---
title: "Google Ads Match Types: Broad, Phrase and Exact Compared"
description: "Google Ads match types compared: what broad, phrase and exact let through, how close variants work and which one pays at your margin."
canonical: "https://www.productmetrics.io/blog/google-ads-match-type"
pageType: article
language: en
publisher: "Product Metrics"
author: "Berend Vrakking"
datePublished: 2026-10-30
dateModified: 2026-10-30
image: "https://www.productmetrics.io/og/blog/google-ads-match-type.png"
---

> Content index: https://www.productmetrics.io/llms.txt

# Google Ads Match Types: Broad, Phrase and Exact Compared

## Key takeaways

- Google Ads has three live match types: exact, phrase and broad. Each wider type also reaches the narrower types' searches, and close variants can't be switched off.
- On Optmyzr's 2024 averages, broad match (ROAS 2.77) pays only on products with a contribution margin above 36%; exact match (4.15) pays from 24%.
- Choose the type by the commercial job, then read search terms weekly: keep variants that sell the same product, exclude the wrong buyer or item.

A Google Ads match type is the rule that sets how closely a search must relate to your keyword before your ad can enter the auction. There are three live types: broad, phrase and exact.

The symbol you type is an intent boundary, not a filter on the exact string. Close variants apply to every type, so real control comes from the search terms report and negatives.

I checked each rule against Google's help pages in October 2026 and priced the types with Optmyzr's 2024 keyword data.

## What are the Google Ads match types?

The Google Ads match types are broad, phrase and exact, and [Google's keyword match type guide](https://support.google.com/google-ads/answer/14102374?hl=en) ranks them from most reach to most control. You set the type with punctuation around the keyword.

| Match type | Syntax | What can trigger the ad | Reach | Control |
|---|---|---|---|---|
| Broad | lawn mowing service | Searches related to the keyword's meaning | Widest | Lowest |
| Phrase | "lawn mowing service" | Searches that include the keyword's meaning | Middle | Medium |
| Exact | \[lawn mowing service] | Searches with the same meaning or intent | Narrowest | Highest |

Google's own example uses one keyword, lawn mowing service. As a phrase match keyword it can show on "lawn mowing service near me" or "hire company to mow lawn".

As broad match it can also show on "lawn aeration price", a related search without the keyword's words.

The types stack rather than split traffic. Phrase can reach every search exact can, plus more; broad can reach both, plus related searches.

So a wider type is a bigger net over the same water, not a separate pool.

### How common is each match type?

Broad match covered 36.67% of keywords in 2024, up from 33.12% in 2022, in [Optmyzr's analysis](https://www.optmyzr.com/blog/choose-best-google-ads-keyword-match-type/) of 992,028 keywords across 15,491 Google Ads accounts (November 2024).

**Figure: Share of Google Ads keywords by match type, 2022 and 2024.** Broad match is the only type that grew, from 33.12% to 36.67% of keywords, while exact fell to 34.35% and phrase to 28.98%.

![Bar chart of keyword share by match type: broad grew from 33.12% in 2022 to 36.67% in 2024, exact fell from 37.11% to 34.35%, phrase from 29.77% to 28.98%.](https://www.productmetrics.io/images/blog/google-ads-match-type/match-type-share.webp)

| Match type | 2022 | 2024 |
| --- | --- | --- |
| Broad | 33.12% | 36.67% |
| Exact | 37.11% | 34.35% |
| Phrase | 29.77% | 28.98% |

_Source: Optmyzr keyword study, 992,028 keywords across 15,491 accounts, November 2024_

Exact and phrase each lost a little share over the same two years. Loose matching is already normal in working accounts, which is why reading the queries matters more than the punctuation you typed.

### What happened to broad match modifier?

Broad match modifier is gone: since July 2021 you can't add keywords with the plus-sign syntax (+lawn +mowing) that forced chosen words into the search.

Google folded that traffic into phrase match, and old modifier keywords still in an account now behave like phrase match. Judge their search terms the way you judge a quoted keyword.

## Do close variants mean exact match isn't exact?

Yes: exact match means same meaning, not the same characters.

