Overview
Follow this article to learn how audience clustering works within TikTok Insights and how Audiense organizes audiences into meaningful audience communities called clusters.
Once an audience has been processed, Audiense automatically groups users into distinct, non-overlapping clusters based on shared affinities, interests, and profile characteristics. These clusters help organizations better understand how different communities exist within a larger audience and provide a more actionable way to analyze audience behaviors, interests, and identities. Rather than relying only on demographic splits, TikTok Insights clustering is designed to surface naturally occurring communities based on how people engage with content and describe themselves online.
How Audience Clustering Works
After a TikTok audience is processed, Audiense assigns every audience member to a single cluster.
Clusters are:
- Distinct
- Non-overlapping
- Fully inclusive of the analyzed audience
This means:
- Every audience member belongs to exactly one cluster
- No users appear in multiple clusters
AI-Generated Cluster Naming
Each cluster is automatically assigned a suggested name generated by AI based on the characteristics most strongly associated with the audience community.
These names are designed to provide a quick understanding of the cluster and can be edited at any time to better align with internal terminology or reporting needs.
Choosing the Number of Segments
Users can choose how many clusters should be generated within a report.
TikTok Insights currently supports between 2 and 20 clusters per report. If no cluster count is specified, Audiense automatically selects a default structure based on audience size. Fewer clusters generally create broader audience communities, while more clusters reveal more detailed distinctions within the audience.
What Each Cluster Includes
Each generated cluster includes its own audience profile and insight set.
Cluster outputs include:
- A cluster name generated by AI
- Cluster size as a percentage of the total audience
- Audience insights describing the users within the cluster
