Keyword Clustering: Grouping Terms Into One Post
Keyword clustering means grouping search phrases that one page can satisfy, so you write a single strong post instead of several overlapping ones. The simplest test is to search both phrases and see whether the same pages rank for each.
If the results are largely the same, the phrases belong to one cluster and one post. If they differ, they need separate posts. Here is how to do this by hand without special software.
Why clustering matters
When two posts target nearly the same search, they compete with each other. Google may rotate which one it shows, and neither builds strength. A single, more complete page tends to do better than two thin ones.
Clustering also saves work. Instead of writing ten short posts for ten close variations, you write one thorough post and cover the variations inside it.
Finally, it improves the reader’s experience. A visitor who lands on a page wants the whole answer, not one fragment with a link to another fragment.
The overlap test
The core method is simple. Take two keywords and search each one. Compare the top ten results.
- Most results overlap: the phrases share an intent, so they go in one cluster.
- Some results overlap: the phrases are related, and probably need separate posts that link to each other.
- Few results overlap: they are different searches, and should not share a page.
Google’s own results are the best judge of whether two phrases mean the same thing. Your intuition about wording is less reliable than what actually ranks.
Do it in a spreadsheet
List your keywords in a column. In the next column, write the URL of the top result for each. Phrases whose top results match are candidates for one cluster.
Then add a “cluster name” column and give each group a short label, such as “repotting snake plants.” Pick the phrase with the clearest wording and the most demand as the primary keyword, and list the others as secondary keywords in the same row.
For a list of a hundred phrases, this takes an hour or two. It is slow, but it protects you from a plan full of duplicates.
Choosing the primary keyword
Within a cluster, the primary keyword goes in the title and first paragraph. Choose the phrase that is most natural, since you will have to read it in your own headline, and that has the strongest demand.
Secondary keywords appear in headings, subheadings, and the body where they fit. You do not need to force every variation in. If you answer the question well, the related phrasing tends to appear naturally.
Structure the post around the cluster
A good cluster post is organized so each secondary phrase has a place. If your cluster is “how to repot a snake plant,” “when to repot a snake plant,” and “best soil for snake plants,” your headings might be:
- When a snake plant needs a bigger pot
- What soil to use
- Step-by-step repotting
- Aftercare for the first few weeks
Each heading answers one of the grouped searches, and the reader gets a complete guide. Place the primary answer first so searchers see it quickly.
When to split a cluster
Clusters can grow too large. If a post would pass several thousand words, or if sections serve different readers, it may be better to split. A beginner guide and an advanced troubleshooting guide, for example, often deserve separate pages.
Use the overlap test again on the sub-phrases. If a subset has a distinct results page, give it its own post, and link the two.
Link clusters together
Once you have several cluster posts on a subject, link them in a way that reflects the topic. A broad guide can link out to the detailed posts, and each detailed post can link back. This shows readers and search engines how the subject fits together.
Use descriptive anchor text, such as “how to choose a pot,” not “click here.” The link text tells both the reader and the crawler what to expect.
Revisit clusters after publishing
After a post has been live for a while, check Search Console for the queries it ranks for. You may find phrases that belong in the cluster that you missed, or phrases that suggest a separate post.
Update your spreadsheet when the data changes your view. Clusters are a working model, not a permanent classification.
A short example from start to finish
Suppose your list contains “how to write a cover letter,” “cover letter format,” “cover letter examples,” and “how long should a cover letter be.” You search each and compare the results.
The first two share most of their top results, because both want a walkthrough of what a cover letter contains. They become one cluster with “how to write a cover letter” as the primary keyword. The third mostly shows pages full of samples, so it needs its own post. The fourth shares enough results with the first cluster that it fits as a section inside it.
You end up with two posts instead of four. Each is more complete, and neither competes with the other. The same step repeated across a long list can remove dozens of duplicate plans before you write a word.
Watch for clusters that mislead
Be careful when overlap comes from a single dominant page. If one huge site ranks first for many related phrases, it may be there on authority, not because the phrases truly share intent. Look at positions two through ten as well before deciding.
What to do next
Take the twenty keywords on your current list and run the overlap test on every pair that looks similar. Group the matches in your spreadsheet and assign each group a name and a primary keyword.
Pick the cluster with the best demand and weakest competition. Outline one post with a heading for each secondary phrase, and write it as the complete answer for that subject. Add the other clusters to your schedule, and delete any planned posts that turned out to be duplicates.