Broad vs Phrase vs Exact Match in Google Ads (2026)

Your Google rep will tell you broad match plus Smart Bidding beats a tight exact-match list. At a few hundred conversions a month, they may be right. At eight conversions a month, they are describing a machine-learning system with nothing to learn from — and you are the training data it burns through.

This post is about the inclusion decision: which match type lets a search into your auction. Keeping the wrong searches out is a separate job, and it lives in our negative keywords strategy for service businesses.

What do broad, phrase and exact match actually mean in 2026?

There are exactly three positive match types, and they nest inside each other. Google's keyword matching documentation puts it plainly: 'a broad match keyword will match all the searches of equivalent phrase and exact match keywords, plus additional related searches.' So this is not three tools. It is one dial, and you are choosing how far to open it.

Verdict: for a service business under 30 conversions a month, exact match wins. Google itself concedes exact match 'gives you the most steering over who views your ad.' Steering is exactly what you need when every wasted click is 2% of your monthly budget. Broad match wins later, on volume — not on day one.

Why does exact match no longer mean exact?

Because close variants are switched on for every match type and cannot be turned off. Google's close variants definition is unambiguous: 'By default, all keyword match types are eligible to match to close variants. There's no way to opt out.'

For an exact match keyword, Google lists six categories of close variant it will match: misspellings and plurals, reordered words with the same meaning, added or removed function words, implied words, synonyms and paraphrases, and 'same search intent.' Google's own example: the exact keyword [bathing suits] can show on searches for 'swimming suits.'

Translate that to a plumber. Bracket [emergency plumber] and you are not buying that string. You are buying Google's model of that intent — which may include 'urgent plumber,' 'plumber open now,' or something it decides means the same thing. Exact match in 2026 is a narrow intent bucket, not a literal string match. Plan for slippage even at your tightest setting.

At what monthly conversion volume does broad match start working?

Use 30 conversions in 30 days as the floor, because that is the number Google uses. Its Target CPA documentation says: 'For evaluation, we recommend you measure performance for the last 30 days, including at least 30 conversions.' Same page, higher up, in the section on how the strategy works: 'Advertisers can start using Target CPA with no conversion history.'

Read those two sentences together. Google will happily let you start with zero data, but its own guidance says you cannot judge the result until you have 30 conversions in a 30-day window. Starting and evaluating are not the same thing, and only one of them is free.

Do the arithmetic for a plumber booking eight jobs a month from paid search. Thirty conversions takes you roughly four months to accumulate. By the time you have a readable window, the season changed, your CPCs moved, and you have spent four months of budget buying a verdict you could have bought in one month on exact match.

Why is broad match a spending permission slip at low volume?

Because part of what broad match spends on is invisible to you. Google's search terms report documentation states that 'some search terms that don't have enough query activity are omitted from the search terms report in order to keep with our standards on data privacy.'

That rule is fine at enterprise scale, where the hidden tail is a rounding error. It is not fine at your scale. If your CPC is $30 and you spend $2,000 a month, you are buying roughly 66 clicks. Broad match scatters those clicks across a long tail of one-off queries — precisely the queries most likely to fall under the reporting threshold and never appear in your report.

So the honest description of broad match on a small account is this: you authorize Google to spend your budget on searches it will not always show you, judged by a model that does not yet have enough conversions to know what a good search looks like. That is a permission slip, not a strategy. Get your Google Ads budget working on searches you can name first.

Which match type should you start a new service campaign with?

Start with 10 to 20 exact match keywords in one campaign, add phrase only after four weeks of clean data, and add zero broad keywords until you clear 30 conversions a month twice in a row. That is the whole rule.

Build the exact list from the money queries you would pay for by name: [emergency plumber austin], [water heater repair], [burst pipe repair]. Not [plumber]. Not [plumbing]. Head terms attract price shoppers and homework; the service-plus-problem queries attract people with a wet floor.

Phrase match is the middle setting, and it earns its place once you have four weeks of search terms showing which themes convert. It is the right default for a local service business that has some data but not 30 conversions a month — enough reach to find new query shapes, tight enough that you can still explain every dollar. When we run paid ads for service businesses, that is the shape we start with: narrow, explainable, and expanded only on evidence.

How often should you read the search terms report?

Weekly for the first eight weeks, then every two weeks once the query mix stops surprising you. On a small account this is a 15-minute job, and skipping it is how a $2,000 month turns into a $2,000 donation.

Turn on the 'Keyword' column — Google notes it does not show by default — and the 'Match type' column. Match type is the one people misread. Google's documentation is explicit that 'keywords with broader match types can still match search terms in narrower ways': its example shows the broad keyword 'purple flowers' logged as an exact match search term when someone searched exactly that, and as a broad match search term when they searched 'pink flowers.'

So the match-type column describes the search term's relationship to your keyword, not the setting you chose. Read it as a spectrum of how far Google wandered. The rows that wandered furthest are your negatives list and your evidence for whether broad match deserves a second campaign.

How do match types and smart bidding change each other?

They are a package deal, and Google says so outright: 'It's critical to use Smart Bidding with broad match,' and 'Smart Bidding works best with broad match because the wide set of searches allows it to learn what works best for you.' Broad match on manual CPC is the worst configuration in the account — maximum query variance, one fixed bid.

The chain runs one way. Smart Bidding needs conversions. Broad match needs Smart Bidding. So broad match needs conversions, and if you do not have them, every link in that chain is bluffing. Match type and bid strategy have to be chosen together — the bid strategy automation decision is the other half of this one.

One more thing the model cannot fix: bad conversion data. If your 'conversion' counts every form fill including spam and every 12-second call, Smart Bidding will faithfully learn to buy more spam. Clean the conversion definition before you widen the match type. Garbage signal at 30 conversions a month is worse than no signal at all, because it feels like proof.

What should you do this week?

Pull the last 30 days. Count conversions — real ones, booked jobs or qualified calls, not clicks on the phone icon. Under 30? Pause every broad keyword today, keep exact and phrase, and re-read your search terms. Over 30 for two months straight? Broad match earns a test, in its own campaign, with a budget you can afford to lose.

If you would rather have someone else make that call with your numbers in front of them, our paid ads service is month-to-month with no lock-in, and any channel that has not produced qualified leads in 90 days gets cut. Get my free audit and we will start with your search terms report.