Google Ads Negative Keywords: How to Build a List That Actually Cuts Wasted Spend

Every Google Ads account has a version of this problem. A search term shows up in the report, it looks close enough to what you sell, so nobody flags it. Three months later you pull the search terms report properly and find you’ve spent $4,200 on clicks for “free,” “jobs,” and a competitor’s brand name that never had a chance of converting. Nobody made a bad decision on purpose. The budget just leaked out through a hundred small gaps nobody was watching.

Negative keywords are the tool built to close those gaps, and they’re also one of the most misunderstood parts of Google Ads. Most advertisers know the basic idea — tell Google not to show your ad for certain searches — but far fewer understand the match type behavior, the list architecture, or how the rules have quietly changed for Performance Max and AI Max campaigns over the past year and a half. Get the mechanics wrong and you either leave money on the table or accidentally block traffic you wanted. Get them right and negative keywords become one of the highest-leverage, lowest-effort things you can do to an account.

This is a practical guide to building and maintaining a negative keyword list that actually holds up — not a copy-paste list of 500 generic terms, but the process for finding the negatives that matter for your account specifically, structuring them so they don’t create new problems, and keeping them current as Google keeps changing how they apply.

What Negative Keywords Actually Do

A negative keyword tells Google Ads not to show your ad when a search query contains (or matches, depending on match type) that term. That’s the whole mechanism. The complexity comes from where you add them, what match type you use, and how they interact with the rest of your account.

Negative keywords always take precedence over positive keywords. If a search query would trigger one of your keywords but also matches a negative, the negative wins and your ad doesn’t show, full stop. There’s no bidding your way past a negative keyword and no Quality Score override. This is useful to know because it means negatives are absolute — which is exactly why sloppy ones cause real damage, a point we’ll come back to.

People tend to think of negative keywords purely as a cost-control tool, and that’s true, but it undersells what they do. A tight negative list also improves the quality of the traffic reaching your landing pages, which affects conversion rate, which affects Quality Score, which affects the price you pay for the traffic you do want. If you’ve read our piece on what Quality Score actually measures, you already know relevance is judged relative to the query, not in the abstract. Negative keywords are one of the few levers you have to shape which queries you’re being judged against in the first place.

Where to Find Your Negatives: The Search Terms Report

The generic negative keyword lists you can find online (block “free,” “cheap,” “jobs,” “how to,” “DIY,” and so on) are a reasonable starting point, but they’re not where the real value is. The real value is in your own account’s search terms report, because that’s the only place that shows you the actual queries triggering your ads, not the keywords you set up to target them.

We’ve written a full breakdown of how to read a Google Ads search term report, but the short version for negative keyword purposes is this: pull the report, sort by cost or clicks, and look for two patterns. First, queries with meaningful spend and zero conversions — these are your most obvious candidates. Second, queries that are topically adjacent but wrong in intent. Someone searching “what is CRM software” when you sell enterprise CRM implementation services is not the same buyer as someone searching “CRM software for accounting firms,” even though both queries would match a broad keyword around “CRM software.”

One thing worth knowing going in: you won’t see every query. Google has withheld search terms below an undisclosed activity threshold since 2020, citing user privacy — the rough rule of thumb reported by advertisers is that queries generating very low impression volume in a given window don’t get broken out individually. If a hidden query converts, you’ll still get credit for the conversion; you just won’t know which exact phrase drove it. This isn’t a reason to give up on search term mining — most of the wasted spend sits in the visible, higher-volume queries anyway — but it’s worth knowing the report is not a complete picture, so don’t assume silence in the report means silence in your account.

A cadence that works for most accounts: review search terms weekly for the first month after any major change (new campaign, new match type, broad match test, AI Max rollout), then settle into every two to three weeks once the account has stabilized. Waiting a full month between reviews on a new campaign usually means you’ve already spent past the point where the negatives would have paid for themselves.

The Universal Negatives (And Why They’re Not Enough on Their Own)

Most accounts benefit from a baseline list of low-intent modifiers applied broadly: “free,” “cheap,” “jobs,” “salary,” “career,” “template,” “example,” “definition,” “how does it work,” “DIY,” and similar words that signal someone is researching or job-hunting rather than buying. If you’re in B2B or high-ticket services, add “cheap” and “affordable” cautiously — someone hunting for the lowest possible price is rarely the buyer you want, but exclude too aggressively and you can cut off legitimate price-comparison shoppers who convert just fine.

The mistake is treating this universal list as sufficient. It isn’t. It catches maybe 20-30% of realistic waste. The other 70-80% is specific to your business: adjacent services you don’t offer, locations you don’t serve, product categories that sound similar but aren’t what you sell, and competitor names if you don’t want to bid on them. None of that shows up in a generic list — it only shows up when you actually read your own search terms.

