Answer in brief
Start with the queries your site already appears for. In the Search Console performance report, filter for words that signal a decision, such as a price, a place, a deadline or a type of client, and group what you find by what the person wants. Give each distinct intent one page that answers it completely, link to it from related pages, and judge the result over months and by enquiries, because the report leaves out the rarest queries and small numbers swing widely.
The rarely typed query sits closest to a purchase
A broad query such as “web design” is typed by students, job seekers, competitors and people who are a year away from spending anything. A query such as “website redesign for a dental clinic with online booking” is typed by very few people, and almost every one of them has a project, a budget and a deadline. The more words a person adds, the more of the decision they have already made.
That is why specific queries matter so much to a service business. One client can mean weeks or months of paid work, so a page does not need a crowd; it needs the right handful of visitors. The trade calls these low search volume keywords: tools often report nothing at all for them, which is also the reason competitors leave them alone.
Nothing here is a trick. Google’s own SEO Starter Guide says plainly that it holds no secrets that will automatically rank a site first. The method below is more modest and more reliable than that: find the precise questions buyers already ask, answer each one on a page of its own, connect the pages and measure patiently.
Search Console already holds your first list of buying queries
Open the performance report in Search Console and choose the Queries tab. Google’s help describes queries as the search terms that led to your site, so this is a record of demand that has already touched you. Add a filter for queries containing a decision word: price, cost, quote, hire, the name of your city or of an industry you serve. Repeat it with each word and keep what you find.
Read the list with one caveat in mind. The same help page explains that some queries are omitted from the report to protect user privacy, and calls them anonymised queries. It adds that Search Console stores and shows only the most important rows, so not every query appears in the table, and that the most complete list comes from bulk data export. The rarest searches, which are exactly the ones this article is about, may therefore never be shown by name.
So add two sources that no tool can see. The first is your inbox: the exact sentences people write when they enquire. The second is the sales conversation: the questions asked before signing. The Starter Guide observes that people who know a subject search with different words from newcomers, and gives “charcuterie” and “cheese board” as its example. Write the client’s word down next to the professional one.
Sort the list by what the searcher is about to do
Group the queries by intention, not by vocabulary. For services, three groups cover most of the list. Some people are choosing a provider and add words like agency, studio, company or best. Some are checking terms and add price, cost, how long or contract. Some describe their own situation: a type of business, a city, an urgent date, a failed first attempt.
To decide whether two queries belong together, imagine one page with one opening paragraph. If it would satisfy both searchers, they share an intent and need one page. If the answer to the first would be noise to the second, they need two. A difference in wording alone is no reason to split: the Starter Guide notes that Google’s language matching can understand how a page relates to many queries even when the exact terms are absent.
Then rank the groups by how close they are to a decision and by how well you can deliver, not by impressions. A query you cannot serve properly does not deserve a page. Google’s guidance on people-first content offers a useful filter here: would your existing or intended audience find the page useful if they came to you directly, with no search engine in between?
One intent, one page: why splitting and merging both fail
Splitting is the more common mistake. Five pages written for five wordings of one need compete with each other, and none of them becomes the obvious answer. Bing’s Webmaster Guidelines describe the effect from the search engine’s side: publishing the same or substantially similar content across several addresses reduces its confidence in selecting a preferred version.
Merging fails in the opposite direction. A single services page that lists everything the firm does answers nobody’s particular question. The same Bing guidelines advise focusing each address on one primary topic, keeping unrelated concepts off a single page and aligning the title, the headings and the intent of the content.
A workable rule sits between the two. A page earns its own address when the service, the client or the terms differ in a way that changes the answer: another process, another way of pricing, other deliverables, other proof. When only a noun changes, write a paragraph, not a page. This is also the line between a useful specific page and the practice Google’s guidance warns against, producing lots of content on many topics in the hope that some of it performs.
Five answers a ready-to-buy visitor looks for first
Someone who typed a precise query arrives with the decision half made and checks five things quickly. Is this exactly the service I need? Is it meant for a business like mine? What does it cost, or how is the price formed? What happens after I get in touch, and how long does it take? Who has already done this with you? Bing’s guidelines add an instruction about order: place essential information near the top and avoid long introductions.
Price is where most service pages go silent. If a figure cannot be published, publish the logic: what makes the job larger or smaller, what is included and what is not. Describe the process as steps, with what you need from the client at each one. For proof, show a comparable piece of work, named with permission or described without the name.
Google’s self-assessment questions make a good final check. After reading, will the visitor feel they have learned enough to achieve their goal, or will they have to search again for better information? Length is not the measure. The Starter Guide states that there is no magical word count target, minimum or maximum, so the page is long enough when the five answers are on it.
