Google Ads Keyword Match Types: Choose Broad, Phrase or Exact

Choose exact match when you want to concentrate on a defined customer need, phrase match when useful variations of that need deserve exploration, and broad match when you are ready to test wider demand with dependable measurement. The decision should reflect what your business can serve and afford to learn.
We judge a match-type choice by the suitable inquiries or purchases it produces. More searches and cheaper leads can still leave you with fewer worthwhile opportunities. This guide explains the settings, gives you a decision worksheet and works through a fictional test where the headline lead cost points toward the wrong conclusion.

What broad, phrase and exact actually mean
A keyword is a targeting term you add; a search term is the customer's query. Google's keyword matching guidance describes three positive match types:
Exact: searches with the same meaning or intent. Format: [office cleaning service].
Phrase: searches that include the keyword's meaning, including an implied meaning. Format: "office cleaning service".
Broad: related searches, including searches beyond the keyword's direct meaning. Format: office cleaning service.
Broader types include the searches covered by equivalent narrower types, plus more. Putting the same keyword into three match types does not create three separate groups of search demand.
Google's close variants guidance explains why exact is not a letter-for-letter restriction. Paraphrases and similar intent can qualify. Phrase does not simply mean an unchanged phrase with extra words around it.
The labels describe eligibility. They do not confirm that a customer can afford your service or that your landing page answers their question.
Choose the starting point from your constraints
Write one customer need: “An office manager seeking recurring cleaning for premises in our service area.” Then list the boundaries: commercial work, recurring contracts, supported locations and realistic availability.
Use those boundaries to choose a starting approach:
Defined need, little room for adjacent work: consider exact for priority service terms. The tradeoff is less scope to discover other useful ways customers describe the need.
Defined service, varied customer language: consider phrase where you can assess the resulting variations. Keep the ad and page specific enough to explain the offer.
Expansion question and room to investigate: consider a broad-match test with clear spending and outcome rules.
A small budget strengthens the case for prioritising what you need to learn first. It does not establish one correct match type. Decide which uncertainty is worth paying to resolve.
Pass four checks before testing broader reach
Google's matching guidance calls for broad match to be used with Smart Bidding, Google's conversion-focused automated bidding approach. Pairing those settings still leaves important business decisions with you.
Usable measurement: test whether the recorded action represents a real inquiry or purchase. A clicked button and a successfully submitted request are different events.
A qualification rule: specify the service, location and other essential conditions. Have someone assess inquiries consistently.
An exploratory allowance: agree the amount you can spend on this question before another approval or review.
A decision point: define the review date, the time leads need to mature and the evidence that would justify continuing.
If you cannot distinguish a genuine inquiry from an accidental action, improve the measurement first. A wider search pool will not resolve uncertainty about what the campaign is producing.
A fictional test: cheaper leads, weaker economics
Meridian Office Care is a fictional business selling recurring office-cleaning contracts. Its team defines a qualified inquiry as a business premises within its coverage area seeking recurring work it has capacity to quote. It chooses a provisional ad-cost ceiling of $110 per qualified inquiry from its own acquisition plan.
The team has a focused baseline and allows $600 to investigate broader reach. It holds the service offer, page and qualification rule steady and records differences in campaign conditions. The following invented figures illustrate a review, not a forecast or a statistically reliable result:
Focused baseline: $600 spend, 12 unique inquiries, six qualified. Ad cost per inquiry is $50; ad cost per qualified inquiry is $100.
Broader-reach test: $600 spend, 30 unique inquiries. At the first review, four qualify, 24 do not and two still need assessment.
At $20 per inquiry, the broader group appears cheaper. The team nevertheless stops additional exploration when its $600 allowance is used and finishes assessing the open inquiries. One qualifies; one does not. The completed group has five qualified inquiries at $120 each: $600 divided by five.
The completed comparison does not support expansion under the team's $110 rule. It keeps the focused activity under review and investigates why the broader group attracted unsuitable work before authorising another test. The threshold and allowance are choices for this example, not industry benchmarks.
The extra 18 inquiries are not evidence of extra customers. The groups can differ in timing and traffic; this comparison cannot isolate match type as the cause. Follow qualified opportunities through to sales and include other acquisition costs before judging profitability.
Diagnose the mismatch before narrowing everything
In the fictional review, “office cleaning jobs” signals an employment need, while “weekly office cleaning quote” suggests a service inquiry worth assessing. “Clinic cleaning contract” needs a capability check before the business decides whether it belongs. These are intent examples, not promises about which match type will serve an ad.
Separate three problems:
Wrong need: the person wants employment, supplies or a service you do not offer. Review targeting and exclusions using our guide to choosing negative keywords without blocking useful demand.
Right need, wrong expectation: the ad attracts the intended customer but leaves them expecting an unsupported price, service or next step. Use our guide to writing Search ads that match your offer to inspect the promise and destination.
Right need, incomplete outcome: a suitable inquiry is waiting for contact or a quote. Resolve the handoff before labelling that demand unproductive.
Narrowing may be appropriate, but first name the problem it is intended to solve.
Complete a match-type test worksheet
Use this short brief before changing the account:
Customer need: what the searcher wants and the service or product that answers it.
Current evidence: available search information, spending and assessed customer outcomes.
Proposed change: the keyword group and match-type change, with the reason.
Learning question: for example, whether wider reach finds suitable inquiries within your agreed cost ceiling.
Limits: exploratory spend, essential exclusions, unchanged parts of the offer and a review date.
Outcome owner: who assesses each inquiry, by when, and how unresolved cases stay visible.
Decision: continue, narrow, fix the journey or finish assessment before another test.
Keep the change small enough to explain. If bids, offer, page and targeting all change together, record that the comparison evaluates the combined change.
Make the next decision about useful demand
Keep a setting when the evidence supports its role. Narrow repeated irrelevant demand. Test wider reach when the business can fund and assess the exploration. Treat an inconclusive sample as a reason to decide what evidence is missing, rather than declaring a universal winning match type.
Our Google Ads management service can help you make that choice. Bring your priority services, current keyword settings, recent search concerns and anonymised lead outcomes to a strategy call. We can discuss whether the next useful move is tighter targeting, a clearer offer or a bounded expansion test.
Comments