GOOGLE ADS KEYWORD MATCH TYPES EXPLAINED

The honest version of google ads keyword match types explained starts with an admission: the way these match types work in 2026 is meaningfully different from how they were documented three years ago, and the gap between documentation and reality is wide enough to cost real money if you are still operating on older assumptions. I have been running accounts where all three match types coexist, broad, phrase, and exact, and the interaction between them is not what most guides describe. This article is built around a specific 90-day test I ran on a B2B SaaS account, where I deliberately isolated the three google ad keyword match types into separate campaigns and tracked CPA, CTR, and conversion volume independently. The numbers surprised me in a few ways. I will show you exactly what happened and what I changed afterward.

Before the data, a fast definitional pass, not because you need a dictionary entry, but because the definitions have shifted. According to Google's official documentation on keyword matching options, the three match types dictate how closely a keyword must match the user's search query for an ad to enter the auction. The broad match type is the loosest, it includes searches that include semantically related terms the keyword itself does not contain. The phrase match type requires the meaning of the keyword to appear within the query in sequence. Exact match targets queries with the same meaning or intent as the keyword, including close variants. What that last clause, "including close variants", means in practice is something I will get into in detail below, because it is where most practitioners underestimate how much reach exact match now has. One thing worth noting from the start: every match type controls the gap between the keyword you write and the actual search that triggers your ad to show, and in 2026, that gap is wider for all three types than it used to be.

Broad match: the match type most accounts misuse

Broad match is the default google ads keyword match type assigned to every keyword you add without a syntax modifier. That fact alone explains why so many accounts bleed budget on irrelevant searches: advertisers add keywords without specifying a match type and end up in the broad match type without intending to. When google ads keyword match types explained correctly, broad match is not a problem, it is a tool with a specific use case. The problem is that most accounts use it accidentally rather than intentionally. The critical distinction is between deploying the broad match type as a deliberate reach strategy toward a well-defined target audience, versus landing in it by default and watching match keywords your ads trigger on drift far outside anything commercially relevant.

What broad match actually matches in 2026

In 2026, the broad match type does not just expand to synonyms and related terms. It uses signals from the entire account context, landing pages, other keywords in the ad group, user search history, and audience data, to determine what searches that include sufficient relevance deserve an entry into the auction. A broad match keyword like project management software can cause an ad to show on queries like "best tools for team coordination," "alternatives to Asana," or "software to track tasks remotely." None of these contain the keyword. That reach is exactly the point, but it also means that match types broad without a structured negative keyword list paired with broad match is one of the fastest ways to discover irrelevant searches you did not know existed in your account. The search terms report under a broad match campaign is always illuminating, and sometimes alarming. I pull it weekly on every account where broad is running, without exception, because the gap between what you intended and what the broad match type actually decided to surface can be substantial.

When broad match earns its place

The scenario where I actively choose the broad match type is when I am running a Smart Bidding strategy, Target CPA or Target ROAS, and the account has sufficient conversion history for the algorithm to work with. Google's own guidance makes this connection explicit: broad match paired with Smart Bidding allows the system to use auction-time signals that manual or enhanced CPC setups cannot access. In the 90-day test I describe below, the broad match campaign was the only one using Target CPA. The exact and phrase campaigns ran on manual CPC with bid adjustments. That difference in bidding strategy is a confounding variable I want to be transparent about, it means the comparison is not perfectly clean, but it does reflect real-world conditions, because in practice you would not run broad match without a Smart Bidding layer.

Phrase match: the middle ground that is shrinking

The phrase match type is the google ad keyword match type I find hardest to justify in 2026. It used to occupy a clear position between the openness of broad and the precision of exact: it required the keyword phrase to appear in sequence within the query, which gave it predictable reach without the unpredictability of broad. That position has eroded from both sides. Exact match has expanded through close variants. Broad match with Smart Bidding has become smarter. What is left for phrase is a narrower middle lane than it had three years ago, and in several account types I manage I have phased it out entirely in favor of a broader set of exact match keywords supplemented by broad match in a separate campaign.

Where phrase match still works

The use case where the phrase match type still performs reliably for me is local service businesses. A keyword like "plumber in Chicago" in phrase match will capture "emergency plumber in Chicago," "best plumber in Chicago north side," and similar local intent variants while ensuring your ad does not show on "plumber school Chicago" or "Chicago plumber licensing exam", the kinds of tangential queries that broad match attracts. In these accounts, the geographic qualifier in the keyword acts as a natural boundary that phrase match enforces. For national or international campaigns with no geographic anchor in the keyword itself, phrase match provides much less containment and I start questioning whether it is adding value over a well-structured exact match pool. For anyone looking at google ads keyword match types explained from a strategic angle, phrase match in 2026 is not a default choice, it is a deliberate one justified by specific account characteristics.

