What your customers actually type, and why it is rarely what you would guess

Ask a business what it sells and you will get the trade word. Ask a customer what they were looking for and you will get a description of a problem.

A roofer talks about flat roof refurbishment. The customer types “water coming through bedroom ceiling”. A firm selling window treatments describes made-to-measure solutions. The customer types “how to stop sun glare on my telly”. Both are describing the same job. Only one of them is what anybody searches for.

Nearly every disappointing outcome in search comes down to that gap. Not to technical faults, not to competition, but to a site written in a vocabulary its customers do not use.

You already own the best data

You do not need to guess, and you do not need an expensive tool to start. Four sources, all of them yours.

Search Console query data. This is the good one. It lists the actual phrases people typed before your site appeared, how often you appeared, and how often anybody clicked. Read a few hundred of those and the trade-versus-customer gap becomes obvious immediately. Pay particular attention to phrases where you appear often and nobody clicks, because that usually means you turned up for something you do not really do.

Your site’s search box. Every phrase typed into it is somebody telling you what they expected to find, in their own words, on your site.

Your inbox and your phone. The questions people ask before buying are the questions they searched for beforehand. Keep a list for a fortnight.

Your reviews. Customers describing what they bought to other customers, which is the most natural language you will ever get for free.

Between those four you will have a better vocabulary list than most keyword research produces, because every word in it came from somebody real.

Four different jobs, not one

Two people can type phrases about the same product and want completely different things. It helps to think in terms of what the search is for.

Working out what the problem is. “Why is my boiler dripping”. These people are not ready to buy, and a sales page is the wrong answer. An article that genuinely helps builds the trust that gets you the call later.

Comparing options. “Combi vs system boiler”, “best X for Y”. They have accepted they need something and are deciding what. A page that compares honestly, including where you are not the right answer, does very well here.

Ready to act. “Boiler repair Hertford”, “buy X online”. This is where enquiries come from, and it is the intent most worth building proper pages for.

Looking for you specifically. Your name, possibly misspelled. Small in volume, enormous in value, and the one people forget to check they are winning.

The mistake is trying to serve all four on one page. A page that opens by explaining what a thing is, then tries to sell it, then compares it with alternatives, does none of the three well and reads as though it was written for a search engine rather than a person.

Why the obvious phrase is usually the wrong target

Every business wants to rank for the two-word phrase that describes its trade. It is the phrase with the most searches, and it is the one every competitor is also chasing, including national companies with budgets you cannot match.

It is also, quite often, a phrase with muddled intent. Somebody typing a broad trade term might be a customer, a student, a competitor, or somebody looking for a job. Ranking for it is expensive and converts poorly.

The specific phrases are the opposite. Fewer people type them, far fewer businesses have bothered to answer them properly, and the person typing has told you exactly what they want. Twenty of those, each bringing a handful of the right people, beats a long unsuccessful campaign for one flattering phrase.

This is also why volume figures deserve less respect than they get. A phrase with ten searches a month, all of them from people in your town who need what you do this week, is worth more than a phrase with ten thousand.

Matching pages to language

Once you have the list, the work is mapping rather than writing.

Group the phrases by what the person is trying to do. Then, for each group, ask which page on your site currently answers it. Some will have an obvious home. Some will have three pages half-answering them, which is a sign to merge. Some will have nothing, and those gaps are your list of pages worth writing.

Two rules keep this from going wrong. One subject per page, so a page can be genuinely good at one thing rather than mediocre at four. And no page created purely because a phrase exists, because a page written to catch a search rather than to help somebody reads exactly like what it is.

Where the words actually go

Use the customer’s language where it does work: in the page title, in the first paragraph, in headings, and in the sentences that describe what you do. Not because a search engine counts them, but because a person scanning the page needs to recognise their own problem in the first few seconds.

Keep the trade term as well, particularly where customers will encounter it later. Explaining that what they call a dripping overflow is what the trade calls a failed float valve is genuinely useful, and it means the page serves both vocabularies.

What earns the position is answering the question properly, which is the subject of the next part of this series.

Where to start

Open Search Console, go to the performance report, set it to the last three months, and read the queries. Not the summary numbers, the actual list.

Write down every phrase that surprises you, and every phrase where you appear regularly and nobody clicks. Those two lists between them will tell you more about your customers than any amount of guessing at the office.

If you would like help turning that into a plan for which pages to write and which to merge, get in touch, or read more about how we approach search.