Guides
What is keyword clustering? A plain answer with an example
Grouping searches that one page can answer, why it matters, the three methods, and how seogeoaeo.ai does it from a pasted list.
Keyword clustering is the practice of grouping search phrases that one web page can answer together, so a site builds one strong page per topic instead of many weak ones. You can do it by hand in a spreadsheet, or paste the list into a tool such as the seogeoaeo.ai Keyword Clusterer, which sorts it into groups with a suggested page path for each.
That is the short answer. The rest of this page covers why it matters, the three ways it is done, and what a cluster looks like.
Why it matters
Search engines rank pages, not keywords. People phrase the same need in many ways: "cheap flights to Rome", "Rome flights deals", "budget flights Rome". Write a separate page for each and those pages compete with each other, while none of them gets enough depth to rank well. Put them on one page and that page covers the topic properly.
Clustering also tells you how big a job is. A list of 400 keywords might turn into 40 pages, and you can plan and brief those 40.
A worked example
Start with this list:
- running shoes
- best running shoes
- trail running shoes
- running shoes for beginners
- email marketing
The first four share "running shoes" and serve a shopper comparing options, so they make one cluster and one page. "Email marketing" has nothing in common with them and belongs elsewhere. In a real list the groups are rarely this obvious, which is why people use software.
Three ways to cluster
| Method | The test for "same group" | Strength | Weakness |
|---|---|---|---|
| Shared words | Keywords contain the same meaningful words | Fast, cheap, easy to check | Misses synonyms with no shared words |
| Search result overlap | Google shows many of the same pages for both | Close to how Google sees intent | Slower, needs live search data |
| Meaning | A language model or embeddings rate them as similar | Catches synonyms | Harder to see why a keyword landed where it did |
seogeoaeo.ai's clusterer uses the shared-words test. It works from token overlap only, with no embeddings, language model or live search results, and it suggests a parent page path for each group. Treat that path as a starting point for checking intent, not a finished site map.
Tools such as Keyword Insights and Keyword Cupid use search result overlap. When a list is full of synonyms, or the same words hide different intents, that method is the better choice.
Keyword clusters and topic clusters
The two terms are related but not the same. A keyword cluster is a group of searches for one page. A topic cluster is a set of pages: one broad pillar page linked to several narrower pages around it. You usually build keyword clusters first, then arrange the resulting pages into topic clusters.
Doing it yourself
- Put every candidate keyword in one list.
- Remove duplicates and plural or hyphen variants. seogeoaeo.ai's Keyword Deduplicator does this, and its clusterer's page recommends running it first.
- Run the clusterer, at 3 credits a run.
- Read each group and split any that mix buyers with people still learning.
From there, each group gets a focus keyword and a URL, and you have a content plan.