[Google's close variants documentation](https://support.google.com/google-ads/answer/9342105?hl=en) says every match type is eligible for close variants by default and there is no way to opt out.

For an exact match keyword, Google lists these close variants, with its own examples:

| Close variant | Keyword | Search it can match |
|---|---|---|
| Misspelling, plural or stem | \[floor] | flooring |
| Reordered words | \[shoes mens] | mens shoes |
| Function word added or removed | \[men shoes] | shoes for men |
| Implied word | \[daydream vr headset] | daydream headset |
| Synonym or paraphrase | \[bathing suits] | swimming suits |
| Same search intent | \[images royalty free] | free copyright images |

Exact still gives tighter intent control than phrase or broad. Read the brackets as "keep this commercial intent".

A synonym that swaps brand, size or model is a different product, even when Google treats it as the same meaning.

The Google Ads API makes the gap visible. Search terms carry classifications such as NEAR_EXACT and NEAR_PHRASE for close-variant hits, while the [Google Ads API match type enum](https://developers.google.com/google-ads/api/reference/rpc/v26/MatchTypeEnum.MatchType) stores only BROAD, PHRASE and EXACT on the keyword.

Keep the search term next to its classification, and don't infer what sold from the keyword alone.

### How do you sort the queries a keyword catches?

Sort every search term into one of three groups: the exact offer you wrote, a close variant where the form changed but the sale is the same, and a broader expansion where the meaning drifted.

Then compare [conversion value](https://www.productmetrics.io/glossary/conversion-value "Conversion value is the amount you tell Google Ads a conversion is worth, such as sales revenue or profit margin."), margin and new-customer rate across the groups.

A reordered product phrase usually keeps the intent. A synonym can change compatibility or model, and that sells a different item.

Negative keywords stay necessary here, because no keyword list can name every query Google treats as the same meaning.

## Which match type has the best ROAS?

Exact match had the highest average [ROAS](https://www.productmetrics.io/glossary/roas "Return on ad spend (ROAS) is the revenue your ads generate for every euro you spend on them.") in Optmyzr's November 2024 study, at 4.15, against 3.13 for phrase and 2.77 for broad.

Exact also led on click-through rate (21.66% against 8.5% for broad), while phrase had a 9.31% conversion rate.

**Figure: Average ROAS by match type against break-even at a 30% margin.** Exact match averaged a ROAS of 4.15, but phrase (3.13) and broad (2.77) both sit below the 3.33 a product with a 30% contribution margin needs to break even.

![Bar chart of average ROAS by match type: exact 4.15, phrase 3.13, broad 2.77, against a break-even line of 3.33 for a 30% contribution margin; only exact clears it.](https://www.productmetrics.io/images/blog/google-ads-match-type/match-type-roas.webp)

| Match type | Average ROAS |
| --- | --- |
| Exact | 4.15 |
| Phrase | 3.13 |
| Broad | 2.77 |

Reference line: Break-even at a 30% margin (3.33).

_Source: ROAS: Optmyzr keyword study, November 2024. Break-even calculated as 1 ÷ 0.30_

Those averages read differently once you add margin. [Break-even ROAS](https://www.productmetrics.io/glossary/break-even-roas "Break-even ROAS is the revenue-to-ad-spend ratio at which contribution profit exactly covers advertising costs.") is 1 ÷ [contribution margin](https://www.productmetrics.io/glossary/contribution-margin "Contribution margin is the revenue from a sale minus its variable costs, such as product cost, shipping, payment fees and returns."), so a product with a 30% margin needs a ROAS of 3.33 before an ad pays for itself.

The same check in profit terms is [POAS](https://www.productmetrics.io/glossary/poas "POAS (profit on ad spend) means contribution profit before advertising divided by ad spend: the profit each euro of ads leaves.") (ROAS × contribution margin), where 1.0 is break-even, as [POAS vs ROAS](https://www.productmetrics.io/blog/poas-vs-roas) explains.

| Match type | Average ROAS (Optmyzr, 2024) | POAS at a 30% margin | Margin needed to break even |
|---|---|---|---|
| Exact | 4.15 | 1.25 | 24.1% |
| Phrase | 3.13 | 0.94 | 31.9% |
| Broad | 2.77 | 0.83 | 36.1% |

Read from the margin side, broad match at its average ROAS only pays on products with at least a 36.1% contribution margin. Exact match pays from 24.1%.

