Negative Keyword Match Types: Which One to Use and When

Negatives use the same three match types as positive keywords, but the rules invert in ways that quietly cost you traffic if you pick wrong.

· QueryCut

The short version

  • Negatives use three match types (broad, phrase, exact) but never match plurals, misspellings, or synonyms the way positive keywords do.
  • Blocking too much is invisible and often permanent; blocking too little is visible and recoverable. Default to the tighter match type.
  • Negative broad is for single disqualifying words only; negative phrase is your everyday default; negative exact is a scalpel for one bad query inside a good theme.
  • Never apply a generic downloaded list at the account level without checking it against your own converting search terms.
  • There is no report for queries a negative blocked, so keep negatives auditable and revisit any broad ones.

Negative keywords come in three match types: broad, phrase, and exact. Default to negative phrase match for most blocking, use negative exact when you want to kill one specific query without touching close variants, and reserve negative broad match for single throwaway words you never want to serve on. The critical thing to know: negative match types behave differently from positive ones, and the difference is what trips people up.

The three types and exactly what each blocks

You set the match type the same way you do for positive keywords, using the surrounding punctuation. In Google Ads, when you add negatives under Campaigns > Audiences, keywords and content > Search keywords > Negative keywords (or at the account level under Tools > Shared library > Exclusion lists), the syntax is: no punctuation for broad, quotes for phrase, and square brackets for exact.

Negative broad match blocks a search only when the query contains every word in your negative, in any order. The negative broad keyword running shoes blocks "buy running shoes" and "shoes for running" but does not block "running" alone or "jogging shoes." Note the key departure from positive broad match: negatives do not match synonyms, close variants, or misspellings. "running shoe" (singular) will not be blocked by the negative broad running shoes.

Negative phrase match blocks searches that contain your exact phrase in order. The negative phrase "running shoes" blocks "best running shoes" and "running shoes cheap" but not "shoes for running" (wrong order) and not "running shoe" (again, no plural or variant matching).

Negative exact match blocks only the exact query, nothing more. The negative exact [running shoes] blocks the search "running shoes" and nothing else. "buy running shoes" still serves. "running shoe" still serves. It is a scalpel.

The one rule that makes negatives behave differently: no close variants

This is the mechanism that catches experienced people off guard. Positive keywords, since the 2019 close-variant changes, match plurals, misspellings, abbreviations, and reordered words automatically. Negative keywords do none of that. Google will not silently expand your negative to catch "the obvious variant." It blocks literally what you typed, subject to the match-type rules above.

Why does Google build it this way? Because the cost of an over-eager negative is invisible and permanent. If a positive keyword matches too loosely, you see the wasted spend in your search terms report and can add a negative. If a negative blocks too loosely, you never serve, you get no impression, no click, no data, and nothing in any report tells you the sale you never got. Google keeps negatives literal so that blocking is a deliberate act, not an inference the system makes on your behalf.

The practical consequence: one negative rarely covers a theme. If you want to block the concept "free," the negative broad free does not catch "freebie," "for free," spelled variants, or "gratis." You have to enumerate. This is the single biggest reason negative lists need maintenance rather than one-time setup.

The asymmetric risk that should drive every match-type choice

Here is the reasoning that makes the default choices safe. A negative keyword is a one-way door: it removes queries from your eligibility. The two ways to get it wrong are not equally costly.

Block too little, and the failure mode is visible and recoverable. Junk terms show up in Campaigns > Insights and reports > Search terms, you spend a few dollars, you see them, you add the negative next week. The damage is a small, measurable amount of wasted spend that you can always claw back.

Block too much, and the failure mode is invisible and often permanent. A too-broad negative silently suppresses queries that would have converted. There is no line item for it. You will not find it in a report, because a blocked query generates no search-term row. You lose the revenue and you lose the evidence that you lost it. That asymmetry is why the correct default leans toward the narrower, more predictable match type. Phrase over broad, exact over phrase, whenever the tighter type still catches what you actually want gone.

This principle is predictive. Any time you are unsure whether a negative might catch a query you'd want, choose the tighter match type, because the cost of leaving a bad term running for a week is smaller than the cost of never knowing you killed a good one.

Decision rules you can apply this afternoon

Use these defaults and you will be right the large majority of the time.

