Exact Match vs Broad Match vs Phrase Match: The Case for Staying Exact
A plumbing company in Denver ran Google Ads for six months. Their agency put them on broad match with Smart Bidding. They got clicks. Lots of clicks. They also paid for “how to fix a leaky faucet yourself,” “plumbing school near me,” and “free plumbing advice.” Their cost per lead was $340. When they switched every keyword to exact match, it dropped to $89. Same budget. Same city. Same service. Different match type.
That is not an edge case. That is broad match doing exactly what it is designed to do — find you traffic. Not the traffic you want. Traffic.
What the three match types actually do in 2026
Google has changed what these mean over the years. Here is where they stand now.
Broad match shows your ad for any search Google considers related to your keyword. “Related” is doing a lot of work in that sentence. If you bid on emergency plumber, broad match will show your ad for DIY pipe repair, plumbing apprenticeships, what does a plumber charge, and anything else the algorithm decides is in the neighborhood. Google calls this “finding new demand.” It is also how you pay for research traffic, competitor navigational traffic, and informational queries from people who will never buy.
Phrase match requires that the meaning of your keyword appear in the search query, in roughly the same order. If you bid on "emergency plumber Denver", you might show for affordable emergency plumber in Denver or best emergency plumber Denver CO. You will also show for emergency plumber Denver reviews — someone comparing options, not booking. Phrase match is tighter than broad, but “meaning-based” matching means Google still has room to interpret.
Exact match — in theory — shows your ad only when someone searches for your keyword or a close variant. In practice, Google has expanded close variants to include different word orders, implied words, and paraphrases. [emergency plumber] might match plumber emergency or urgent plumber. It is not perfectly exact anymore. But it is still the tightest control available. When you bid on exact match, you are making a deliberate choice about which intent you want to pay for. The other match types make that choice for you.
The case for staying exact
The conventional wisdom from Google, from agencies, from most of the industry: start broad, let Smart Bidding learn, scale what works. The logic sounds reasonable. The data does not support it.
Optmyzr analyzed performance across a large sample of Google Ads accounts and found that exact match outperformed broad match on CPA, ROAS, and CTR for the majority of accounts. Not by a little. Exact match delivered lower costs per acquisition, better returns on ad spend, and higher click-through rates — the three metrics that determine whether a campaign is actually working.
Here is why. When someone types [emergency plumber Denver] into Google, their intent is clear. They have a pipe problem, they are in Denver, they want a plumber now. When you pay for that click with exact match, you are paying for that specific intent. Every dollar goes toward someone who is likely to convert.
Broad match spreads that same dollar across the full range of what Google calls “related.” Some of those searches are buyers. Many are not. Studies have found that 50 to 73 percent of broad match spend goes to queries with no realistic chance of conversion when there are no hard guardrails in place. That is not waste you can optimize away — it is the mechanism working as intended.
What broad match waste looks like in practice
These are real search term categories that appear in broad match campaigns for high-intent commercial keywords:
- A company bidding on
business insurance quotepays foris business insurance required by law— an informational search from someone not yet ready to buy. - A software company bidding on
project management softwarepays forproject management certification online— a completely different product. - A dentist bidding on
dental implants costpays fordental implant failure stories— someone scared of the procedure, not booking one. - A home security company bidding on
home security systempays forhome security tips— a content search, not a purchase search.
Each of those clicks costs the same as a click from someone ready to convert. Broad match cannot distinguish between them unless you block every bad query with negative keywords — a job that is never finished, because Google keeps finding new “related” searches.
The Smart Bidding problem
Google’s response to all of this is Smart Bidding. The pitch: broad match + Smart Bidding learns from your conversion data and automatically avoids wasted spend. Google claims this combination drives 25% more conversions than exact match alone.
There is a catch. Smart Bidding requires at least 30 conversions per month to optimize properly. Most small and mid-size businesses do not hit that threshold. Google’s own documentation acknowledges this. Below 30 conversions, the algorithm is guessing. It does not have enough signal to distinguish a buyer from a browser, so it defaults to chasing volume — which is another way of saying it defaults to the same behavior as broad match without the learning layer that supposedly makes it valuable.
If you are running a service business with a $3,000 average contract value, you might close two or three clients per month from paid search. That is not 30 conversions. That is not enough data for Smart Bidding to do anything useful. What it will do is spend your budget while it waits for signal that never comes.
The 25% more conversions claim also comes with fine print: it assumes a high-volume account with consistent conversion data, correct attribution, and a well-structured campaign. It does not apply to most businesses running Google Ads.
When phrase match is acceptable
We are not absolutists. There are narrow cases where phrase match earns a place in a campaign.
Geographic variations you cannot fully enumerate. If you serve 40 different cities and the keyword behavior is consistent across all of them, phrase match for "emergency plumber [city]" can work — provided you have strong negatives for the terms you know are waste.
Discovery phase before you have search term data. When you are launching a brand-new campaign in a category you have no historical data for, phrase match can surface the exact queries you should be targeting. Run it for 30 days, mine the search terms report, convert the winners to exact match, negative out the rest.
Highly specific multi-word keywords where intent is stable. A seven-word technical keyword with narrow audience and clear commercial intent is unlikely to match broadly even in phrase mode.
In all three cases, the approach is the same: phrase match is a temporary tool for building toward exact. It is not a permanent strategy. The moment you have enough data to know which exact queries convert, you move to exact match and keep the control.
Broad match is never acceptable in a budget-constrained campaign where the cost of waste is visible. Which is most campaigns.
The control argument
The real reason to stay exact is not just the Optmyzr data. It is what control means when you are running a paid channel for a real business.
When every keyword is exact match, you know exactly which queries triggered your ads. You know which search terms are converting and at what cost. You can make decisions based on facts. You can adjust bids, pause underperformers, and scale winners with precision.
When you run broad match, you are making a tradeoff: you give Google the wheel and hope the destination is somewhere useful. Sometimes it is. More often, you are buying traffic that feels like progress — clicks, impressions, activity — while the queries that actually matter are buried in a search terms report you have to dig through to find.
No broad match bleeding your budget into irrelevant queries. No phrase match catching tangential traffic. You chose a keyword because the intent is right. Exact match ensures that is the only intent you pay for.
That is not a limitation. That is the point.
What this looks like in practice
A campaign built on exact match keywords looks different from a typical Google Ads account. Fewer keywords. Higher CPCs on individual terms. Smaller reach. Those are not problems to fix — they are features. You are paying more per click on a smaller set of searches because every search on that list is worth paying for.
The math works out. Lower conversion volume, but a much higher percentage of clicks that convert. Lower wasted spend. Better CPA. Better ROAS. A campaign that is genuinely profitable rather than one that looks busy.
When we build campaigns for clients, we start with exact match and stay there. We find the keywords where intent is unambiguous — the searches that mean someone is ready to buy, not just curious. We do not try to manufacture scale by loosening match types. We find the right searches and own them.
If you want to see what this looks like applied to your specific account — which keywords have the right intent, what your current match type mix is costing you, and where the actual opportunities are — the audit is a good place to start. It takes your Google Ads data and shows you exactly where you stand.