Google Ads Keyword Match Types: What Changed in 2026

In 2026, Google Ads offers three keyword match types — broad, phrase, and exact. Each one functions as a bidding-strategy coupling decision, not a standalone precision dial. Broad match has been the default for new Search campaigns since July 2024. The question is not which type gives control — it is which bidding-and-match-type combination fits the account's conversion data volume.

Match types govern whether an ad is eligible to show for a given search query. The legacy framing — broad match for reach, exact match for precision — no longer describes how these types work. Since July 2024, Google sets new Search campaigns to broad match by default when a conversion-based Smart Bidding strategy is selected, a change that redefined how advertisers must think about keyword architecture from the moment a campaign is created. Match types are now decision points about automation coupling, not just query expansion breadth.

The three match types differ in how they interpret a user's search query. Broad match expands to related meanings and implied intent — Google's AI interprets the query at auction time and decides whether the keyword is relevant. Phrase match shows for queries containing the keyword meaning in the same general order; since the 2021 migration, it is intent-based and covers paraphrased queries where meaning is equivalent. Exact match covers queries with the same meaning or same intent as the keyword — not the same characters. Close variants including misspellings, plural forms, stems, abbreviations, and same-intent paraphrases are included by default with no opt-out available.

The "broad equals wasteful, exact equals safe" framing that dominated pre-2019 Google Ads strategy is obsolete. The real variables are conversion history depth, Smart Bidding signal maturity, and negative keyword coverage. An account running broad match with 150 conversions per month and Target CPA has the data infrastructure for Smart Bidding to make per-auction bid decisions that filter off-target queries. An account with 20 conversions per month running broad match has no such infrastructure. This pillar covers the match type behavior matrix, negative keyword architecture, Smart Bidding pairing logic, Keyword Planner accuracy, and the four misconceptions that generate the most wasted spend. For the bidding side of this equation, see Google Ads Bidding Strategies. For campaign type architecture decisions, see Google Ads Campaign Types.

Key Takeaways

Match Type Stats at a Glance

Broad Match + Smart Bidding: Google's 2026 Default Architecture

Broad match is Google's default for new Search campaigns as of July 2024, but only when a conversion-based Smart Bidding strategy is selected. Without Smart Bidding, broad match serves off-topic queries with no per-auction bid calibration. The pairing is the condition — not the match type alone.

Google's explicit framing is that broad match is correct only when paired with one of four Smart Bidding strategies: Target CPA, Target ROAS, Maximize Conversions, or Maximize Conversion Value. Without one of these strategies active, broad match serves queries with no per-query bid calibration to control cost. The AI that interprets query intent at auction time is the same AI that sets the bid — remove the bid automation and the intent filtering loses its cost-control mechanism entirely.

The performance data Google publishes: switching from phrase match to broad match with Target CPA delivers a published average of +25% more conversions while meeting the cost target; switching with Target ROAS delivers a published average of +12% more conversion value while meeting the return target. Both are Google-published averages — they are industry guidance, not guarantees, and Google has not disclosed the methodology or sample composition behind these figures. Additionally, Google's AI improvements to broad match quality delivered a +10% performance improvement, cited in Google's July 2024 announcement. These numbers establish the ceiling of what the pairing is designed to produce; actual results depend on conversion history depth, Smart Bidding signal maturity, and negative keyword architecture.

The campaign-level "broad match keywords" setting forces all keywords in a campaign to broad match and is available only when a conversion-based Smart Bidding strategy is active. This setting is designed for full-funnel campaigns where granular keyword list management is replaced by the combination of audience signals, bidding automation, and negative keyword exclusions. It is not a shortcut — it requires that the account has the conversion history to support full automation.

The conversion threshold question is widely discussed and poorly documented. The industry broadly uses 30 conversions per month as the minimum floor for Smart Bidding signal to be reliable; for Target ROAS specifically, 50 conversions per month is the practitioner benchmark. Neither number is officially published by Google — Google's Smart Bidding documentation uses the phrase "sufficient conversion history" without specifying a threshold. The 30 and 50 conversion benchmarks originate from practitioner consensus, not Google documentation. In accounts below these thresholds, phrase or exact match keeps budget predictable while conversion data accumulates to the level where Smart Bidding can function as designed.

