Keyword research goes wrong when it becomes data collection instead of decision-making. The goal is a short list of pages worth building, each with a clear job.
Step 1: Gather from more than one place
- Search Console: queries you already receive impressions for are the cheapest opportunities available
- Competitor rankings: what brings them traffic that you have no page for
- Autocomplete and people-also-ask: real phrasing and follow-up questions
- Your sales inbox and call notes: the words customers actually use
Step 2: Classify by intent
Sort every query into informational, commercial, transactional or navigational. Then look at the current results — Google's own answer tells you what type of page it considers correct. If the first page is all guides, a service page will not rank there no matter how well optimised.
Step 3: Judge winnability honestly
Ignore the difficulty score for a moment and look at who ranks. If the top ten is entirely marketplaces, directories and national brands with thousands of referring domains, a new local site will not break in this year. Find the adjacent query where the competition is beatable.
Step 4: Map one intent per page
Assign every priority query to exactly one URL. If two pages target the same intent, merge them. Cannibalisation splits your signals and both pages underperform.
Step 5: Write briefs, not keyword lists
For each planned page, note the intent, the questions to answer, the format the results reward, the internal links to include and the action you want the reader to take. That brief is the actual output of keyword research.
Questions about applying any of this to your own website? Request a free consultation.