Work out your own floor with the [break-even ROAS calculator](https://www.productmetrics.io/break-even-roas-calculator) before you pick a target.

> **Sidenote:** These are averages across thousands of accounts. Exact keyword lists often hold brand terms and best sellers, so part of exact match's lead comes from what advertisers put in it, not the match type alone.

The catch is that one search term rarely sells one product.

A broad query for trail running shoes can sell a high-margin own-brand shoe on Monday and a discounted branded one on Tuesday. Checking that by hand, query by query and product by product, stops working past a few dozen products.

> **Grow revenue with the products that perform best.** Every product segmented on profit, so you know which ones get more priority. [Analyse your products for free](https://app.productmetrics.io/) Log in with Google, no credit card.

## When should you use broad, phrase or exact match?

Use exact match when the query is the buying decision, phrase match when qualifiers are welcome, and broad match for discovery once bidding has conversions to learn from.

Pick the type from the commercial job, not from a habit of always starting exact.

| Match type | Use it for | Example keyword | Watch out for |
|---|---|---|---|
| Exact | A brand, model number or product that must not widen | \[trail running shoes] | Synonyms that swap brand or model |
| Phrase | A product meaning plus useful qualifiers | "running shoes" | Qualifiers that change the item, such as kids' sizes |
| Broad | Demand your keyword list missed | running shoes | Spend on related but unsellable searches |

Phrase match keeps the product and lets "near me", "sale", a size or "for women" come along without listing every combination. You still exclude the qualifiers that change the item, such as the wrong width.

Broad match needs [Smart Bidding](https://www.productmetrics.io/glossary/smart-bidding "Smart Bidding is a set of Google Ads bidding strategies that use Google AI to optimise for conversions or conversion value in every auction.") to work. [Google's broad match documentation](https://support.google.com/google-ads/answer/12159290?hl=en) says it reaches more relevant searches than phrase or exact, uses additional signals to judge intent and gives Smart Bidding more data.

Pair it with a value-based Smart Bidding strategy with a target ROAS, and use it only where the campaign has enough conversions to price the extra reach.

### How do you test broad match on profit?

Test broad match on incremental margin, not click-through rate. Hold creative, landing pages, budget and the bid target as steady as you can, then compare:

- incremental qualified conversions, not raw clicks;
- contribution margin, not platform ROAS alone;
- new query themes you couldn't have listed in advance;
- the share of spend that needed a negative.

Sparse conversions make the extra reach hard to price. Join the new queries to margin and stock before you scale, as in [improving PPC campaigns on profit](https://www.productmetrics.io/blog/how-to-optimize-ppc-campaigns).

When a theme is wrong, [choose negative keyword match types](https://datahuntersagency.com/google-ads-negative-keywords/) that block the intent without wiping out the useful variants.

### Should you add every keyword in all three match types?

No: because coverage stacks, copying every keyword into broad, phrase and exact mostly buys the same searches three times.

The copies split the conversion history Smart Bidding learns from, and your own keywords compete for the same query.

Say a retailer runs \[trail running shoes] as exact in one campaign and trail running shoes as broad in another, with the same ads and the same target ROAS.

The broad keyword already covers most of the exact traffic, so the exact campaign only thins out the bidder's data. Keep the exact copy only if that query needs its own budget, and handle wrong brands and incompatible models with negatives.

## How do you audit search terms by match type?

Audit search terms by comparing each keyword with the searches it actually triggered, then keep, exclude or move each search by intent.

[Google's search terms report help](https://support.google.com/google-ads/answer/2472708?hl=en) shows the keyword next to the user's search and marks close variants in the Match type column, as in "purple flowrs" matching \[purple flowers].

