What are keyword match types

A keyword is not what the user types. It is the term an advertiser adds to an ad group to tell Google Ads what it wants to compete for. What the user types is called a search term, and it almost never matches the keyword letter for letter. The match type is the rule that joins the two: it defines how much distance the system allows between the real search and the keyword you bought.

It helps to separate it from other settings it gets confused with. It does not change the ad text, it does not alter the bid and it does not decide position. It only delimits the set of eligible searches. Inside that set, the auction still rules.

The current system has three options. Until 2021 there was a fourth, the broad match modifier, written with a plus sign in front of every required word. Google retired it that year and moved its behaviour into phrase match. Anyone who learned to target with that modifier now works under different rules, even though the interface looks similar.

The other change is conceptual. All three options stopped comparing strings of text and started comparing meanings. Even exact match, which sounds literal by name, covers searches with the same intent even when the words change.

Broad match is the default and is written without any symbol: running shoes. It triggers the ad on searches related to the keyword, including some that do not contain its direct meaning. To decide this, Google Ads draws on signals from the account itself: the content of the landing pages, the other keywords in the ad group and the user's recent search activity.

Phrase match is written in quotation marks: "running shoes". It shows the ad on searches that include the meaning of the keyword, and that meaning may be implied. Word order matters when it changes the sense, so a phrase such as removals from Barcelona to Madrid should not trigger on the reverse search.

Exact match is written in square brackets: [running shoes]. It offers the most control and still does not require a literal coincidence: it covers searches with the same meaning or the same intent as the keyword.

All three types allow close variants that cannot be switched off. These include typos, singulars and plurals, words from the same stem, acronyms, accents, reorderings that do not change the sense, prepositions and articles added or removed, and synonyms such as swimming trunks against swimsuit.

When several keywords in the account fit the same search, they do not compete with each other. Only one enters the auction: the one with the highest Ad Rank. Two nearly identical broad match keywords are treated as duplicates, so the system uses only one and performance is not split.

The choice of match type decides how much budget you hand over to Google's automatic interpretation. An account on broad match gains volume and loses predictability; an account on exact match caps spend and leaves traffic on the table. That is the real decision, and it is taken ad group by ad group, not once for the whole account.

The instrument for reviewing it is the search terms report. It shows what people typed before they saw the ad, not which keyword was bought. It is the only source that confirms whether broad match is bringing useful queries or noise.

Here comes the point that tends to surprise: that report is not complete. Google leaves out search terms without enough query activity in order to meet its data privacy standards, a restriction tightened in 2020 and readjusted in 2021. Low-volume queries are grouped together and not listed one by one.

The practical consequence is uncomfortable. Part of broad match spend happens on searches you will never see written down, and that invisible share grows the more open the match type is. The decision is not only about reach, it is also about auditability.

• Start with phrase and exact match in the ad groups that already convert, and keep broad match for discovering new demand on a capped budget.
• Isolate broad match in its own campaigns, with its own budget and bid, so that its spend is readable without cross-checking reports.
• Review the search terms report weekly during the first month of a new campaign, and at least monthly after that.
• Feed broad match with precise landing pages: the system uses their content as a signal, and a generic page widens the interpretation.
• Avoid duplicating the same keyword across ad groups with different match types unless you genuinely differentiate the ad and the landing page.
• Compare total campaign spend with the spend added up in the search terms report: the difference estimates how many queries stay hidden.

• Assuming exact match means literal coincidence, and never checking which searches actually triggered those keywords.
• Getting the notation wrong: typographic quotation marks or a missing square bracket turn the keyword into broad match with no warning.
• Running the account on the logic of the broad match modifier, retired in 2021, and expecting the level of control it used to give.
• Reading the search terms report as a complete record of spend, when it leaves out low-volume queries.
• Mixing broad, phrase and exact match in the same ad group and crediting the result to one type without being able to separate it.