Skip to main content

Keyword Clustering

Group keywords into semantically related clusters automatically — to plan site structure, assign keywords to pages, and streamline content creation.

Written by Mike

Group Keywords by Meaning — Automatically

The Keyword Clustering tool groups a large list of keywords into semantically related clusters — based on the URLs that appear together in Google search results for those keywords. Each cluster represents a topic that can be targeted with a single page, helping you plan site structure, assign keywords to existing pages, and prioritize content creation. Unlike manual grouping — which can take 2–3 days for 1,000 keywords — Serpstat clustering runs automatically in minutes.

What you can achieve:

  • Plan site structure from keyword data: Distribute your keyword list into clusters that map directly to pages, sections, and categories — giving you a data-driven site architecture.

  • Identify the right keyword for each page: See which keywords share SERP overlap and can be targeted together, versus which need separate pages.

  • Find unclustered keywords worth acting on: Keywords in "Unsorted" either need their own page or reveal gaps in your current content plan.

Getting Started

  1. Navigate to Keyword Clustering & Text Analysis in the left sidebar and click Add a new project.

  2. Enter a project name and optionally your domain — if you add a domain, Serpstat will automatically assign the most relevant page from your site to each cluster.

  3. Select search engine, country, and optionally region and city.

  4. Choose your Strength and Cluster Type settings (see Key Metrics below for guidance).

  5. Upload or paste your keyword list (3 to 50,000 keywords; CSV or TXT). Keywords must be clean — no duplicates, special characters, search operators, or extra spaces.

  6. Click Save to start clustering. Once complete, review clusters, edit or merge them as needed, and export in XLSX or CSV format.

Key Metrics Explained

  1. Strength (Weak / Medium / Strong): Controls how many common URLs two keywords must share in the top-30 search results to be grouped together. Weak = 3 common URLs, Medium = 8, Strong = 12. Use Weak for diverse keyword sets (e.g. multi-category stores); use Strong when keywords are already closely related.

  2. Cluster Type (Soft / Hard): Soft means a cluster forms if at least one pair of keywords meets the URL threshold — produces fewer, broader clusters. Hard means all keywords in the cluster must meet the threshold — produces more clusters of tightly synonymous terms. Strong + Hard gives the most granular, high-precision groupings.

  3. Connection Strength: A score from 0–100 showing how closely each keyword relates to its cluster's topic. Higher is better — low-scoring keywords may belong in a different cluster.

  4. Homogeneity: A score from 0–100 reflecting the semantic consistency of the whole cluster. A high homogeneity score means the cluster is tight and focused on one topic.

  5. Metatop: A list of the most frequent competitor pages appearing in SERP for keywords in the cluster. Higher-ranked pages in Metatop are the most relevant benchmarks to study when preparing content for that cluster.

  6. Unsorted Keywords: Keywords that didn't match any cluster under the selected settings. They may lack semantic similarity to the rest of the set, or they may need individual pages. You can manually move them to an existing cluster or leave them for separate planning.

Pro Tip: Choose Settings Based on Your Keyword Set

The right Strength + Cluster Type combination depends entirely on how semantically related your keywords already are.

  1. Tightly related keywords (e.g. sneaker brand variants, synonyms for one product): use Strong + Hard. This produces many small, precise clusters — each ideal for a dedicated page or category.

  2. Diverse keyword sets (e.g. keywords for a multi-product store or full-service agency): use Weak + Soft. This produces fewer, broader clusters that reflect natural topic groupings across a wide content scope.

  3. After clustering, look at unsorted keywords — if too many ended up there, consider running a second project with looser settings for that subset.

  4. Once clusters are ready, launch Text Analysis on any cluster directly from the clustering interface to get content recommendations for the assigned page.

⚠️ Important: Credits for clustering are shared with Text Analysis, Domain batch analysis, and Keyword batch analysis. Each keyword costs 5 credits: 20 keywords = 100 credits. Changing cluster type or strength without adding new keywords does not use additional credits. Adding new keywords or changing the country triggers a recalculation and charges credits for the added keywords.

❓ Frequently Asked Questions

Q: What is the minimum and maximum number of keywords I can cluster?
​A: You need at least 3 keywords to start a project. The maximum depends on your plan and available credits — up to 50,000 keywords per project.
​

Q: What happens if I change project settings after clustering?
​A: Changing only Strength or Cluster Type without adding keywords does not charge credits — but you must restart clustering to get updated results. Adding new keywords or changing the country charges 5 credits per keyword when the project is refreshed.
​

Q: Which plan includes Keyword Clustering?
​A: The tool is available from the Individual plan. Check the Tools block on the Plans & Pricing page to see how many monthly credits your plan includes.
​

Next Steps

  • Keyword Lists — Collect and prepare keyword sets for clustering using batch analysis.

  • Keyword Research — Overview — Find keywords to build your clustering project from.

  • Text Editor — Launch Text Analysis from any cluster to get content recommendations for your target page.

Did this answer your question?