**Audit search terms by match type**

1. **Pick the match type from the job.** Exact when the query is the offer, phrase when qualifiers are safe, broad when you want demand your list missed and bidding has conversions to price it.
2. **Export search terms with the keyword column.** Keep the triggering keyword and the user's search side by side. An export of keywords alone hides which searches they actually caught.
3. **Label each search by intent.** Mark it as an exact match, a close variant or a broader expansion, judged by what the person wants to buy, not by how similar the words look.
4. **Keep the variants that still sell.** A reordered phrase or a plural that leads to the same product is not waste. Leave it running.
5. **Add negatives for the wrong buyer or item.** Jobs, a competitor brand, the wrong size or an incompatible model.
6. **Check margin before you scale broad match.** Compare incremental conversions and contribution margin, not clicks. Extra revenue on low-margin or out-of-stock products can still lose money.

Take the lawn mowing keyword through that audit. The question on each row is whether the person still wants to buy what you sell.

| Keyword | Search | Keep it? | Why |
|---|---|---|---|
| \[lawn mowing service] | grass cutting service | Usually yes | Same hire-a-service intent, as a close variant |
| \[lawn mowing service] | lawn mowing jobs | No, add a negative | The searcher wants work, not a contractor |
| "lawn mowing service" | affordable lawn mowing service near me | Yes | Keyword meaning with price and place qualifiers |
| lawn mowing service | book a gardener to cut the lawn | Check first | Related meaning, but "gardener" can mean planting |
| +lawn +mowing (legacy) | lawn mowing service prices | Treat as phrase | Modifier keywords now follow phrase match |

The "Match type" column is where [Google's guide to analysing ads](https://business.google.com/us/resources/articles/how-to-analyze-google-ads-successfully/) points too: it shows how each keyword's match type affects performance, so you can tighten the ones that catch the wrong searches.

The same keyword in two campaigns can still produce different search terms when negatives, bids or account context differ.

Label each row by who is buying and what is being sold, the same lens as in [PPC research for ecommerce](https://www.productmetrics.io/blog/ppc-research).

Shared negative lists help when several campaigns sell the same products, and [managing negative keywords](https://www.productmetrics.io/blog/google-ads-negative-keywords) covers where each list belongs.

## Final thoughts

A match type only decides which searches can enter the auction; whether they pay depends on the margin of what the click sold.

Pick the type by the job, read the search terms every week and judge broad match on profit, not clicks.

For Shopping and Performance Max, where there are no keywords, the same discipline moves to the products.

[Product Segmentation](https://www.productmetrics.io/product-segmentation) segments every product on its own performance and writes the result to Merchant Center as labels, so the products that pay get more priority.

Match types are a Search setting. If you also buy [advertising across Google's display network](https://baikou.agency/blog/google-ads-display-network/), targeting there runs on audiences and placements rather than queries, so the search terms audit above doesn't cover it.

If the sale happens offline, or the conversion you count is a lead rather than an order, the query that looks best in Google Ads may not be the one that paid.

A [Google Ads management playbook with offline conversion imports](https://crescade.com/blog/google-ads-management) shows how to feed the real outcome back so bidding learns from it.

## Frequently asked questions

### Is broad match the default match type in Google Ads?

Yes. A keyword typed without quotation marks or square brackets is a broad match keyword. Google lists broad as the default for all new keywords, so check the symbols before you upload a list.

### Do negative keywords match close variants?

No. Negative keywords don't match close variants, so a negative for flowers still lets red flower trigger your ad. Google handles casing and misspellings, but add plurals and synonyms as separate negatives.

### Do Shopping and Performance Max campaigns use match types?

No. Shopping and Performance Max campaigns have no keywords to set a match type on. Google matches searches to your products from the product feed, so the levers there are the feed, negative keywords and how much priority each product gets.

## Keep reading.

- [POAS In The Real World Passing Profit, Composing Conversion Value, And Bidding With Confidence](https://www.productmetrics.io/blog/poas-conversion-value-bidding)
- [How to optimise PPC campaigns: a margin-first playbook](https://www.productmetrics.io/blog/how-to-optimize-ppc-campaigns)
- [PPC research: keywords, competitors and the break-even CPC](https://www.productmetrics.io/blog/ppc-research)

---

Written by Berend Vrakking, founder of Product Metrics. Last updated 2026-10-30.

1M+ products analysed · Nominated at the Dutch Search Awards for Best Measurement Campaign (2023) and Best Use of Data (2024)

HTML version: https://www.productmetrics.io/blog/google-ads-match-type