Keyword clustering and topic research
Group keywords that one page can answer. The results page shows you which keywords belong together.
- Time
- 18 min
- Level
- intermediate
- Reviewed
- 5 November 2026
After this lesson you can
- Group keywords into topics by meaning, intent and results page overlap.
- Decide whether two keywords need one page or two.
- Turn clusters into a topic structure for a site.
After the last two lessons you have a long list of keywords. Most of them are not separate targets. They are different ways of asking for the same thing. Clustering groups them so that each group can be answered by one page.
Three tests for "same cluster or not"
- Meaning: would a person accept the same answer for both queries?
- Intent: are both queries informational, both transactional, both local?
- Results overlap: search both and compare the top 10. If many of the same URLs rank for both, Google treats them as the same need.
Example
| Keyword | Intent | Cluster |
|---|---|---|
| nešiojamo kompiuterio remontas | Transactional, local | laptop repair |
| nešiojamų kompiuterių taisymas | Transactional, local | laptop repair |
| laptop repair siauliai | Transactional, local | laptop repair |
| nešiojamas kompiuteris neįsijungia | Informational | laptop will not turn on |
| ką daryti jei kompiuteris neįsijungia | Informational | laptop will not turn on |
| nešiojamo kompiuterio ekrano keitimas kaina | Transactional | laptop screen replacement |
The first three can be answered by one service page. The next two need a helpful article, which can link to the service page. The last one may deserve its own page if the business offers screen replacement and people search for it specifically.
From clusters to site structure
Clusters often form a hierarchy:
- Broad topic: computer repair, which can be a hub or category page.
- Subtopics: laptop repair, desktop repair and data recovery, each a service page.
- Supporting questions: articles that answer specific problems and link to the relevant service.
This structure is often called a topic cluster or hub-and-spoke model. The content module covers how to build one. For now, aim for a list of clusters where each could clearly become one page.
Tools for clustering
Some paid tools cluster keywords automatically by comparing results pages. They save time on large lists, but check their groupings. For a list of a few hundred keywords, sorting by hand with the three tests above works well and teaches you more about your market.
Practice
Read the lesson, pass the quiz and do the exercise to complete it.
Knowledge check
4 questions. Get at least 4 right to pass.
Exercise: Cluster your keyword map
Assign every relevant keyword in your Workspace keyword map to a topic. For at least three pairs you are unsure about, compare the top 10 results and count the shared URLs. Mark one keyword in each topic as the main keyword.
Deliverable: A keyword map where every keyword has a topic, plus a note with the overlap counts for the three pairs and your decision for each.
Sources
Checked against these sources on 5 November 2026. Next review by 5 May 2027.
- How results are automatically generated and ranked (meaning of the query) (Google Search (How Search Works), official)
- SEO Starter Guide (expect your readers' search terms) (Google Search Central, official)
- A guide to Google Search ranking systems (neural matching, BERT) (Google Search Central, official)