Pick keywords by intent and value, not search volume

Fibvo editorial · 2 September 2026 · 5 min read

A keyword tool returns hundreds of rows sorted by monthly volume, and the temptation is to start at the top. That is how a business ends up ranking for a broad term that brings in students and competitors, while the phrase its buyers type sits further down the list, unranked. Volume counts how many people type a phrase. It says nothing about who they are, or whether the first page is open to you.

Sort by intent before you sort by anything else

The results page is the best available record of what Google believes a search means. Run the search and look at what already ranks. Not the positions, the type of page. A first page of guides tells you that people who type this want to read, not buy.

• Informational. The searcher wants to understand. Results are guides and explainers. A service page will not rank here, and would not convert if it did.
• Commercial. The searcher is comparing options. Results are comparisons, reviews, shortlists and directories. Worth having, but it takes a comparison page, not a sales page.
• Transactional. The searcher wants to act now. Results are service, product and pricing pages, usually with paid listings above them. These are the terms worth fighting for.
• Navigational. The searcher wants one organisation. Leave these alone unless it is yours.

Modifiers do the signalling. How, why and what point to reading. Best, versus and alternatives point to comparing. Price, quote, near me and the name of a city or sector point to buying. A term with no modifier is usually too broad to deserve a page.

If the page you would write looks nothing like the pages already ranking, you are not competing for that keyword. You are arguing with the results page.

Work out what one visit is plausibly worth

Commercial value is not a property of the keyword. It is the distance between the search and the invoice. Write in one sentence what the person typing a term must do next to become a customer. Someone looking for a supplier in their own city and sector has one step left. Someone still deciding whether the problem is worth solving has several. Both are worth having, but not the same investment or the same reporting timeline.

• What advertisers pay. If a phrase has volume and nobody bids on it, it usually does not lead to revenue. Advertisers have tested it with their own money.
• What your own data says. The search terms report in your ads account and the query report in Search Console name the phrases that have already produced enquiries. That is evidence. A tool's estimate is a guess.
• How narrow the term is. Adding a place, a sector or a job title cuts the volume and raises the value, because it removes the people who were never going to buy.

Judge difficulty by reading the pages that rank, not the score

Difficulty scores blend link metrics into one number. They are useful for sorting a long list and unreliable for deciding anything. The judgement happens when you open the results and read them.

• Who holds the slots. If marketplaces, national publishers, aggregators and government sites fill the first page, it is not open to a small site with one new article.
• How much screen is left. Look at what sits above the first organic result: paid listings, shopping units, a local pack, an AI summary. Ranking first underneath all of that is not ranking first.
• Whether the ranking pages are good or merely old. A thin page on a heavily linked domain can be beaten with a better one. A thorough, maintained page on the same domain cannot, not soon.
• Whether the results page is confused. A mixture of guides, directories and product pages means the intent is unsettled, and an unambiguous page can take a slot.
• What you already have. Related pages, internal links on the topic and a page sitting just off the first page put you in a different position from a cold start.

Then be honest about time. A competitive commercial term takes a programme of publishing, internal linking and earned coverage that runs for months before it moves. Plan a mix: terms winnable this quarter, which usually means long, specific and low in volume, and terms worth a year because the searcher arrives ready to buy. A term that is neither does not belong on the list.

One page per intent, and only one

Once the shortlist is set, map it. The unit is the intent, not the keyword. Every phrasing of the same want belongs on one page: singular and plural, the abbreviation and the full term. Two pages chasing one intent split their internal and external links, then leave Google to choose between them. You do not get a say in which.

• The intent, written as a sentence a person would recognise. Wants an agency to run Google Ads for a D2C brand.
• The single URL that owns it, existing or still to be built.
• The primary term, plus the variants that page also covers.
• The page type the intent demands, which the results page has already told you.
• The next step the page offers, matched to the intent rather than your sales target. A guide earns a download. A service page earns an enquiry.

Check for conflicts before writing anything. Search your own domain for each term with a site: query. If two pages come back, merge them, redirect one, or point it at a different intent and rewrite its title and headings to match. Then point internal links on that topic at the owning page, with anchor text that names the intent.

Write the map down and check it against what happens

A keyword map is a scope document. It records which pages exist, what each is for, and what would count as it working. Writing that down before work starts is what makes the programme reviewable later. Review it each quarter against enquiries rather than positions, because a term can rank and still send nobody who buys. When a target is missed, the map shows where: the intent was misread, the first page was never available, or the page ranks for a phrase nobody acts on. We report the months that miss target too. The map stays with the client, like the analytics and ad accounts the work is judged in.

---

Written by Fibvo — Bengaluru and Erode, India — working with clients in other time zones. Questions, or think we have it wrong? Tell us.