Keyword clustering tool

Why group keywords by Google's results

Keyword clustering answers one question: should these keywords share a page, or each get their own? Guessing from the words alone goes wrong both ways. "cat tree for large cats" and "big cat tower" look different but are the same search to Google. "seo audit" and "seo audit template" look nearly identical, yet one wants a guide and the other wants a downloadable file.

The most reliable referee is Google itself. If Google ranks mostly the same pages for two keywords, it already treats them as one need, and one page can rank for both. If the top results barely overlap, Google sees two different needs, and one page will struggle to rank for both.

That's the method here. For each keyword we take Google's top 10 results, then compare every pair. Two keywords that share at least 3 of their 10 results belong together. We also require every keyword in a group to share at least 3 results with every other one, so a keyword that sits between two topics can't drag them into one oversized group.

How to read the groups

• Each card is one page to write or improve. It's named after its most searched keyword, which is usually the best main keyword for the page.
• One page for these lists the keywords that page can target. Use the main keyword in the title and headline, and the others naturally in headings and copy.
• Its own page means Google ranks different results for that keyword than for anything else on your list.
• Write tells you what kind of page Google ranks for the group: a "best of" list, a how-to guide, a product page and so on, read from the titles and addresses of the shared results. If the results are mixed, Google hasn't settled on one kind of page, which is often an opening.
• The shared results show which pages already rank for the group. Those are the pages to beat.

Common clustering mistakes

• One page per keyword. It's the most common mistake. Ten thin pages for ten variations of one search compete with each other, and usually none of them ranks well.
• One page for everything. The opposite mistake: a giant page that tries to cover separate needs ranks for none of them properly.
• Clustering by meaning alone. Grouping by topic in a spreadsheet, or with an AI tool that never looks at Google, often merges keywords Google keeps apart, and splits ones it treats as the same.
• Ignoring the page type. Two keywords in one group still need the right kind of page. If Google ranks product pages, a blog post won't break in, however good it is.

What this tool can't see

The results are Google in the US, in English, on desktop, at the moment you run it. Results shift over time, so a group on the edge (exactly 3 shared results) can split or merge on another day.

It doesn't know your site. Before writing a new page for a group, check whether you already have a page that targets it; improving an existing page is usually faster than starting over. And it groups only the keywords you give it, so start from a good list: our keyword research tool and related keywords finder are good places to build one.