Search terms vs keywords: the Google Ads gap

A keyword is what you bid on. A search term is what someone typed. These are different things, often very different, and the gap between them is the single most important concept in Google Ads.

If you understand this distinction, the rest of Google Ads makes sense. If you do not, almost nothing about the search terms report or negative keywords will feel intuitive.

The simplest possible version

You bid on the keyword plumber phoenix. The matcher serves your ad against the search term emergency plumber phoenix az because of broad-match expansion. Two different strings. One you wrote, one a user typed.

The keyword is your input. The search term is the matcher's output. The match type you chose for the keyword (broad, phrase, exact) controls how aggressively the matcher will turn one keyword into many search terms.

That is the whole concept. The rest of this piece is the implications.

Why this distinction was created

Originally, online advertising required exact keyword matching. If you bid on "plumber" and the user typed "plumbers," your ad did not show. If they typed "plumber service," your ad did not show. This was unworkable for advertisers because users type queries in thousands of variations on the same intent.

Match types were the response. Phrase match let you bid on "plumber service" and serve against any query containing those two words in that order. Broad match let you bid on a concept and serve against synonyms and close variants. Each level of expansion traded precision for reach.

The asymmetry that makes this confusing: the advertiser thinks in terms of keywords (the input), while the spend and the report are organized around search terms (the output). The translation between the two happens inside the matcher and is partially hidden.

Keyword

what you write in the bidding screen

Where it lives

The keywords screen

Examples

plumber phoenix, "emergency plumber", [ac repair]

Match type

Broad, phrase, or exact

Quantity

Tens to thousands per account

Search term

what a real user typed

Where it lives

The search terms report

Examples

"emergency plumber phoenix az", "plumber salary 2026", "free ac repair near me"

Match type column

Shows how the matcher classified the match

Quantity

Often 2-6x more than keywords

What the match type column in the report actually means

When you read the search terms report, every row has a "Match type" column. This is showing you not how you bid, but how the matcher classified the match.

If you bid on the broad-match keyword plumber phoenix and the user typed "plumber phoenix," the row shows up as "Exact match." Wait, exact match? But you bid broad. The "Exact match" classification means the actual query exactly matched your keyword, even though your keyword's match type was broad.

If the same broad-match keyword caught "emergency plumber phoenix az," the row shows up as "Broad match" because the query expanded beyond your literal keyword.

The takeaway: the report's match type column tells you how loose the match was for that specific query. It does not echo what match type you set for the keyword. This is information about the matcher's behavior. Your own keyword settings live elsewhere.

Where the gap matters most

The gap between keywords and search terms matters in three specific ways.

Negative keyword decisions. When you read the report and decide to add a negative, you are blocking against the search-term side. The keyword side is unaffected. The negative salary blocks any query containing the word "salary," regardless of which of your keywords got matched to that query. This is also why one universal negative can clean up junk across many keywords.

Performance attribution. A keyword that "performs well" might be performing well because the matcher is sending it good search terms, or because the keyword's literal text is itself converting. Two different stories. The honest version requires looking at the search terms breakdown for that keyword rather than relying on the keyword's rolled-up metrics alone.

Bid optimization. When you raise or lower a bid on a keyword, you are changing the bid for every search term the matcher routes to that keyword. If 80% of those search terms are bargain hunters and 20% are real buyers, raising the bid amplifies both. The fix is structural (split the keyword by intent into separate ad groups, or apply negatives) rather than tactical (just changing the bid).

A worked example

Imagine a residential plumbing campaign with these keywords:

plumber                      [broad]
"emergency plumber"          [phrase]
[24 hour plumber phoenix]    [exact]

Three keywords, three match types. Over thirty days, the search terms report shows the following queries served against this campaign:

Search terms · last 30 daysillustration

	Search term	Match	Clicks	Cost	Conv.
	emergency plumber phoenix	broad	24	$184	4
	free plumber repair	broad	18	$72	0
	emergency plumber near me	phrase	22	$176	3
	24 hour plumber phoenix	exact	14	$112	2
	plumber jobs phoenix	broad	11	$44	0
	best plumber tempe az	broad	9	$72	1

Walk through what happened.

The exact-match keyword [24 hour plumber phoenix] produced one search term: "24 hour plumber phoenix." Exact match is tight; it produced exactly the literal query.

The phrase-match keyword "emergency plumber" produced "emergency plumber near me." Phrase is looser; it allowed "near me" to be appended to the phrase.

The broad-match keyword plumber produced four search terms: "emergency plumber phoenix" (which also matched the phrase keyword above, but the matcher chose this keyword), "free plumber repair," "plumber jobs phoenix," and "best plumber tempe az." Broad is loosest; the matcher pulled in queries that share a single word with the keyword.

The two negative-flagged rows ("free plumber repair" and "plumber jobs phoenix") came in via the broad keyword. They are why broad match needs vigilance: the wider the match type, the more junk slips through.

The "added/excluded" column in the report

Each search term row has an Added/Excluded column. This is where the relationship to your keywords becomes visible.

Added. This search term has been added as an explicit keyword. So if "emergency plumber phoenix" is bid on as a keyword and also appears as a search term, the row shows Added. The matcher routed the query to your explicit keyword rather than expanding from a different one.

Excluded. A negative keyword is in place that blocks this query. Useful for confirming negatives are working.

None. Most rows. The search term came in via match-type expansion from a keyword you do not explicitly have. This is the column to read for finding new keyword candidates.

A common workflow built on this column: filter for rows where Added/Excluded is "None" and conversions is greater than zero. These are the converting queries you are not explicitly bidding on. Each is a candidate to add as an explicit keyword (usually exact match) so you have direct bid control over a query that is paying off.

When to bid on a search term as a keyword

If a search term is converting and you are not bidding on it directly, you can promote it to a keyword. Should you?

Yes, when:

• The search term has converted multiple times over a sustained period.
• You want tighter bid control over that specific query (often exact match).
• The search term came via broad match and you want to stop relying on broad-match expansion to keep matching it.

Probably not, when:

• The search term has only converted once or twice. Statistical noise.
• The search term is a close variant of an existing keyword. Adding it would create internal competition.
• Adding it to one ad group would cannibalize another ad group's traffic.

The Added/Excluded column makes this workflow possible because you can tell at a glance which converting queries are not yet explicit keywords.

Why understanding this saves the most time

Most operator confusion in Google Ads traces back to conflating keywords and search terms. Examples we have heard:

• "I added the negative keyword 'free' but a query with 'free' still showed up." (The negative was added to the wrong campaign or the wrong level.)
• "My exact-match keyword is showing up as 'broad match' in the report." (The keyword is exact; the query that matched was a close variant, which the report calls broad.)
• "Why am I bidding on this query? I never added it." (You did not; the matcher served against it via broad expansion of a different keyword you did add.)
• "I want to bid on this exact query as a keyword." (You can; it would still produce search terms that vary from the literal keyword via close-variant matching.)

Each of these confusions evaporates once the keywords-versus-search-terms distinction is solid. The mental model becomes: I write keywords, the matcher produces search terms, the report shows me what got matched, and I adjust both sides until the search terms look like queries I want to pay for.

That is most of Google Ads, in three sentences.