The raw material is a list of queries. It comes from the business itself (what customers ask, how they describe the problem), from the search engines' own suggestions and related-search boxes, from competitors' pages, and from tools that hold search data. Google's Keyword Planner, built for advertisers, gives an estimate of monthly searches, a competition level and the average cost for an ad to show for each query; the SEO suites from Semrush, Ahrefs and others add an estimated difficulty of ranking organically and the pages that currently do. The output is a spreadsheet of queries with four columns: what it means, how many, how hard and how much.
- Informational: the searcher wants to know something (how does a limited company pay tax). Served by guides; converts slowly or never.
- Navigational: the searcher wants a specific site (hmrc login). Not worth competing for.
- Commercial: the searcher is comparing (best accountant for contractors). Served by comparisons and reviews; converts well.
- Transactional: the searcher is ready to act (hire accountant leeds). Served by a service page with a price and a form; converts best and costs most.
Intent is the column that decides everything else. The four classes above are the standard reading, and a results page usually shows which one the engine has decided on: a query answered by ten guides is informational whatever the business wishes it were, and a service page will not rank for it. Long-tail queries, the specific phrases with few searches each, make up most of the volume in aggregate and most of the conversions, because specificity is intent. A page built for one head term and a dozen of its long-tail variants outperforms a page built for the head term alone.
The same research feeds two channels. For pay-per-click it becomes the keyword list and the negatives; for SEO it becomes the page plan, with one page per intent cluster rather than one per keyword. Semrush's Content-Led SEO certification and the SEO courses teach the tool side; the Google Ads Search certification teaches it from the paid side, with match types that decide how loosely a keyword matches a query. The skill that neither tool teaches is reading the results page to see what the searcher wanted.
Reading a results page
Search the query and look at the ten results before the numbers: guides mean informational, comparisons mean commercial, product and service pages mean transactional. A page type the engine does not show for the query will not rank for it.