How to Interpret the Content Clustering Visualisation in Screaming Frog

Screaming Frog’s Semantic Content Cluster Diagram is a game-changing visualisation tool that helps SEOs quickly understand the true structure of a site’s content by understanding the semantic relevance of each page. This blog post walks you through exactly how to interpret the visualisation and what the colours, dots, and distances actually mean.

Need to understand more about semantic SEO and entities? Read our guide to Google's knowledge graph.

What Is the Semantic Content Cluster Diagram

The Semantic Content Cluster Diagram is generated by analysing the text on each page of your website using natural language processing. Each page is converted into a high-dimensional vector that captures its semantic meaning, and then reduced into a 2D space for easy visualisation.

In simple terms: it's a map of your content based on what the pages are about, both topically and by the ideas they communicate.

What Do the Dots Represent?

• Each dot = one page (URL) from your site.

• Proximity between dots = semantic similarity. Pages that are close together express similar ideas.

• Distance = dissimilarity. Pages that are far apart are about different things.

So, if you see a tightly packed group of dots, those pages all cover similar types of content. If you see a lone dot off to the side, that's called an outlier, which is a page that isn't semantically aligned with the rest.

What Do the Colours Represent?

By default, Screaming Frog assigns a different colour to each semantic cluster. These clusters are calculated using algorithms like k-means that group similar vectors together.

There are 3 methods of clustering that you can set in the options:

• K-means - Groups your content into a fixed number of clusters. By default this is set to 5, but can be increased or decreased depending on the size of your site. This can be useful for quick and consistent auditing
• X-means - The tool automatically detects the optimal number of clusters to group your content into. This can create more accurate clustering and make it easier for you to spot outliers, but it can get messy especially for larger sites.
• Segments - These are the segments that you define as part of your Screaming Frog settings (go to Configuration > Segments to set up). This is probably the most effective method to allow you to identify content that doesn't fit with your overall strategy. It also has the added benefit of adding a legend to your graph. However, it can take time and be tricky to set up.

Quantitative Ways to Interpret the Diagram

You can get real insights from the diagram by interpreting it in a structured way:

Visual Element	What It Means
Dot Closeness	Pages are semantically similar
Cluster Density	High density = strong topic group; scattered = weak focus
Number of Clusters	Reflects topical breadth/diversity
Outliers	Unconnected dots = content that may be irrelevant, thin, or unique
Colour Consistency	Uniform colour in one area = tightly defined topic

 

Semantic similarity vs. topical similarity

Imagine your site has these two pages:

• "How to stay motivated when working out alone"

• "How to stay motivated when studying for exams by yourself"

These would likely be shown close together in the diagram, even though they are from different topics. That’s because semantically, they cover the same concept: self-motivation in solitary tasks. This is what makes the diagram more powerful than traditional category-based site mapping.

Alternatively, if your website had a tutorial for the Google Ads dashboard and one for Google Search Console, the topics would be PPC and SEO but these pages would still be semantically similar and there may be an opportunity to link between them.

Now consider two blog posts that are both topically grouped under SEO:

• “How to fix crawl errors in Google Search Console”

• “The history of SEO: From keyword stuffing to semantic search”

These would both fall under the SEO category, and might even be in the same blog folder. But semantically, they are quite different — one is a hands-on technical guide, while the other is a historical overview. They may have the same colour dot on the graph but will likely be spaced quite far apart.

What Can You Learn from This?

1. Find Content Gaps - Are there important topics missing clusters?

2. Prune your Content - Are there outliers that should be removed or rewritten?

3. Improve Internal Linking - Are semantically similar pages linked? Are there opportunities to link between different topics?

4. Consolidate Similar Pages - Are too many pages covering the same intent?

5. Evaluate Content Strategy - If your content clusters suggest an unfocused content strategy, you may need to go back to the drawing board.

Pro tip: Export the raw data from your Screaming Frog content clusters with our free tool