| name | keyword-research |
| description | Research search demand by connecting real audience language, topics, query variants, intent, volume or trend evidence, difficulty context, product relevance, and content opportunity without chasing keywords detached from customer value. |
Keyword Research
Use when search strategy needs evidence about how people describe and seek a problem, product category, comparison, task, or answer.
Procedure
- Define the audience, product or content area, geography, language, and customer journey stage the research should support.
- Seed research from customer language, support, sales, forums, search suggestions, existing analytics, competitor topics, and product terminology rather than one keyword tool alone.
- Group query variants by underlying topic and intent so singular and plural or phrasing differences do not become artificial separate strategies.
- Gather current demand, trend, SERP, competition or difficulty, and click-opportunity evidence from available search tools while recording date and source.
- Evaluate business relevance, ability to satisfy the intent, product fit, and likely conversion or authority value alongside headline search volume.
- Identify long-tail, comparison, problem, educational, branded, local, or transactional clusters relevant to the audience journey.
- Map clusters to existing pages or planned content and flag cannibalization, coverage gaps, and terms the product should not target.
- Prioritize topics using demand, relevance, achievable quality, strategic value, and evidence rather than one opaque tool score.
Decision rules
- Search volume without audience or product relevance is not opportunity.
- Keywords are observations of language and intent, not instructions to repeat phrases unnaturally.
- Tool estimates are directional and should not be treated as exact traffic forecasts.
- Verify current search evidence when making decisions because demand and SERPs change.
Quality gate
Research is ready when terms are grouped by real intent, demand and trend evidence is current enough for the decision, product relevance and existing coverage are visible, priority reflects more than volume, and each target cluster has a credible user need the content can satisfy.
