All articles

5 September 20263 min read

Keyword research and search intent

How I decide what a page should target: intent before volume, one page per intent, and the terms that are not worth a page at all.

  • Keyword research
  • Search intent
  • Content strategy

Most keyword research I inherit is a spreadsheet sorted by search volume. That tells you what is popular. It does not tell you what to build, and it is how sites end up with four pages competing for the same query and none of them winning.

Intent first, volume second

Before I look at a number I look at what the searcher wants. In practice there are four answers, and each one produces a different page.

  • They want to learn. A guide or an explainer.
  • They want to compare. A comparison, a pricing page, a case study.
  • They want to buy or enquire. A service or product page with a clear next step.
  • They want a specific place or brand. A location page, a contact page, a branded landing page.

The fastest way to check is to search the term and read what already ranks. If the results are all guides, a service page will not break in no matter how well it is built. That is not a ranking problem, it is a format mismatch.

One intent, one page

My rule is one page per intent, and never two pages for the same intent. When two of your own pages target the same thing, you split internal links and external links between them and confuse which one should be shown. Search Console usually shows the symptom clearly: the same query cycling between two URLs across weeks.

The fix is a decision, not more content. Pick the stronger page, fold the useful parts of the weaker one into it, and redirect.

What I do with volume and difficulty

Volume and difficulty are useful, with two conditions attached. Use the same exact term and the same country for both, and treat them as evidence rather than a forecast. Low reported volume does not mean nobody searches, especially for high value commercial terms in a narrow market. Low reported difficulty does not mean you will rank, because it says nothing about whether your site has any authority in the subject.

For a business selling in Britain I use British data. If I have to reason from another country's numbers, I say so and I say that local demand is unproven.

Mapping terms to templates

The output of research is not a keyword list, it is a map. Each row states the term, the intent, the page that should own it, whether that page exists, and what has to change on it. Where no page exists, that is a brief. Where two pages exist, that is a merge.

This is the document that makes the work reviewable. Anyone should be able to open it and see why a page is being written before it gets written.

Terms not worth targeting

I will argue against a keyword when any of these is true. The results are entirely dominated by aggregators or marketplaces the site cannot realistically displace. The intent does not match anything the business actually sells. The term is a near duplicate of something already ranking, so the honest move is to strengthen that page instead. Or the search is answered fully in the results themselves, so even ranking first earns very little.

Saying no to a keyword is cheaper than writing a page for it and then maintaining that page for years.

How I know it worked

Not by rankings alone. I watch impressions and clicks for the mapped terms on the mapped URL, and whether Google is showing the page I intended for the query I intended. When it shows a different page, the map was wrong and I go back to it.

Want this run on your site?

Tell me the domain and the pages that matter. I will come back with what I would fix first and what it should move.

Start a project