Titles and snippets written in the client’s own words
The title is the first thing a searcher reads. According to the Starter Guide, a good title is unique to the page, clear and concise, and accurately describes the contents. On a specific page that means naming the service together with its qualifier, whether that is the client type, the place or the condition, in the words the client used and not in internal jargon.
Below the title sits the snippet. Google explains that it is taken from the content of the page and only occasionally from the meta description, which should be short, unique to the page and contain its most relevant points. If the query asks about cost or timing, let that sentence appear early in the text. Search Console’s help suggests one more use of the query list: look for queries with high impressions and a low click-through rate, because they point to titles and descriptions worth improving.
Restraint matters as much as precision. The Starter Guide warns that repeating the same words over and over is tiring for users and that keyword stuffing is against Google’s spam policies. It also says that keywords in a domain name or URL path have hardly any effect on their own. Use the phrase where an editor would, in the title, the opening and one heading, and then write normally.
Internal links that carry a decided reader to the right door
A specific page that nothing links to is practically invisible. The Starter Guide says the vast majority of new pages Google finds every day are found through links, and that link text tells both users and Google something about the page being linked. Bing’s guidelines ask that every important address be reachable through crawlable internal links with relevant anchor text.
Start with the links that point in. The general service page should list its specific cases and link to each. Articles, guides and case studies that touch the subject should mention the page where a reader would naturally want it. Write anchor text that names the destination, for instance “website redesign for clinics”, never “read more” or “click here”.
Then add the links that point out. From the specific page, link to the general service, to neighbouring cases and to the contact step, so that a visitor who landed slightly off target is one click from the right page. Make it a habit: every time a new page is published, name three existing pages that should link to it and edit them the same day.
Reading the results when the numbers are small
Give the page time before judging it. The Starter Guide says that some changes take effect in a few hours and others take several months, and suggests waiting a few weeks before assessing any change. Low search volume keywords gather evidence slowly, so compare quarters and not weeks. Search Console’s help notes that weekly or monthly granularity smooths out daily fluctuations and makes long-term trends easier to see.
Look at the page before the query. Filter the report by the page’s address and read its totals; because the rarest queries are hidden from the table, page-level figures tell you more than individual rows. Treat average position with care. As an invented illustration: if a page was shown four times in a month, a single appearance far down the results moves its average more than any improvement you made.
The measure that matters is the enquiry. Record which page each enquiry came through and ask new clients what they searched for. A page that brings one suitable client a quarter may be the most valuable on the site. If a page shows no impressions after several months, check that it is indexed and linked and that it answers a question people really ask, then improve it, merge it or remove it. No method guarantees a position.
Practical checklist
- Export the queries your site already appears for and mark every one that names a cost, a location, a date or a kind of client.
- Collect the exact wording clients use in enquiries and calls, and add it to the list beside the professional terms.
- Group the list by intent and assign every group to one existing or planned page, so that no two pages answer the same question.
- Check each page for the five answers: the service, who it suits, the price logic, the process and timing, and the proof.
- Link every specific page from at least the general service page and one related article, with anchor text that names the destination.
Questions and answers
Is a page worth building for a query with almost no search volume?
Often yes, when the query describes work you actually deliver and a need that differs from your other pages. Volume estimates for rare searches are rough, and even Search Console leaves very rare queries out of its table. What decides the matter is the value of one client and whether the page says something your other pages do not. A page that differs only by one swapped word is not worth building.
How can I find long-tail queries that Search Console does not show?
Use the report for what it does show, then go beyond it. Filtering by page gives totals that include visits from queries the table hides. For the fullest list, Google’s help points to bulk data export. Outside the tool, read enquiry emails, call notes, on-site search terms and the questions clients ask before they sign; these are the same people using the same words.
How many keywords should one service page target?
Think in intents, not in keywords. One page should serve one need, however many ways people phrase it. Google’s Starter Guide says its systems can connect a page with searches that never use its exact words, so a separate page for each wording is unnecessary. Mention the natural variants where they fit and spend the effort on answering the need fully.
How long before a new specific page starts bringing enquiries?
Nobody can promise a date. Google says changes may show within hours or take several months, and advises waiting a few weeks before drawing conclusions. For searches that are typed rarely, allow longer still, because there are simply fewer occasions for the page to be shown. In the meantime confirm that the page is indexed and that other pages link to it.
Should pages with very few clicks be deleted?
Not because of the click count alone. First ask whether the page has produced enquiries, whether it is indexed and linked, and whether it repeats another page. Fold duplicates into the stronger page and redirect the old address. Keep a page that answers a real question for a few valuable visitors; that is the whole point of working with specific queries.