The phrase-to-broad cannibalization problem

One consistent finding in accounts that run phrase and broad on the same keywords simultaneously, even in separate ad groups, is query overlap. The search terms report regularly shows the same query being matched by both the phrase match keyword and the broad match keyword, sometimes in the same campaign, sometimes across campaigns. When this happens, Google's ad rank determines which keyword wins the auction, which means your most carefully written ad group copy may be bypassed by the broad match variant. Adding negatives is the practical mechanism to prevent your ads from showing on queries that the wrong match type picks up, but negatives alone are not enough without structural campaign separation. Understanding how to change keyword match type in Google Ads at scale, particularly when migrating an existing structure, is the first operational skill this kind of setup requires.

Exact match: still the tightest control, but not what it was

Exact match is the google ads keyword match type I recommend as the starting point for any new campaign where budget is limited and conversion data is scarce. The reasoning is straightforward: exact match gives you the clearest signal about what is working. When a conversion comes from an exact match keyword, you know the query that triggered the ad to show. When it comes from a broad match keyword, you need to dig into the search terms report to find out, and in accounts with high query volume, that attribution becomes noisy quickly. Exact match is also the easiest type to use as a meaningful control group: because the searches that include your keyword (or very close variants) are tightly bounded, you can make bidding and copy decisions with higher confidence in the data behind them.

Close variants: what exact match actually includes

The expansion of exact match through close variants is the single most important evolution in google ads keyword match types explained for practitioners who have been running accounts for more than a few years. An exact match keyword like [keyword match types google ads] will also trigger on queries like "google ads keyword match types," "google ad keyword match types," and "match types in google ads keywords", reorderings, plural variations, and functional word changes that Google determines carry the same intent. According to Google's documentation on how keyword matching works, the system interprets the query's meaning rather than its literal wording before matching it to keywords. In practice, this means exact match today reaches a meaningfully wider set of searches that include intent-equivalent variants than the syntax implies, and advertisers relying on exact match for strict spend control need to verify reach assumptions against the actual search terms report rather than the theoretical keyword definition.

Exact match as the baseline for expansion

The way I use exact match in new accounts is as a learning layer, not a permanent ceiling. I start with exact match on the highest-priority keywords, accumulate conversion data for 60 to 90 days, and then use the search terms report to identify high-performing queries that the exact match keyword is already capturing through close variants. Those queries become new exact match additions. Only after this phase do I consider introducing phrase or broad match, and when I do, the existing exact match keywords act as a control group that I can compare against. This sequencing is one of the most operationally useful things I can recommend for anyone trying to understand google ads keyword match types explained as a working framework rather than a theoretical taxonomy.

90 days of data: what actually happened across all three match types

The 90-day test I ran was on a B2B SaaS account in the project management software space. The account had 18 months of conversion history and was spending approximately €8,000 per month before the test. I restructured it into three parallel campaigns, one per google ad keyword match type, with the same keyword set, matched budgets, and identical ad copy. The only intentional variables were match type and bidding strategy: broad on Target CPA, phrase and exact on manual CPC with equivalent target bids. Negative keywords were shared across all three campaigns to prevent the same irrelevant queries from burning budget in multiple places.

The numbers after 30 days

At the 30-day mark, the pattern that emerged matched my expectations, but the magnitude was larger than anticipated. The exact match campaign delivered 41 conversions at a CPA of €62. The phrase match campaign delivered 67 conversions at €89 CPA. The broad match campaign delivered 94 conversions at €118 CPA. Volume scaled with match type looseness, and so did cost per conversion. CTR told a different story: exact match had a 7.4% CTR, phrase had 5.1%, and broad had 3.2%. The exact match ads were reaching users with higher intent, and those users clicked more often once they saw the ad. The broad match campaign was generating impressions from a wider query set where the ad was less directly relevant, lower CTR, higher volume, higher CPA. So far, google ads keyword match types explained in a fairly conventional way. The interesting part came later.