This is also where account context matters. A local plumber and a SaaS company both might want to exclude “free,” but the plumber also needs to exclude every neighboring city they don’t service, while the SaaS company needs to exclude every adjacent product category their competitors rank for. If you serve a specific vertical, our guides for local service businesses and for ecommerce accounts cover some of these vertical-specific patterns in more depth.

Match Types for Negative Keywords (Where Most of the Damage Happens)

Negative keywords have match types too — broad, phrase, and exact — and they don’t behave the way positive keyword match types behave. This trips up more accounts than almost anything else in this guide.

Negative broad match blocks a query if it contains all the words in your negative keyword, in any order, but does not include close variants, synonyms, or expanded matching the way positive broad match does. If you add “running shoes” as a negative broad match, it blocks “best running shoes,” “running shoes for men,” and “shoes for running” — but it does not block “running sneakers” or “jogging shoes,” because those don’t contain both exact words. This is actually less aggressive than people assume, which is a common source of “why isn’t this negative working” confusion.

Negative phrase match blocks a query that contains your negative phrase in the exact word order, allowing other words before or after. “Running shoes” as phrase match blocks “cheap running shoes online” but not “shoes for running.”

Negative exact match blocks only the precise query, word for word (with minor allowances for close variants like plurals and misspellings in some cases). This is the tightest and safest option, but also the one that requires the most volume of individual entries to have real coverage.

The mistake that costs people traffic: adding a single broad-match negative for a word that’s also part of a phrase you want to rank for. A classic example is a company selling job-posting software adding “jobs” as a negative broad match to filter out job-seekers — and inadvertently blocking “post jobs online” and “manage job applications,” which were exactly the queries they wanted. When a negative keyword is broader than it needs to be, use phrase or exact instead, and always sanity-check a new negative against your existing positive keyword list before adding it at broad match.

Why This Matters More Once Smart Bidding Is Involved

There’s a second, less obvious reason negative keywords matter more than their direct cost savings suggest: automated bidding strategies learn from your conversion data, and irrelevant traffic pollutes that signal. If Target CPA or Maximize Conversions is optimizing toward a pattern that includes a chunk of clicks from searches that were never going to convert, the algorithm isn’t just wasting the spend on those clicks — it’s also drawing slightly wrong conclusions about what a “good” click looks like for your account, and adjusting future bids accordingly. We go into the tradeoffs between bidding strategies in more depth in our comparison of Target CPA, Target ROAS, and Maximize Conversions, but the short version here is that a cleaner query stream doesn’t just save money directly — it makes the automated systems downstream of it work with better information.

This is part of why negative keyword hygiene tends to matter more, not less, as accounts lean further into automation. When you controlled match types tightly and bid manually, a handful of irrelevant clicks were a nuisance. When a bidding algorithm is making thousands of micro-decisions a day based on the patterns in your account’s conversion data, the same irrelevant clicks are quietly nudging every one of those decisions in a slightly wrong direction. It’s a small effect per click, but it compounds, and it’s invisible unless you’re specifically looking at the search terms feeding the model.

What Negative Keywords Can’t Fix

It’s worth being honest about the limits here, because negative keywords get treated sometimes as a catch-all fix for poor targeting, and they’re not. They’re a reactive tool — they clean up after a keyword or a campaign has already attracted the wrong traffic. They don’t fix a keyword list that’s fundamentally too broad for what you’re selling, and they don’t fix an ad group that’s trying to cover too many distinct buyer intents with one set of ads.

If you find yourself adding dozens of negatives a month to the same ad group just to keep it pointed at the right audience, that’s usually a sign the ad group itself is too broadly themed, not a sign you need more negatives. The fix in that case is usually restructuring — splitting the ad group into more tightly themed groups so each one naturally attracts a narrower set of queries — rather than trying to negative your way to precision from a structure that was never precise to begin with. Negative keywords are the fine-tuning layer on top of a reasonably well-structured account, not a substitute for one.

The same logic applies to broad match specifically. Broad match keywords rely heavily on Google’s matching algorithm to find relevant queries, and they can pull in a wider spread of traffic than phrase or exact match by design. That’s not automatically a bad thing — broad match with strong negative keyword coverage and Smart Bidding can genuinely outperform tightly restricted match types in some accounts — but it does mean broad match campaigns need a more actively maintained negative list than exact match campaigns do, simply because there’s more surface area for the algorithm to get wrong. If you’re running broad match without a negative list that’s kept current, you’re relying entirely on Google’s matching quality to stay relevant, with no backstop of your own.

The Conflict Problem: When a Negative Silently Blocks Your Own Keyword

Because negative keywords always override positive keywords, a badly scoped negative can silently switch off a keyword you’re actively bidding on — and Google Ads won’t necessarily warn you. The keyword’s status in the interface will still say “eligible”; it just won’t ever actually serve, because every query that would trigger it also trips a negative first.

