Google treats two words as the same search when they mean the same thing. B2B buyers use different words because they want different things.
Google Ads close variants let your keywords match searches that Google judges to have the same meaning or intent, including synonyms, and you cannot switch that behavior off. For most consumer advertisers that is a gift. For a B2B advertiser it quietly merges buyers who want different products, at different prices, from different kinds of providers.
This post is for the person running a B2B search account who has seen lead quality drift without any change they made. It covers how to test whether two similar terms are really one buyer, how to pick the word that governs your ads and landing pages, and how to keep two close terms apart inside the account once you know they are different.
Google decides what your synonyms are, and there is no opt-out
Close variants are Google's list of searches it considers equivalent to your keyword. The definition page is blunt about the scope: "By default, all keyword match types are eligible to match to close variants. There's no way to opt out." (Keyword close variants: Definition)
That list has grown well past misspellings and plurals. For exact match keywords alone, Google lists synonyms and paraphrases (its example is bathing suits matching swimming suits) and same search intent (images royalty free matching free copyright images) as valid close variants. The keyword matching page puts it in one line: "Ads may show on searches that have the same meaning or same intent as the keyword." (About keyword matching options)
So the question of which words count as the same is answered by a model trained on billions of searches, most of them from consumers. Bathing suits and swimming suits really are the same purchase. In B2B, the closest-sounding pair of terms often is not.
Ready for paid ads that pay off?
Book your free auditWhy B2B synonyms are usually two different buyers
In B2B, the word a buyer types tells you what kind of help they think they need, and that decides whether you can sell to them. Two terms can describe the same category on paper and still come from people shopping for different things.
A few illustrative pairs show the pattern:
→ Payroll software and payroll service. One searcher wants a tool their team will run. The other wants someone to take the job off their desk. Same category, opposite offer.
→ Penetration testing and vulnerability scanning. Close enough that a model will happily bridge them. One is usually a scoped project with a real budget, the other is often a search for a cheap or free scanner.
→ Contract manufacturer and supplier. The first buyer wants a production partner. The second may just want a parts quote.
In every pair, one term attracts the buyer you can close and the other attracts someone who will fill in the form and then go quiet on the first sales call. The click costs the same. The lead does not.
The pattern I see most often in B2B accounts is a team picking its main term because it has the bigger number in Keyword Planner. The bigger number is frequently the broader word, and the broader word is frequently the one that signals the wrong kind of provider. Volume picks the wrong word more often than it picks the right one.
A three-part test for whether two terms are one buyer
Two terms are the same buyer only if they produce the same kind of result page, the same quality of lead, and the same words in your sales calls. Run all three checks before you decide to merge or split.
1.) Read both result pages. Search each term in a clean browser and look at who is ranking and advertising. If one page is full of software vendors and the other is full of agencies, consultancies or marketplaces, Google's own index thinks these are different needs, even if its ad matching does not.
2.) Split your leads by search term. The search terms report shows the actual query behind each click, and its "Match type" column flags when a search matched as a close variant of your keyword (About the search terms report). Pull 90 days, group the queries into the two terms, and compare the share of leads that reached your qualified CRM stage. A gap of a few points is noise. A gap where one term's leads qualify at roughly twice the rate of the other is a different buyer.
3.) Ask sales which word closed customers used. Listen to five recorded first calls, or ask the reps directly. The word a customer uses to describe their own problem is the word that belongs in your headline. Buyers rarely use the vendor's category name for themselves.
If all three checks point the same way, you have your answer. If the result pages differ but the lead data is too thin to read, trust the result pages and the sales calls over a spreadsheet with eleven leads in it.
Let the closing buyer's word govern the copy, not the search volume
Pick the term your best customers use, and make it the word in your top headline and your landing page's main heading, even when a synonym carries more searches. This is the decision most accounts get backwards.
The higher-volume synonym feels like the growth play. In practice it does two kinds of damage. It attracts the adjacent buyer directly, and it teaches the people you actually want that you might be the wrong kind of provider. A buyer looking for a service who lands on a page that says software assumes you sell software and leaves.
