Frequently Asked Questions About TikTok Insights
Find answers to common questions about TikTok data coverage, audience analysis, reporting methodology, Interest Categories, and MCP capabilities in Audiense Insights.
Overview
This article answers frequently asked questions about TikTok Insights reports, including how audience data is collected, how insights are generated, and how to interpret the information displayed in your reports.
Use the sections below to better understand TikTok audience analysis, reporting limitations, and available features.
Data and Coverage
What data are TikTok Insights reports based on?
TikTok Insights reports are built using publicly available TikTok content and the public profiles of users who publish it.
New profiles are added and existing profiles are updated as users publish content, helping keep the underlying dataset current.
Important: TikTok audiences represent users who actively publish content. Users who only watch or engage without posting are not represented, and comments are not collected or analyzed.
Is TikTok data sourced compliantly?
Yes. Audiense uses publicly available content accessed through official platform access and licensed data from trusted partners.
These sources are continually developed to provide reliable audience insights while maintaining data quality and compliance.
How much of TikTok does Audiense cover?
TikTok Insights provides global coverage across markets where TikTok operates.
The dataset launched with hundreds of millions of audience members and continues to expand, with approximately 7–10 million additional users added each month.
Is the data biased toward particular countries?
Audiense's geographic distribution closely reflects TikTok's global user presence.
China is not included because TikTok operates as a separate platform in that market.
The primary consideration when interpreting results is that the dataset represents users who publish content, rather than TikTok's entire viewing population.
How do TikTok Insights reports differ from X reports?
Both TikTok and X reports use publicly available content, AI, and machine learning to generate audience insights.
However, TikTok exposes less public profile information than X. As a result, more demographic and behavioral attributes in TikTok reports are inferred rather than obtained directly from profile data.
How Audience Insights Are Generated
How does Audiense determine age, gender, and location?
TikTok does not publicly expose structured age, gender, or location information in the same way as some other platforms.
Audiense uses AI models to estimate these attributes from available public profile information and published content.
For location, country information is available for many profiles. When it is unavailable, Audiense estimates the location using available signals.
Note: These attributes are probabilistic estimates intended to describe audience characteristics in aggregate, not verified information about individual users.
What content does Audiense analyze?
TikTok Insights analyzes two primary sources from published videos:
- Post captions: Text written by the content creator.
- Video transcriptions: AI-generated transcriptions of spoken audio within videos.
These inputs help identify topics, entities, content characteristics, and creator or brand affinities.
Audiense does not analyze comments or use TikTok's own AI-generated post summaries.
How is a profile's language determined?
Audiense identifies languages based on the content a profile publishes.
A profile can be associated with multiple languages, with one designated as its primary language.
What time range is available for TikTok Conversations reports?
New TikTok Conversations reports support a date range of up to 120 days, with available dates beginning in May 2026.
You can select a reporting window of up to 120 days within the available historical period.
Understanding Your TikTok Insights Report
Why don't my cluster sizes add up to the total audience?
Audience clusters are created using publicly available signals, such as profile biographies and interests.
If an audience member does not have enough information to be assigned to a cluster, they remain part of the full audience but are not included in an individual cluster.
This is expected behavior and does not indicate a reporting error.
The unassigned population is typically small, although it may be more noticeable in reports with smaller audiences.
Why do benchmarks in the Topics chart sometimes show 0.00%?
At the full-audience level, Topics are compared against a broad geographic baseline.
Because the number of possible topics is very large, individual topics may represent an extremely small percentage of the baseline audience, causing their values to round to 0.00%.
This is expected behavior.
When viewing an individual cluster, the comparison uses the full audience as its baseline, which may produce more visible benchmark values.
What is each audience compared against?
By default, Audiense Insights uses the following comparison structure:
- Full audience: Compared against the default baseline, typically a geographic audience.
- Individual clusters: Compared against the full audience.
Custom comparison baselines are also available.
Is my TikTok Insights report updated automatically?
No. Each report is a snapshot of the audience at the time it was generated.
The report does not automatically update as audience characteristics or activity change.
The same principle applies to comparison baselines, although Audiense regularly refreshes its default country baselines.
Can I compare audience changes over time?
Yes. Although individual reports are snapshots, you can compare different reporting periods.
Create separate reports using the same audience criteria but different date ranges. You can then use one audience as the comparison baseline for the other to explore how its characteristics have changed.
Why isn't TML available in the Influencers & Brands, Public Figures, and Hashtags tabs?
These insights are based on open-ended results rather than a fixed list of categories.
The same creators, public figures, or hashtags may not appear in both your selected audience and its comparison baseline, making a reliable Times More Likely (TML) calculation difficult.
For this reason, these dashboard tabs display audience percentages to show how strongly each result is represented within the selected audience.
Understanding Interest Categories, Topics, and Entities
What's the difference between Interest Categories, Topics, and Entities?
These insights describe different aspects of audience behavior and content.
- Interest Categories: Broad audience interests organized into predefined categories.
- Topics: Subjects and themes identified from the content an audience publishes. These are not restricted to a fixed list and may provide more specific insights.
- Entities: Named people, brands, products, or other identifiable subjects mentioned in published content.
In short, Interest Categories describe broad affinities, Topics describe conversation subjects, and Entities identify specific named things.
Where does the Interest Categories taxonomy come from?
The TikTok Insights Interest Categories taxonomy is developed by Audiense, not TikTok.
It is based on a refined subset of the Interactive Advertising Bureau (IAB) taxonomy, with categories adapted to reflect interests relevant to TikTok audiences.
Interest Categories are inferred from the content users publish.
Although X Insights also uses a refined IAB taxonomy, the TikTok and X taxonomies were developed separately and are not currently identical.
Are Interest Categories the same across Audiense products?
The underlying concept is similar, but the methodology and available categories can differ depending on the product and data source.
Audiense is working toward consistent terminology across its products.
| Product | Terminology |
|---|---|
| Insights (X) | Interest Categories |
| Insights (TikTok) | Interest Categories |
| SoPrism | Interest Categories (formerly Topics) |
| Top Interests | |
| Demand | Interests |
Does Influencers & Brands show the accounts an audience follows?
No. Unlike X, TikTok does not provide a publicly accessible social graph for this analysis.
The Influencers & Brands tab identifies inferred affinities based on published content and public profile information.
These results represent creators and brands associated with the audience's content, not verified following relationships.
Using TikTok Insights with Audiense MCP
Can I create TikTok audience reports through MCP?
Yes. Audiense MCP supports creating reports using conversation-based audience definitions for TikTok and X.
It also supports audience definitions based on selected profile attributes through Advanced Audience Manager.
Currently supported profile attributes include:
- Country
- Follows any handle
- Age
- Gender
- Biography
Note: MCP does not yet support every audience-definition option available in the Audiense Insights interface. Additional properties are planned for future updates.
Can I use custom baselines with Audiense MCP?
Yes. Custom comparison baselines are supported across MCP tools.
By default, MCP uses the same comparison logic as the Audiense Insights dashboard:
- Full audiences are compared against the default baseline.
- Clusters are compared against their full audience.
You can specify a custom baseline when a different comparison is needed.
Does Audiense MCP support Times More Likely (TML) scores?
Audiense MCP supports TML scores to help identify insights that are more or less prevalent in an audience compared with its baseline.
TML values help distinguish audience characteristics that are especially relevant or distinctive, rather than simply common.
The availability of individual metrics may depend on the insight type and the underlying comparison data.