This is genuinely hard to catch by eye once an account has more than a handful of campaigns and a few hundred keywords. Google’s Recommendations tab sometimes surfaces a “remove conflicting negative keywords” suggestion, but it’s inconsistent and shouldn’t be your only check. The more reliable approach is exporting your negative keywords and your active keywords and cross-referencing them systematically — either manually for a small account, or with a script for anything larger. A few agencies publish free Google Ads scripts that do exactly this: they scan every active keyword against every applied negative and flag exact overlaps and broad-match collisions, then drop the results into a spreadsheet.

Whichever method you use, the point is to check this periodically, not just when you set the account up. Negative lists grow over time as more people add to them, campaigns get restructured, and keywords get added that nobody cross-checks against the existing negatives. A quarterly conflict check is a reasonable minimum; monthly is better if you have more than one person touching the account.

Structuring Your Lists: Account, Shared List, Campaign, and Ad Group Level

Negative keywords can live in four places, and where you put them matters:

• Account-level negative keyword lists (shared lists): A shared list can hold up to 5,000 keywords, and an account can have up to 20 such lists, applied to as many campaigns as you want. This is the right home for negatives that should apply everywhere — universal exclusions, competitor names you never want to bid on, locations you’ll never serve.
• Campaign-level negatives: Specific to one campaign, useful when a term is irrelevant for this campaign but perfectly fine for another. A common example: a company running separate campaigns for “installation services” and “replacement parts” needs to negative “installation” out of the parts campaign without blocking it account-wide.
• Ad group-level negatives: The most granular option, used to stop keyword overlap within a single campaign — for instance, making sure your broad “running shoes” ad group doesn’t also fire for queries that should route to your more specific “trail running shoes” ad group.
• Performance Max negatives: Handled separately from the above, and worth its own section, because the rules here changed substantially over the past 18 months.

The practical structure that works for most accounts: build one or two shared lists for universal exclusions and known irrelevant categories, apply them account-wide, then layer campaign-level and ad group-level negatives on top for anything that’s only irrelevant in a specific context. Resist the urge to build negatives ad group by ad group from scratch — it’s slow, and it’s the structure most likely to produce the conflicts described above, since nobody has visibility into what’s already excluded elsewhere in the account.

Negative Keywords in Performance Max: What Changed

For a long time, Performance Max was effectively a black box on this front — advertisers had no self-serve way to add negative keywords at all, and had to submit a support request to Google to get anything excluded, capped at a token 100 terms. That changed through 2025. Google began rolling out self-serve campaign-level negative keywords for PMax in December 2024, and by January 23, 2025 the feature was fully available in the standard Google Ads interface, letting advertisers add and manage PMax negatives themselves. In March 2025, Google raised the campaign-level cap from 100 to 10,000 keywords, and account-level PMax negatives (applied across every PMax campaign) now sit at a separate limit of roughly 1,000.

If you’re running Performance Max and haven’t checked whether your negative keyword settings reflect this, it’s worth a look — plenty of accounts still have PMax campaigns running with either no negatives at all, or with the same handful that were submitted through the old support-ticket process years ago and never updated. Given that PMax spans Search, Display, Shopping, YouTube, and Discovery inventory simultaneously, a stale or empty negative list there tends to cause more wasted spend than a stale list on a single Search campaign, simply because of how many surfaces it touches. If you’re trying to understand what’s actually happening inside a Performance Max campaign more broadly — since Google gives you comparatively little visibility into it — our guide on how to audit a Performance Max campaign covers what you can and can’t see, and where negative keywords fit into that picture.

Negative Keywords and AI Max for Search

AI Max for Search adds a different wrinkle. Google has been explicit that standard negative keywords are still respected under AI Max — they’re applied before the AI’s semantic query-expansion logic runs, meaning a negative keyword will still block a query even if that query is otherwise a strong semantic match for your ad. That’s the reassuring part.

The messier part is what happens with brand inclusion and exclusion settings. Those used to be a broad match feature; they’ve now effectively been folded into AI Max, meaning you need AI Max’s keywordless targeting enabled to use brand controls at all. Independent reporting on early AI Max rollouts has found real leakage even with brand exclusions active — brand terms occasionally matching to non-branded queries, non-branded terms occasionally matching to competitor queries, and in some cases brand queries slipping through to competitor-related terms. Brand filters help, but misspellings and close variants still get through in some accounts, which means a solid negative keyword list isn’t optional under AI Max — it’s a more necessary backstop than it was under standard Search campaigns, not a less necessary one.