Use negative broad match only for single words that are unambiguously disqualifying no matter the context: free, cheap, jobs, salary, diy, torrent, used, wholesale (if you sell only new retail). One-word broad negatives are the exception where broad is correct, because there is no word order or variant risk to worry about with a single token, and the word poisons any query it appears in.

Use negative phrase match as your default for multi-word junk themes: "how to make," "is it worth it," "customer service number," competitor product lines you don't want to bid against. Phrase respects word order, so it blocks the theme without the risk that a broad negative's any-order matching sweeps in a legitimate query. Example: the negative broad car insurance would block "insurance for my car business"; the negative phrase "car insurance" would not, because the words are not adjacent in order.

Use negative exact match for surgical removal of one specific query you've seen in your search terms report that converts poorly but sits inside a theme you otherwise want. If "[brand] reviews" wastes money but "[brand] pricing" converts, exact-negative the review query and leave everything else eligible.

  • Single disqualifying word: negative broad (free, jobs, diy)
  • Multi-word junk theme in order: negative phrase (default)
  • One specific bad query inside a good theme: negative exact
  • When unsure if a negative is too wide: pick the tighter type

Traps, limits, and what the tools do not tell you

The trap that costs the most: applying an aggressive negative list at the account level via a shared exclusion list before you've checked it against your own converting search terms. Account-level negatives apply everywhere, including campaigns you weren't thinking about. A generic "free negative keyword list" pulled off the internet is built for someone else's business. If you sell free-trial software, the word free is a buying signal, not junk. Read every list against your actual converting terms before you apply it, and prefer applying at the ad group or campaign level first so a mistake is contained.

Known limits worth stating plainly: a single negative keyword is capped at 10 words. Anything longer will not save. Google recommends keeping negative keyword lists to a manageable size (the shared-library limit is large but not infinite), and very large lists become impossible to audit by eye, which is its own risk because an un-auditable list is where the silent over-blocking hides.

What you genuinely cannot know: you cannot measure the queries a negative blocked. There is no "negatives blocked this" report. The only defense against over-blocking is to keep the match type as tight as the job allows and to revisit any broad-match negative periodically by asking, "would I be comfortable if this word appeared in a query I've never seen?" If the answer is no, tighten it to phrase.

A note on generators. Tools that scan your search terms and propose negatives (QueryCut is one; it works from your own account's search-term data rather than a generic list) solve the enumeration problem, since they surface the actual variants your traffic contains instead of the ones you guessed. But no tool can see the converting query a proposed negative would suppress on future traffic. Treat every suggested negative as a proposal to review, not an instruction, and the tighter match type is always the safer way to accept one.

Questions people also ask

What is the best match type for negative keywords?

Negative phrase match is the best default for multi-word junk themes because it respects word order and won't sweep in legitimate queries the way negative broad can. Use negative broad only for single disqualifying words like free or jobs, and negative exact when you need to remove one specific query without touching the rest of a theme.

How do negative exact match keywords work?

A negative exact keyword, written in square brackets like [running shoes], blocks only that precise query and nothing else. Unlike positive exact match, it does not extend to plurals, misspellings, or reordered words. Use it to surgically remove a single bad-performing query that sits inside a theme you otherwise want to keep serving on.

Do negative keywords match plurals and misspellings?

No. Negative keywords never match close variants: plurals, misspellings, abbreviations, or synonyms are all ignored. The negative running shoes will not block "running shoe" or "runing shoes." This is why one negative rarely covers a theme and you often have to enumerate the variants explicitly.

Is it safe to use a free negative keyword list generator?

A downloaded generic list is built for a hypothetical business, not yours, so words that are junk for most advertisers (like free or cheap) may be buying signals for you. Check every entry against your own converting search terms before applying, and prefer campaign or ad group level over account level so a mistake stays contained. Tools that generate negatives from your own account's search terms are safer because they reflect your real traffic, but you should still review each suggestion.

What is the word limit on a negative keyword?

A single negative keyword can contain at most 10 words. Anything longer will not save. For long or repeated exclusions across campaigns, use a negative keyword list under Tools > Shared library > Exclusion lists, but keep it small enough that you can still audit it by eye, since un-auditable lists are where silent over-blocking hides.

More from the blog

Negative Keyword Match Types: How Broad, Phrase, and Exact Actually BlockFree Negative Keyword Tools: What They Actually Do and When to Skip ThemFree Negative Keyword Tools: What Actually Works and How to Use OneExact Match Negative Keywords: When They Pay and When They Backfire