Broad match keywords: what Google matches, why it needs Smart Bidding, and how to switch it on safely
The syntax for all three match types, the signals broad match reads, the duplicate rule that saves you an afternoon, and a two week test with its own budget.
Published by AI Knows Us · Updated 12 October 2026
A broad match keyword is a keyword typed as plain words, with no quotation marks and no brackets, which lets Google show your ad on searches related to that keyword rather than only on the keyword itself. It is the default match type in Google Ads, it reaches more searches than phrase or exact, and it hands the matching decision to Google. That makes it the most powerful match type in an account that has conversion tracking, a working negative keyword list and somebody reading the search terms every week, and the most expensive one in an account that has none of those.
The three match types, written out
| Match type | How you write it | What it reaches |
|---|---|---|
| Broad | running shoes | Searches related to the keyword, including some that do not share its direct meaning |
| Phrase | "running shoes" | Searches that include the keyword's meaning, including more specific versions of it |
| Exact | [running shoes] | Searches with the same meaning or intent as the keyword |
Google's documentation on keyword match types, checked on 12 October 2026, states that broad match is the default because it is the most comprehensive, and that the broader types capture everything the narrower ones capture plus more. So a broad match keyword already covers the searches its phrase and exact equivalents would have matched. The reference is Google's keyword match types page.
What broad match actually looks at
Broad matching is not keyword stuffing in reverse. According to the same documentation, the system may consider the person's recent searches, the content of your landing pages and assets, and the other keywords in the ad group when it decides whether your ad is relevant to a query. Two practical consequences follow from that, and both are easy to miss.
First, your landing page is part of your targeting. A vague page about several services widens the set of searches Google thinks you are relevant to, and a specific page narrows it. If broad match is pulling in queries you do not sell to, the page is often the cause rather than the keyword.
Second, the ad group is part of your targeting. Mixing unrelated keywords in one group gives the system a confused picture of what the group is about. Keep one theme per ad group, with the page that matches it, and broad match behaves far better.
Word order, duplicates and the thing you can stop worrying about
Google's page states that near identical keywords are treated as duplicates: "red car" and "car red" are the same keyword, and only the one with the higher Ad Rank is used, which the documentation notes does not hurt costs or performance. That means adding every word order, singular and plural, and every obvious variant of the same idea adds administration without adding reach. Build a short list of distinct ideas instead, and spend the saved time on negatives.
Why Google tells you to pair it with Smart Bidding
The documentation is direct about this: "It's critical to use Smart Bidding with broad match", because each query needs its own bid, and it adds that Smart Bidding "works best with broad match because the wide set of searches allows it to learn what works best for you". The logic holds up. Broad match delivers a wide and varied set of queries with very different values, so a single manual bid applied across all of them overpays for the weak ones and underpays for the strong ones. Automated bidding sets a price per auction using the signals attached to it.
Which gives you the precondition for the whole strategy: conversion tracking that records the thing you actually sell, with a value attached where you can, imported from the CRM for a lead based business. Broad match plus automated bidding plus a conversion action that fires on every form view is how accounts spend money on the wrong traffic efficiently.
When broad match is the right call, and when it is not
| Use broad match when | Use phrase or exact when |
|---|---|
| Conversion tracking is clean and the account has conversion history | The account is new and has nothing to learn from yet |
| You are looking for demand you have not thought of | Budget is small enough that a wide query set would spread it thin |
| Your margin tolerates some waste while the negative list matures | Your product has a name that collides with another meaning |
| Somebody reads the search terms report weekly | Nobody will look at the account for a month |
| The landing page is specific about what you sell | Legal, medical or regulated copy means every query has to be qualified |
How to switch it on without losing a month
- Put broad match in its own campaign or its own ad group with its own budget, so you can read what it brought and stop it without disturbing the rest of the account.
- Start with the negative list you can write from product knowledge today: free, jobs, DIY, the competitor names you do not want, the adjacent category you do not serve.
- Give it two to three weeks before judging it, and leave the bidding strategy alone during that window.
- Read the search terms report on day three, day seven and day fourteen. Google's documentation says it shows terms used by a significant number of people that triggered impressions and clicks, and that low volume terms are omitted, so treat it as a strong sample rather than a full log. Details are on Google's search terms report page.
- Add negatives at the level that matches the problem: the campaign for one bad theme, a shared list for the ones that apply everywhere. Remember that negative match types behave differently from positive ones, so check how a negative phrase will block before you add it broadly.
- Keep the queries that converted. Those are the ideas worth their own ad group, their own ad copy and their own page.
For the auction mechanics underneath all of this, including what you pay per click and how Ad Rank is formed, see our page on Google Ads pay per click.
Broad match questions that come up in account reviews
Is broad match the same as the old broad match modifier?
No. The modifier was a separate syntax with a plus sign that required certain words to be present. Today's broad match is meaning based and considers the context signals listed above, so the useful comparison is against phrase match, which is where the modifier's job largely went.
Why is my ad showing on searches that have nothing to do with my keyword?
Three usual causes, in this order: the ad group contains unrelated keywords, the landing page is general rather than specific, or the negative list has not caught up with the query set yet. Fix the group and the page first, because negatives alone turn into an endless list.
Does broad match cost more per click?
Cost per click is set in the auction rather than by the match type. What changes is the mix: broad match brings in a wider range of queries, some with higher competition and some with lower, so the average moves with the mix. Judge it on cost per conversion across the campaign rather than on cost per click.
Should every keyword be broad match now?
Keep your proven converters in phrase or exact so they hold their own ads and pages, and run broad match alongside to find new demand. That combination gives you stability where you have evidence and discovery where you do not.
How many negative keywords is normal?
Enough that the search terms report reads mostly like your business. Build one shared list for permanent exclusions and keep campaign level negatives for themes specific to that campaign, then review both monthly and delete the ones that have stopped being needed.
Can I use broad match with a small budget?
Start narrow with a small budget, prove which queries convert, then open a separate broad match test with a fixed daily amount you are prepared to spend on learning. The test has a purpose and a boundary, which is what makes it affordable.
Where match type stops being the lever
Broad match exists because Google models what a search means rather than what it says, and that is the direction every query surface has moved. The newer one takes it further: somebody describes their situation to an assistant and gets back two or three company names with a sentence each, assembled from pages, listings and reviews. There is no keyword to add, no match type to choose and no bid to raise, so the only lever is the quality and clarity of the material those answers are built from. That is a slower kind of work than editing a keyword list, and it keeps paying when the campaign is paused. AI Knows Us shows which companies the assistants currently name in a category and which sources each answer leaned on, so the work has a starting point instead of a guess.