists, like Search.
A category has zero conversions. Should I block it?
Not as a whole. Open the terms behind it first. A zero-conversion category can contain buyer searches that simply haven't converted yet, especially for expensive services with long sales cycles. Block only the terms with clear non-buying intent, such as 'reviews', 'salary' or 'diy', and leave the rest to collect more data.
What are search term insights actually good for, then?
Planning and checking. They show what kinds of demand a Performance Max campaign is reaching, point you to the categories worth opening first, and let you check a few weeks later whether your negatives reduced spend without reducing conversions.
Clean your search terms. Save time and money.
Every night, PPCleaner reviews the searches your campaigns actually paid for, groups them by intent, and excludes the ones that never convert. No spend threshold to wait for, nothing touched but negatives — and one click undoes any of it.
Start 7 days freeSearch, Shopping and Performance Max · Cancel in one click
Related articles

Performance Max Wasted Spend on Display: How to Spot It
Performance Max spends across Search, Shopping, Display, YouTube, Gmail and Discover, and negative keywords only reach the first two. Here is how to find the split, judge whether the non-search part is waste, and know what you can actually block.

Performance Max Search Terms That Never Convert
Performance Max gives you a capped set of campaign-level negatives and no shared lists. Here is how to decide which non-converting search terms deserve a slot first, and which to leave alone.

Performance Max Brand Cannibalization in Your Search Terms
Performance Max often reports conversions from people who typed your brand name and were going to buy anyway. Here is how to match its search terms against a brand list and work out how much of its result was already yours.