Broad match is not the right architecture for every account state. Brand protection campaigns carry competitor-query overlap risk when broad match interprets branded queries as broadly related to competitor brands. New accounts under 30 conversions per month lack the Smart Bidding signal depth to bid accurately per query. Local service businesses in tight geographies — HVAC in a single suburb, a dental practice in a single zip code — face the additional problem that Smart Bidding does not reliably suppress out-of-area queries in sub-metro geotargeting. For industry-specific keyword architecture in tight geos, see HVAC PPC and the Flagstaff HVAC PPC local guide. For the bidding strategy layer, see Google Ads Bidding Strategies.

Match Type Behavior Matrix (2026)

The three match types differ in query expansion breadth, Smart Bidding coupling requirements, close-variant scope, and account fit. This table is the current state in 2026 — post-BMM retirement, post-close-variant expansion, post-July-2024 default change.

Syntax notation — [brackets] for exact match, "quotes" for phrase match, no wrapper for broad match — matters in Google Ads Editor and the API. In the Google Ads UI, match type is set via a dropdown when creating or editing a keyword; the syntax characters are not required in the UI. The syntax does not change behavior; it is a notation convention. For how match type interacts with Quality Score and Ad Rank, see Google Ads Quality Score and Ad Rank.

Exact Match in 2026: Close Variants You Cannot Opt Out Of

Exact match has not functioned as a character-for-character keyword lock since 2019. Google's current definition is same meaning or same intent — not same words. Close variants are included by default. There is no setting to disable this. A campaign running only exact match keywords still triggers on queries that were not on the original keyword list.

"Exact match gives you exact control." This was accurate before 2019. It has not been accurate since. Google introduced close variants for exact match in stages — initially misspellings and plurals, then same-meaning paraphrases in 2019. The current scope is broad enough that a campaign built entirely on exact match keywords regularly shows for queries that do not contain any keyword term verbatim. The 2026 close variant scope includes seven distinct expansion types.

The practical approach: run exact match keywords as your primary control layer, conduct weekly search-term report reviews (weekly is more effective than monthly for catching close-variant expansion early), and add exact-match negative keywords for specific variants you want to exclude. The seven close-variant types cannot be disabled globally — they can only be countered query by query through the negative keyword list. For how search-term data feeds into performance analysis, see Google Ads Metrics and KPIs. In high-CPC verticals where close-variant leakage has material CPA impact, the stakes are higher: see Legal PPC and Personal Injury Law PPC.

Negative Keywords: The Real Control Lever in 2026

Negative keywords do not match close variants. A negative broad keyword blocks a query only when the search contains all the negative terms — not just some of them. Each unwanted form requires its own entry. Plurals, synonyms, and semantically related terms are not blocked automatically.

The asymmetry between positive and negative keyword matching is the most frequently missed detail in keyword architecture. Positive keywords (broad, phrase, exact) all include some form of variant matching. Negative keywords do not. A negative broad match for "free" does not block "freely" or "complimentary" — each variant is treated as a distinct term. The June 2024 update changed one piece of this: negative keywords now automatically cover common misspellings of the negated term. But the coverage stops there. Synonyms, plurals, and adjacent terms still require individual entries. A negative for "cheap" does not cover "cheapest" or "budget" or "discount" — those must each be added explicitly.

The three negative match types follow the same broad / phrase / exact naming convention as positive match types, but their blocking logic is distinct. Negative broad blocks when a query contains ALL negative terms, in any order — but if a query contains only some of the negative terms, it is not blocked. Negative phrase blocks when all terms appear in the same order, with extra words permitted — different word order is not blocked. Negative exact blocks only when the query exactly matches the terms, in the same order, with no extra words.

A tiered negative keyword structure is the architecture that scales. The three tiers: account-level shared negative list (universal exclusions — irrelevant verticals, competitor brands that should never appear, content categories) applied across all campaigns; campaign-level negatives (theme-specific exclusions that prevent a campaign from serving queries better suited to another campaign); ad-group-level negatives (exact match exclusions preventing ad groups within the same campaign from cannibalizing each other's traffic). Account-level negative keywords are a distinct mechanism from shared negative lists — they apply directly to all campaigns in the account without the shared list infrastructure.

MB Adv Agency has found that accounts running broad match without a tiered negative keyword architecture accumulate off-target spend faster than most teams expect. The search-term report is the only reliable diagnostic — broad match without weekly search-term hygiene is a budget drain, not a reach strategy. For local service verticals where out-of-vertical queries are common, see Plumbing PPC, HVAC PPC, and the Plumbing PPC Missoula MT local guide.

