Keyword research is the process of identifying what your audience actually searches for, understanding the intent behind those searches, and prioritizing which ones are worth targeting based on relevance, realistic ranking opportunity, and business value — not just raw search volume.
Step 1: Start from real questions, not guesses
Before any tool, gather the actual questions your audience asks:
- Support tickets and sales call questions
- Questions asked in your AI chatbot (if you have one — see What Is an AI Chatbot?)
- Comments and questions on competitor content
- "People also ask" and related-search boxes on search results pages
This grounds your keyword list in real demand instead of assumptions.
Step 2: Expand with a keyword research tool
Use a keyword tool to expand your seed list and see approximate search volume, but treat volume as directional, not precise — actual numbers vary between tools and change over time. Look for:
- Related terms you hadn't thought of
- Question variations ("what is X" vs "how does X work" vs "X vs Y")
- Volume trend (growing, flat, or declining interest)
Step 3: Classify by search intent
For each keyword, determine which of the four intent types it represents — informational, navigational, commercial investigation, or transactional (see Search Intent Explained). This determines what kind of content you should build, not just what to title it.
Step 4: Assess realistic opportunity
For each keyword, check what's currently ranking and ask:
- Can I realistically compete? Look at whether top results are from large, highly authoritative sites across the board, or a mix including smaller/newer sites.
- Is there a genuine gap? Are current top results outdated, incomplete, or poorly structured? That's your opening.
- Does it fit your topical cluster? A keyword you can rank for but that's unrelated to your core topic areas contributes less to overall topical authority.
Step 5: Prioritize by relevance × opportunity × value
Don't just sort by search volume. A lower-volume keyword that's highly relevant to your business and has a realistic ranking path is often better to target first than a high-volume keyword you have little realistic chance of ranking for yet.
A simple prioritization framework:
| Factor | Question |
|---|---|
| Relevance | Does this genuinely fit what we do and want to be known for? |
| Opportunity | Is there a realistic path to ranking given current competition? |
| Value | If we rank, does the resulting traffic matter to the business (leads, signups, ad revenue)? |
Step 6: Group into clusters, not a flat list
Organize your prioritized keywords into topic clusters — a pillar topic with related subtopics — rather than treating each keyword as an isolated article target. This is what makes internal linking and topical authority possible, and it avoids accidentally writing multiple thin, competing pages for near-duplicate keywords.
A common mistake
Chasing the highest-volume keyword in a topic area first, rather than building the full cluster around it. A single high-volume "pillar" article without supporting cluster content usually underperforms a well-connected cluster of moderate-volume pages, both for traditional rankings and for the topical authority that matters for AI search — see What Is GEO?.