If you’re testing or running AI Max, the sequencing that seems to hold up best is: build a comprehensive negative keyword list before turning AI Max on, not after. Add competitor brand names and competitor product names as negatives explicitly, since AI Max’s broader query matching makes it more likely your ad shows for a competitor-branded search than tighter match types would. If you haven’t already looked at how AI Max changes campaign behavior more broadly, we cover that in more depth in what AI Max actually changes, and what you still control.

Common Mistakes That Undo the Work

A few patterns show up repeatedly in accounts that have negative keywords in place but still leak spend:

• Set-and-forget lists. A negative list built once during account setup and never revisited stops reflecting the account’s actual traffic within a few months, especially if you’ve launched new products, entered new markets, or changed match types.
• Over-broad negatives that quietly cut off good traffic. As covered above, this is usually a match-type mistake, and it’s genuinely hard to notice because a keyword that’s blocked by a conflicting negative just looks like a keyword that isn’t getting impressions — which reads as “low demand” rather than “silently disabled.”
• Negatives applied at the wrong level. Adding a genuinely universal exclusion at the ad group level means it has to be re-added to every new ad group manually, forever. Adding a campaign-specific exclusion to a shared list means it now applies somewhere it shouldn’t.
• Ignoring Performance Max and AI Max because “negatives don’t really work there.” That was true a couple of years ago. It’s substantially less true now, and treating those campaign types as ungovernable means leaving one of your biggest spend categories with the least oversight.
• Copy-pasting someone else’s list wholesale. Generic negative lists circulating online are a fine starting point for universal terms, but plenty of them include category-specific exclusions that make sense for the account that originally published them and actively hurt yours. Always read a borrowed list before applying it, not after.

Building a Maintenance Cadence That Actually Holds

The single biggest determinant of whether a negative keyword list stays useful isn’t how well it’s built on day one — it’s whether anyone keeps looking at it. Accounts that get this right tend to treat negative keyword review as a standing part of account monitoring rather than a project that gets scheduled once and then quietly drops off. We’ve written separately about how often you should actually be checking your Google Ads account, and the honest answer is more often than most people budget for, particularly in the weeks after any structural change.

A workable cadence, roughly:

• Weekly for the first month of a new campaign, a new match type test, or a rollout of AI Max or Performance Max.
• Every two to three weeks once an account has stabilized and spend patterns are predictable.
• Quarterly for a full negative keyword conflict check across the account, cross-referencing every active keyword against every applied negative.
• Immediately after any major account change — new product line, new geography, new campaign type, a shift in bidding strategy — rather than waiting for the next scheduled review.

This is also, honestly, where a lot of accounts fall down — not because the process is complicated, but because nobody’s actually watching consistently enough to catch drift early. This is part of why we built Growera’s Google Ads management platform around continuous daily account monitoring rather than the monthly-report model most tools default to: catching a search term that’s quietly burning budget in week one is a very different outcome than catching it in month three, and daily visibility is what makes that possible without someone manually pulling reports every day. It’s not a replacement for understanding what you’re looking at — everything in this guide still applies — but it does mean the gap between “a bad query starts converting your budget into nothing” and “someone notices” shrinks from weeks to a day or two.

A Simple Framework to Start From

If you’re building this from scratch today, here’s a reasonable order of operations:

• Pull your search terms report going back 90 days (or as far back as you have meaningful spend) and sort by cost.
• Flag every query with spend and zero conversions, and every query that’s topically related but wrong-intent.
• Build one shared negative keyword list for universal, account-wide exclusions and apply it to every campaign.
• Add campaign-specific and ad group-specific negatives for anything that’s only wrong in one context.
• Check your Performance Max campaigns specifically — most accounts have under-maintained negatives there simply because the feature is newer.
• If you’re running AI Max, build out competitor and brand-adjacent negatives before enabling it, not after.
• Run a conflict check to make sure nothing you just added is silently blocking a keyword you still want.
• Put a recurring date on the calendar to repeat the steps above. This last step is the one people skip, and it’s the one that determines whether any of this holds up six months from now.

The Bottom Line

Negative keywords aren’t a “set it up once” feature, and treating them that way is probably the single most common reason accounts keep leaking budget on searches that were never going to convert. The mechanics matter — match type choice determines whether a negative is too loose or accidentally too tight, list placement determines whether an exclusion applies where it should, and the rules for Performance Max and AI Max have genuinely changed enough in the last year that assumptions from 2023 don’t hold anymore.

The good news is that none of this requires guesswork. Your own search terms report tells you exactly where the waste is; you just have to look at it regularly and act on what you find. If you’re not sure where your account currently stands, a decent gut check is running through the framework above against your own campaigns this week rather than waiting for the quarterly review — the accounts we see with the cleanest data are almost always the ones where negative keywords get looked at as often as bids and budgets do, not treated as a one-time setup task. You can see how Growera approaches this kind of ongoing account oversight, including current plans and pricing, on the Growera homepage.