Exact match is not exact: what close variants mean for trades

No. Google now matches exact-match keywords to searches with the same meaning or intent, not only the same words. SoudCoh's analysis of reported trade campaign data found close to half of what exact match produced came through wording the business never chose.

Key facts

• 47% of the enquiries that exact-match keywords produced came from searches worded differently from the keyword.
• Phrase match carries 69% of keyword spend across SoudCoh's research sample, exact about 30% and broad about 2%.
• Phrase-match clicks cost 17% less than exact-match clicks, and convert less often.
• The two effects cancel out: a phrase-match enquiry cost about the same as an exact-match one (5% less).
• Where broad match was used, its enquiries cost 1.09× as much as exact-match enquiries.
• Phrase match, including its close variants, produced 34% of enquiries across the sample.

How we measure. An exact-match enquiry counts as a close variant when the search behind it was worded differently from the keyword. An enquiry is a recorded phone call or a submitted form.

Google's own list of close variants includes misspellings, singular and plural forms, words with the same stem, abbreviations, accents, reordered words with the same meaning, small words added or dropped, implied words, paraphrases and searches with the same intent. In a trade account that looks like this:

• An exact keyword for hot water system repair can show for repair hot water system, hot water systems repairs or hotwater system repair.
• An exact keyword for switchboard upgrade can show for upgrade switchboard, switch board upgrade and wording Google judges to mean the same job.
• An exact keyword for end of lease cleaning can show for end of lease clean and other phrasings a tenant uses for the same job.

Google decides what counts as the same meaning, auction by auction, and the only record of its decisions is the search terms report.

Mostly it helps. Customers describe the same job in many ways, and close variants catch them without a keyword for every spelling. But it changes three things about how an account has to be run:

• Exact match is not control on its own. Even an account built only on exact match brings in wording the business did not choose, so it still needs a weekly search-term review and a negative list.
• Same meaning is Google's call, not yours. What Google judges to be the same intent may not be the same job for your business: a repair and a replacement, a house and a strata block, a one-off clean and a regular one. Each can deserve its own ad, page and bid.
• Negatives do not stretch the same way. Negative keywords do not match close variants, plurals or synonyms. A negative that blocks one form of a word leaves its plural untouched, so the negative side of an account needs more care than the positive side.

The last point is where accounts get hurt. People add negatives broadly to compensate for loose positive matching, and some of those negatives end up blocking paying customers. Our guide to negative keywords that block customers covers how to check them first.

No. The common belief is that exact match buys tighter, cheaper leads. In SoudCoh's research sample, it buys dearer clicks that convert better, and the two effects land close to level:

• Phrase-match clicks cost 17% less than exact-match clicks.
• Phrase-match clicks were also less likely to turn into an enquiry.
• Net result: a phrase-match enquiry cost 5% less than an exact-match one.

The price of a click and the value of a click move together. Phrase match reaches cheaper, looser searches; exact match pays up for the precise ones. That is why phrase match has become the workhorse for trades, carrying 69% of keyword spend across the sample: it catches the natural variations people type, such as the suburb, the urgency or the symptom, while still requiring the service words to be present.

The lesson is not that phrase beats exact. It is that match types should be judged on cost per enquiry, never on cost per click, and that the search-term data rather than a match-type rule should decide where budget goes.

Late, and only where it has earned a place. Where broad match ran in SoudCoh's research sample, its enquiries cost 1.09× as much as exact-match enquiries. That is not materially dearer, but broad match leans on automated bidding and conversion data to police what it matches; without those, it spends on whatever Google judges to be related.

That is why SoudCoh's builds launch every keyword as an exact-and-phrase pair, with effectively no broad match at launch. Phrase match supplies the discovery in the first weeks, exact match holds the proven core, and broad match is trialled later, if at all, once the account has conversion depth and a mature negative list. Across the sample broad match is about 2% of keyword spend.

For the full launch sequence, see our white paper Launch, discover, control.

Use match types for what they are good at, and let structure do the controlling:

1. One ad group per service. Close variants of one job are then far more likely to land in that job's ad group, with its own ads and landing page.
2. Pair each core keyword as exact and phrase. Exact holds the proven wording; phrase finds the variations.
3. Promote repeat converters. When a search converts more than once, add it as an exact keyword in the right ad group. Google prefers an exact keyword that is identical to the search, so promotion also routes the search to the right ad.
4. Write negatives for every form. Because negatives ignore close variants, add plurals and common misspellings deliberately, and check each negative against searches that have converted.
5. Add broad match last, in one campaign at a time, once tracking and daily review are in place.

See how this works for plumbing businesses, electricians and HVAC businesses, or how we run Search campaigns.

You can do it in the search terms report:

1. Open the search terms report for a Search campaign over the last one to three months.
2. Add the Match type column. Google labels searches that matched as close variants, for example as an exact match close variant.
3. Filter to exact-match close variants and sort by cost.
4. For each of the top rows, compare the search with the keyword that caught it. Ask whether it is the same job, a different job you do, or a job you do not do.
5. Check conversions. Anything converting more than once is a promotion candidate; anything costly with no enquiry is a candidate for a closer look, not an automatic negative.

Remember that the report only names part of what you spend. Google does not show every search, which we cover in why the search terms report hides half your spend. And if you are surprised by how much of your demand arrives in wording you never chose, you are not alone: see most trade enquiries come from searches you never chose.

Does exact match in Google Ads only show for the exact keyword?

No. Exact match now reaches searches with the same meaning or intent as the keyword, including reordered words, plurals, misspellings and paraphrases. In SoudCoh's research sample, 47% of the enquiries that exact-match keywords produced came from searches worded differently from the keyword.

What is a close variant in Google Ads?

A close variant is a search Google treats as the same as your keyword even though the wording differs: a misspelling, a singular or plural, a different word order, an abbreviation, an added or dropped small word, or a paraphrase with the same intent. Google decides what counts, and the search terms report shows what it decided.

Is exact match cheaper per lead than phrase match?

No. In SoudCoh's research sample, phrase-match clicks cost 17% less than exact-match clicks but converted less often, so a phrase-match enquiry cost 5% less than an exact-match one. The two effects roughly cancel out, so judge match types on cost per enquiry, not cost per click.

Do negative keywords block close variants?

No. Negative keywords do not match close variants, plurals or synonyms. A negative for one word does not block its plural, so a negative list needs each form you want to exclude, or a broad-match negative that covers the words in any order.

Should a trade business use broad match?

Only where conversion tracking is solid and search terms are reviewed daily. Where broad match ran in SoudCoh's research sample, its enquiries cost 1.09× as much as exact-match enquiries. It is a tool for well-tracked, well-cleaned campaigns, not a default.