Keyword research template with intent mapping
Volume, difficulty, and intent scored against business relevance — then clustered into the pages you should actually build.
- DataForSEO
- Google Search Console
- Google Sheets
- Sanity / CMS
Run keyword research for "ai agents for marketing". Score against our ICP of B2B growth teams and cluster into pages.
- Expanded 4 seeds into 612 candidatesPulled related terms, questions, and long-tail variants
- Scored relevance against ICPDropped 388 — consumer, design, and no-fit commercial terms
- Clustered 224 survivors19 clusters by shared intent and SERP overlap
- Assigned page types and checked cannibalization11 guides, 5 comparisons, 3 product pages; 2 clusters conflict with live pages
Nineteen clusters are worth building, led by "ai agents for marketing" as a pillar with four comparison clusters beneath it. Two clusters would cannibalize existing pages — fold those into the current URLs instead of publishing new ones.
A keyword research template that stops at volume and difficulty leaves the expensive decision unmade: which of these terms is worth a page. This one scores every candidate against your positioning and buying intent, clusters what remains, and tells you the page type each cluster needs.
What is a keyword research template?
A structured way to move from seed terms to a decision. It records volume, difficulty, and cost per click for each candidate, adds intent and business relevance, then groups survivors into clusters that map to individual pages rather than a flat list you sort by volume.
Expand and qualify seed terms
- Pull related terms, questions, and long-tail variants per seed
- Attach volume, difficulty, and commercial value to each candidate
- Score business relevance against your positioning and ICP
- Drop terms your product cannot credibly serve
Cluster and assign page types
- Group terms by shared intent and SERP overlap
- Assign a format per cluster: guide, comparison, tool, or product page
- Flag clusters that would cannibalize existing pages
Prioritize what to build first
- Rank clusters by expected return against effort
- Separate quick wins on existing pages from net-new builds
- Note the AI answer sub-queries each cluster should also cover
How do you map search intent to page types?
Informational queries want explainers and guides, commercial queries want comparisons and listicles, and transactional queries want product or pricing pages. The template classifies each cluster and assigns the format, which is what stops teams from writing blog posts for queries that need a product page.
How should you use keyword difficulty?
As a filter relative to your own authority, not as an absolute. The template reads the actual ranking pages: if page one is dominated by domains far stronger than yours with no thin results among them, difficulty is real regardless of the score — and a low score with three forum threads ranking is an opening.
Why does keyword clustering matter more than a keyword list?
Because one page can rank for dozens of variants. Clustering groups terms that share intent and SERP overlap, so you build one strong page instead of five thin ones competing with each other for the same query.
Can you export this as a keyword research spreadsheet template?
Yes. The output is structured so you can paste it into Sheets or Airtable with volume, difficulty, intent, cluster, assigned page type, and priority as columns — the difference is that the scoring and clustering arrive already done.
How the keyword research template works across your stack
The agent pulls volume, difficulty, cost per click, and SERP composition from live keyword data, reads your existing pages to check for cannibalization, and uses workspace Knowledge — positioning, ICP, tracked competitors — to score relevance. Output goes to the content roadmap template, a spreadsheet, or straight into briefs.
- DataForSEO
- Google Search Console
- Google Sheets
- Sanity / CMS
Who uses this keyword research template
- Content leads
- Turn a quarter of ideas into a ranked, defensible build list.
- Agencies
- Run the same scoring model across accounts so recommendations stay comparable.
- Founders doing their own SEO
- Find the handful of queries you can realistically win this quarter.
How to run this keyword research template in Metaflow
Give the agent seed topics
Three to ten seeds covering your category, the problem you solve, and your competitors.
Confirm relevance criteria
The agent proposes criteria from your positioning; adjust if a segment matters more this quarter.
Review clusters, not keywords
Each cluster arrives with its terms, intent, assigned page type, and a cannibalization check against your site.
Send winners to briefs or the roadmap
Approved clusters flow into the content brief template or the SEO content roadmap.
What you provide
- Seed topics or competitor domains
- Target geography and language
- Optional: existing page inventory
What you get back
- Scored keyword table with volume, difficulty, and intent
- Clusters with assigned page types
- Cannibalization warnings
- Prioritized build order
Why use this keyword research template?
Scores business relevance, so high-volume irrelevant terms get dropped
Delivers clusters and page types rather than a flat keyword list
Checks cannibalization against pages you already have
Includes the sub-queries AI answer engines fan out to
Keyword research template FAQs
What should a keyword research template include?
Volume, difficulty, cost per click, intent, business relevance, cluster assignment, target page type, and a cannibalization check. The last three are what most spreadsheet templates omit, and they are where the actual decisions live.
How do you find keywords worth targeting?
Filter for queries your product can credibly serve, then check whether the ranking pages are beatable. High volume with no commercial fit is a distraction; moderate volume with weak competitors and clear buying intent is the opportunity.
How many keywords should one page target?
One cluster, which usually means one primary term and ten to fifty variants. Splitting a cluster across pages creates competitors on your own domain.
Does this account for AI search and answer engines?
Yes. Alongside standalone queries, the template captures the sub-queries assistants generate when decomposing a prompt, since those are what your page has to answer to get cited.
Similar SEO templates
Key takeaways
- Relevance filters before volume sorts — most high-volume terms are noise
- Clusters map to pages; keyword lists map to nothing
- Check cannibalization before commissioning any new page