Solutions · Brands and insight teams
Social listening for a platform where the message is spoken.
Listening tools were built for text. On TikTok the opinion is in the voice and on the screen, and the caption is often three hashtags. We measure the video itself.
What is said
Speech and on-screen text are turned into structured records: who or what the video is about, and the stance the speaker takes toward it.
What is seen
The same measure weighted by views. A position held in a thousand small videos and a position watched forty million times are different facts.
What is endorsed
The top comments under a video, weighted by their likes. The audience often ratifies the opposite of what the creator said.
Who gets the views
Share of videos and share of all views, grouped by how many views a video had when sampled.
- Share of videos
- Share of all views
0.5% of videos passed 100,000 views, and they hold 67.8% of all views. The median video had 303 views.
Table
| Views | Share of videos | Share of all views |
|---|---|---|
| 0 | 3.7% | 0.0% |
| 1–99 | 18.5% | 0.2% |
| 100–999 | 63.3% | 6.5% |
| 1K–10K | 11.9% | 8.3% |
| 10K–100K | 2.1% | 17.3% |
| 100K+ | 0.5% | 67.8% |
| Text-first listening tools | Clamor Data | |
|---|---|---|
| Source of meaning | Captions, hashtags, comments | Speech, on-screen text, top comments |
| Unit | Mentions and a sentiment score per post | Stance toward a named target, per region and period |
| Weighting | Counts of posts | Unweighted, view-weighted and like-weighted, kept apart |
| Sampling | Keyword and hashtag queries | A random sample of uploads, so shares mean something |
| Uncertainty | Rarely stated | Sample size and interval on every figure |
| Output | Dashboards of individual posts | Aggregates only: no creator profiles, no post-level export |
Why do views need to be separated from posts?
Because attention on TikTok is extremely concentrated. In our sample the top 1% of videos hold about three quarters of all views. A count of posts tells you what people make; a view-weighted measure tells you what people watch. The two regularly disagree, and the gap is often the finding.
Can I look up what a specific creator said?
No. We publish aggregates with a minimum cell size and do not sell creator-level data, post-level exports or transcripts. If your use case needs to monitor individuals, this is the wrong product.
How do we start?
Request a sample. Tell us the brands, topics or regions you care about and we send a sample dataset on the production schema, with its sample sizes, so your team can judge it before any contract.
Request a sample dataset.
Tell us the market and the question. We return a real slice, with schema, sample sizes and confidence intervals included. Not a brochure.