Negative Keyword Limits: 2025 State

Performance Max campaigns expanded their per-campaign negative keyword limit from 100 to 10,000 in March 2025. Shared negative keyword lists became available for PMax in August 2025. Both changes removed the primary negative-keyword objection to running PMax alongside Search.

Keyword limits define the ceiling of your negative keyword architecture — and the right architecture uses all available tiers. A shared negative keyword list (capped at 5,000 entries) handles universal exclusions that apply across every campaign: irrelevant verticals, competitor brand names you never want to appear for, content categories incompatible with the account. Campaign-level negatives (up to 10,000 per Search campaign) handle theme-specific exclusions. Ad-group-level negatives handle inter-ad-group cannibalization prevention. Knowing the limits for each scope — especially PMax, which changed materially in 2025 — determines how the architecture is built.

A note on the shared list limit: Google Ads liaison Ginny Marvin confirmed that 5,000 per shared negative keyword list remains the official threshold. Some accounts have accepted more than 5,000 entries, but this has not been confirmed as a policy change. Use 5,000 as the planning ceiling for shared lists.

Practical architecture: a shared negative list handles brand exclusions and irrelevant-vertical exclusions uniformly across all campaigns, including PMax from August 2025. Campaign-level negatives handle theme control — preventing one campaign's keywords from pulling in queries that belong in a different campaign. Ad-group-level negatives prevent cannibalization between ad groups within the same campaign. For campaign structure decisions that intersect with this, see Google Ads Campaign Types. For industry applications in roofing and real estate where shared exclusion lists are essential, see Roofing PPC, Real Estate PPC, and the Legal PPC Missoula MT local guide.

MB Adv Agency has found that the March 2025 PMax negative keyword expansion changed how accounts with both Search and PMax campaigns are structured — shared negative lists that previously applied only to Search can now extend coverage to PMax, which closes a significant gap in brand and topic exclusion.

Choosing Match Types: Account Readiness Matrix

Match type selection in 2026 is a function of conversion volume and Smart Bidding signal maturity, not preference. Broad match without sufficient conversion history gives Smart Bidding insufficient data to bid accurately per query. Exact match without search-term hygiene leaks spend on close variants. The matrix below maps account state to the right architecture.

Industry practitioner thresholds (30 conv/mo for Smart Bidding; 50 conv/mo for Target ROAS) are widely cited benchmarks. Google's Smart Bidding documentation uses the phrase "sufficient conversion history" without publishing a fixed number. Source: Google Ads Help answer/7065882; groas.com 2025–2026 best practices guide.

For bidding strategy selection that maps to these thresholds, see Google Ads Bidding Strategies. For market-specific implementation in Austin and Chicago, see PPC Consultant Austin TX and the HVAC PPC Flagstaff AZ guide.

Match Type Performance Data: 16,825 Campaigns (December 2025)

Exact match keywords represent 55% of all keywords in the Adalysis dataset — yet they generate only 27% of impressions while capturing 40% of conversions. The gap quantifies the precision premium: highest conversion efficiency per impression, at the cost of missing more than half of available auction opportunities.

The Adalysis study covers 16,825 Google Ads Search campaigns from December 2025. The dataset skews toward performance-focused accounts actively managing match types — accounts that defaulted entirely to broad match following Google's July 2024 default change are likely underrepresented. The figures below reflect the match type distribution and performance in managed, performance-focused accounts rather than the full universe of Google Ads advertisers.

The +13 percentage-point gap between exact match's 27% impression share and 40% conversion share quantifies the precision premium. Exact match generates the highest conversion efficiency per impression in this dataset — the 1.48 efficiency ratio means each impression point delivers 1.48 conversion points. Broad match inverts this relationship: 46% of impressions from only 25% of keywords (an efficiency ratio of 0.76) illustrates the reach multiplication that Smart Bidding is designed to monetize. Broad match doesn't convert as efficiently per impression, but it surfaces 1.84x more impressions per keyword than exact match. The question is whether the account has the Smart Bidding signal to make those additional impressions cost-efficient. For performance metric frameworks, see Google Ads Metrics and KPIs. For SaaS accounts where keyword efficiency at scale is the central challenge, see SaaS Software PPC.

