Keyword clustering that groups by search results, not just words
You exported 300 keywords and now have no idea which ones belong on the same page. Grouping by similar words gets it wrong: "apple pie recipe" and "apple pie calories" can need different pages. This prompt has your todo.is agent cluster your list by intent and overlapping search results, then map each cluster to one page.
The prompt
- Cluster this keyword list: [ATTACHED FILE OR PASTED KEYWORDS]. The site is [WEBSITE] and targets [MARKET]. Group keywords that should be answered by the same page. Decide by search intent and by checking whether the top Google results overlap for the main keywords in each group, not just by shared words. For each cluster give: a cluster name, the main keyword, supporting keywords, intent (informational, commercial, transactional, local), suggested page type, and whether an existing page on my site already covers it (with the URL) or a new page is needed. Keep keywords with no clear fit in an "unsorted" tab. Return an Excel file with a cluster summary tab and a full keyword-to-cluster tab, sorted by [PRIORITY].
What to change
- [ATTACHED FILE OR PASTED KEYWORDS]: Attach a CSV or Excel export from your keyword tool or Search Console, or paste the list.
- [WEBSITE]: Your site, so your agent can match clusters to existing pages.
- [MARKET]: Country and language, e.g. "US, English" or "Germany, German".
- [PRIORITY]: How to sort, e.g. "total search volume", "business value" or "easiest wins first".
Example result
- Keyword clusters · brewhaus-supply.com
- Market: US, English. 286 keywords in, 31 clusters out, 14 unsorted.
- Cluster 1 · French press guide
- • Main keyword: how to use a french press
- • Supporting: french press ratio, french press brew time, how much coffee for french press, french press instructions
- • Intent: informational
- • Page type: step-by-step guide with a ratio table
- • Status: existing page /blog/french-press-guide covers it. Add the ratio table.
- Cluster 2 · Buy a french press
- • Main keyword: best french press
- • Supporting: stainless steel french press, glass french press, large french press
- • Intent: commercial
- • Page type: comparison or category page
- • Status: new page needed
- • Why separate from Cluster 1: the results for "how to use a french press" and "best french press" barely overlap. One wants instructions, the other wants products.
- Cluster 3 · Grind size
- • Main keyword: coffee grind size chart
- • Supporting: grind size for pour over, espresso grind size, coarse vs fine grind
- • Intent: informational
- • Status: new page. A printable chart could earn links.
- Cluster 4 · Local
- • Main keyword: coffee beans near me
- • Intent: local. Results are map listings.
- • Status: skip for the blog; better handled by a Google Business Profile.
- How the clustering was done
- • Keywords were grouped when their top results showed mostly the same pages.
- • Close variants (plural, word order) were merged into one keyword.
- • Unsorted keywords are too vague, like "coffee", or off-topic for the store.
- Note: search volumes come from your file. Your agent did not add or estimate volumes.
How to do it with todo.is
- Export your keywords as CSV or Excel and attach the file to the prompt.
- Replace the [brackets] and paste it into todo.is, or send it with the file to your agent on Telegram.
- Your agent clusters the list and sends an Excel file with a summary tab and a full mapping tab.
- Ask follow-ups like "write a content brief for cluster 3" to start on the pages.
Tips for a better result
- Cluster before you write. It stops two of your own pages from competing for the same search.
- When the top results for two keywords are mostly the same pages, one page can usually rank for both.
- Keep local and transactional clusters separate from blog clusters. They need different page types.
- Start with clusters that match an existing page. Improving a page is faster than building a new one.
keyword clustering: FAQ
- What is keyword clustering? It is grouping related keywords so each group is targeted by one page. It helps you plan content and avoid several pages competing for the same search.
- What is the difference between semantic and SERP clustering? Semantic clustering groups keywords by meaning or shared words. SERP clustering groups them by whether Google shows the same pages for them, which is closer to how Google treats them.
- How many keywords should be in a cluster? There is no fixed number. Some clusters have 2 keywords and others have 30. What matters is that one page can fully answer all of them.
JavaScript is required to use the todo.is app.