The right word also makes every other filter in the account work better. Ad copy that names the buyer's situation in their own vocabulary repels the wrong click before you pay for it, which is a cheaper filter than any negative keyword list. And a landing page whose heading repeats the term the searcher typed is the cheapest form of message match there is.
None of this means ignoring the bigger term forever. It means the bigger term earns a place by proving its leads close, not by having a larger number next to it in a planning tool.
How to keep two close terms apart inside one account
Once you know two terms are different buyers, give each one its own ad group, its own identical exact keyword, its own landing page and a set of cross-negatives. Each of those pieces closes a different door the matching system would otherwise leave open.
Its own ad group with an identical exact keyword. Google's prioritization rules put "Exact match keywords that are identical to the search term" first when several ad groups could serve the same search, and the footnote matters here: identical "doesn't include plurals or synonyms" (About ad group and asset group prioritization). An identical exact keyword is the only routing you fully control. Anything less identical gets decided by the system.
Its own landing page. For searches that are not identical to any keyword, the same page says relevance "is determined by the meaning and context of the search term, all the keywords in an ad group, and landing pages within the ad group." Two ad groups pointing at one generic page look the same to that system. Two specific pages give it something to tell apart.
Cross-negatives, added one form at a time. Add the service term as a negative in the software ad group and the reverse. The catch is that "Negative keywords won't match to close variants or other expansions," so Google's own page says you need to add synonyms and singular or plural versions yourself (About negative keywords). Add the likely forms in one sitting. Positive keywords stretch on their own. Negatives never do.
A weekly look at the close-variant rows. In the search terms report, filter the match type column to close variants and read only those rows. That is where the system is making the synonym decision for you, and it is the shortest list in the account that tells you whether your walls are holding. It fits inside the weekly negative keyword routine you should already be running.
A brand-new category term needs isolation first
A term buyers have only just started using should sit in its own ad group from day one, so its performance can be read on its own before anyone decides whether it is a synonym. Folded into an existing ad group, its few searches disappear into the bigger term's totals and nobody ever learns whether the new word is a real shift in how buyers search.
Expect the first few weeks to look broken. Google gives a keyword with very little search history a "Low search volume" status, and such a keyword "will be inactive until its search traffic increases" (Low search volume: Definition). The same page says the status is rechecked and that waiting "can be particularly helpful if you're advertising a new brand, term, or product." So an emerging term often shows zero impressions while your broader keywords quietly pick up its searches. Leave it in place, and check the search terms report to see which keyword is actually serving those queries in the meantime.
Then set the merge rule in advance. Once the new term has enough qualified leads to compare, run the three-part test above against the established term. If the leads qualify at a similar rate and the result pages look alike, fold it into the established ad group so the bidding system gets the combined volume. If it brings a different buyer, it stays apart and earns its own copy and page. How much money a new theme gets while it is being read is a separate decision, covered in our post on funding keyword themes in tiers.
When splitting is the wrong call
If two terms pass the three-part test as the same buyer, consolidate them and stop managing the difference. Splitting near-identical keywords across ad groups fragments a thin conversion count, and Google's matching page already says "Broader match types capture all the queries of narrower match types, plus more" (About keyword matching options). That is the argument in our post on Google Ads match types for B2B.
Splitting is also wrong when the account is too small to read either side. A campaign producing fewer than about ten qualified leads a month cannot support a vocabulary experiment, so pick the closing buyer's word from the sales calls, write the copy around it, and move on.
What to do this week
1.) List the three highest-spend non-brand keywords and, for each, write down the closest synonym a buyer might use instead.
2.) Search both versions of each pair and note who ranks and who advertises.
3.) Pull 90 days of search terms, split by the two words, and compare the qualified-lead rate.
4.) For any pair that turns out to be two buyers, build the second ad group with an identical exact keyword, its own landing page heading and cross-negatives.
5.) Put the close-variant filter on the weekly search terms review.
That is an afternoon of work, and it usually turns up at least one pair where the account has been paying for the wrong buyer under the right-sounding word. If you want a second set of eyes on your own search terms, that is a core part of how we run B2B Google Ads at Profit Mill.