Broad Match Modifier Is Gone: What Phrase Match Does Now

Broad match modifier was discontinued in July 2021. Any "+keyword" BMM syntax remaining in an account now functions identically to phrase match. There are three match types in Google Ads, not four. Any documentation still describing BMM as a current match type option is describing a behavior deprecated more than four years ago.

What changed between old phrase match and the 2021 update is significant. Old phrase match required exact word order with no interruptions — a query inserting a word between two keyword terms would not trigger the ad. The 2021 phrase match drops this word-order rigidity in favor of intent: it shows for reordered queries if the intent is equivalent. This is something pre-2021 phrase match would not have done. As Google's language models have improved since 2021, the practical boundary between phrase and broad has narrowed further — phrase match today covers substantially more query variation than phrase match did in 2020.

Any "+keyword" BMM notation in an account functions as phrase match. There is no benefit to maintaining BMM syntax — cleaning it up to explicit phrase match syntax improves account readability without changing behavior. The match type column in Google Ads Editor no longer accurately reflects intent when BMM and phrase coexist as the same effective type.

MB Adv Agency has found that accounts still carrying hundreds of "+keyword" BMM entries gain clarity from a cleanup pass — not because behavior changes (it does not), but because the match type column in Google Ads Editor no longer reflects intent accurately when BMM and phrase coexist as the same effective type. For industry contexts where match type clarity is operationally critical, see Financial Services PPC, Fashion PPC, and PPC Consultant Chicago IL.

Google Keyword Planner: What the Data Actually Shows

Google Keyword Planner shows broad logarithmic volume ranges for accounts without active ad spend. The competition score does not measure organic ranking difficulty — it measures advertiser auction density. Both misreads are common; both distort keyword selection and budget planning decisions.

Accounts without active ad spend see only broad logarithmic volume ranges in Keyword Planner — for example, "1K–10K" or "10K–100K" rather than a specific monthly figure. Active accounts see narrower ranges and, exact monthly figures in certain active accounts. Google has not published a specific spending threshold that unlocks precise data; community reports indicate that even minimal active spend unlocks more granular ranges. Exact monthly figures are not guaranteed even in active accounts — the data depends on query volume levels and available data density for each keyword.

Suited for: initial keyword discovery and grouping; identifying seasonal volume trends across the year; estimating pre-launch budget ranges before a campaign has real data. Keyword Planner excels at the discovery phase — finding keyword themes, understanding relative volume between terms, and building initial keyword lists.

Not suited for: live search term data (60+ days of campaign history in an active account is more reliable once the account is running); competitive keyword gap analysis (Ahrefs and Semrush provide cross-domain keyword data that Keyword Planner cannot); CPC forecasting in competitive verticals (live auction data from the search term report and bid simulator is the only ground truth for actual CPC). For search term analysis beyond Keyword Planner, see Google Ads Audience Targeting and Google Ads Metrics and KPIs. For dental practices where local keyword volume data directly shapes budget planning, see Dental PPC.

Four Misconceptions That Cost Money

The most expensive keyword strategy errors in 2026 share a common root: applying match-type logic that was accurate before 2019 or before July 2021, but has since been superseded by Google's close-variant expansion, the BMM retirement, and the coupling of match types to Smart Bidding automation.

MB Adv Agency has found that the most impactful fix in a new client keyword audit is rarely adding keywords — it is fixing match type architecture and expanding the negative keyword list. Broad match with an undeveloped negative keyword strategy is the most common source of wasted spend in accounts MB Adv inherits. For high-CPC verticals where these errors compound fastest, see Personal Injury Law PPC and Dental PPC.

Frequently Asked Questions: Google Ads Keyword Match Types

Methodology

This pillar draws from Ahrefs keyword data, Google Ads Help documentation, and the Adalysis match type performance study. All numbers are source-attributed.

Data sources: Ahrefs Keywords Explorer (US, May 2026); Google Search Console (March–May 2026); Adalysis match type performance study (December 2025, 16,825 Search campaigns); Google Ads Help documentation current as of May 2026 (answers: 7478529, 9342105, 10195720, 10286719, 2453972, 6372658, 13389795); Search Engine Land; WordStream; Store Growers; groas.ai; groas.com; Ajala Digital. Google internal data on broad match conversion uplift (+25%/+12%) are published averages; Google has not disclosed methodology or sample size for these figures. Last updated: May 31, 2026. Reviewed by MB Adv Agency, May 2026.