The shift between day 31 and day 90

Between days 31 and 90, the broad match campaign's CPA dropped from €118 to €84, a 29% improvement, while conversion volume grew from 94 to 156. The Target CPA algorithm was learning which query patterns converted and was shifting spend toward them. The exact match campaign's CPA was stable at €61–64 throughout the period, and volume grew modestly from 41 to 52 conversions. The phrase match campaign barely moved: €87 CPA at day 90, 71 conversions. The conclusion I drew was that phrase match provided neither the control of exact nor the eventual efficiency of Smart Bidding on broad, it sat in a middle ground that did not justify its dedicated budget once the broad campaign had enough data to optimize. By day 90, I had reallocated 60% of the phrase campaign budget to broad, keeping phrase only for the local intent keywords where it had shown clear containment advantages. The broad match campaign's performance over 90 days aligned with what I consistently observe: the google ad keyword match types hierarchy matters most in the first 30 days, before Smart Bidding has accumulated enough signal to compensate for broad match's inherent noise.

What I changed after the test

The structural change I implemented after this test is the one I now recommend as a default for accounts with sufficient conversion history: one exact match campaign on manual CPC as the control and primary profit driver, and one broad match campaign on Target CPA as the volume and discovery layer. Phrase match survives only where there is a specific containment reason, typically local terms or competitor brand terms where exact intent is critical but query variation is predictable. The negative keyword infrastructure connecting all campaigns is the mechanism that prevents the broad campaign from cannibalizing the exact campaign's most valuable queries. For anyone approaching this structure for the first time, I cover the full negative keyword architecture in the guide on google ads negative keywords best practices, it is the piece of this setup that most practitioners underinvest in and that determines whether the broad campaign helps or hurts the overall account efficiency.

How to structure campaigns when all three match types are running

Running google ads keyword match types explained as a coexistence strategy, not a choice between types, requires explicit structural separation to work. The core problem with mixing match types in the same campaign or ad group is that Google's keyword selection algorithm, when faced with multiple eligible keywords for the same query, does not always pick the one you expect. In theory, the most specific match type wins. In practice, bid levels, Quality Score, and Ad Rank interact in ways that can surface the broad match type over an exact match keyword in the same account. Separate campaigns with separate budgets eliminate the ambiguity and make it far easier to measure which match type is actually contributing value to your specific target audience.

The campaign separation model I use

The structure I apply across the accounts I manage is: Campaign A contains exact match keywords on manual CPC, targeting the highest-intent, highest-converting queries identified from historical search term data. Campaign B contains broad match keywords on Target CPA, targeting the same thematic space but intentionally left open for the algorithm to explore. Both campaigns share a negative keyword list, applied at the account level, that blocks the clearest irrelevant categories: competitor names unless there is a dedicated competitor campaign, navigational queries for other brands, and job-seeking terms. Campaign A additionally uses exact match negatives for the specific queries Campaign B is allowed to match, preventing the broad campaign from cannibalizing the exact campaign's best performers and ensuring your ad from the control campaign is the one showing on your highest-value searches. This model is detailed enough that it deserves a dedicated walkthrough, which you will find in the guide on google ads broad match keywords and when they actually work.

Budget allocation across match types

Budget allocation between match type campaigns is not a fixed ratio, it depends on the account's conversion volume and the Smart Bidding algorithm's learning status. My starting point is 60% to exact, 40% to broad, with zero budget for a separate phrase campaign unless there is an explicit reason. As the broad campaign's CPA stabilizes and approaches the exact campaign's CPA (which in my experience takes 45 to 90 days), I shift budget toward broad incrementally. The signal that the shift is safe is not just CPA convergence, it is also the search terms report showing that the broad campaign is consistently matching high-intent queries rather than exploratory or informational ones. When the broad campaign's top-performing search terms are the same queries that would appear in an exact match set, the algorithm has effectively learned to behave like a smart exact match layer, and that is when it earns a larger share of budget.

The role of negative keywords across the structure

The infrastructure that makes this structure work is negative keywords applied with precision across all match type levels. Without them, the broad campaign will eventually match queries that belong to the exact campaign's core keyword set, split the impression share, and dilute the data signal in both campaigns simultaneously. Negatives are what actually prevent your ads from showing on irrelevant searches at scale, not match type alone. I add negatives in three layers: account-level negatives for irreversible exclusions, campaign-level negatives for structural separations between match type campaigns, and ad group level negatives for intent separations within campaigns. The entire negative keyword architecture for a structure like this is covered in the dedicated guide on google ads negative keywords best practices, for anyone working on a fresh account setup, that guide is the right place to start before touching match types at all.