Sourced analysis · Documented facts are linked to their primary references. Interpretation is identified in the text.
Start with the event being counted
The moment a viewer encounters a drama is not always the moment a platform records a download. An advertisement can show part of an episode. A social clip can circulate independently of the app. A returning viewer can open an existing installation. These are different observations, and a useful audience story begins by naming which one is visible.
Sensor Tower’s short-drama report measures estimated downloads through the App Store and Google Play, with exclusions. That creates a valuable acquisition measure, but it cannot establish how many people finished a particular series. The first distinction in our audience coverage is therefore between reaching a device and retaining a viewer. [Sensor Tower]
Follow the viewer through the funnel
For a single title, the useful sequence is exposure, first play, continued viewing, completion and return. A payment introduces another branch: a viewer can enjoy a free chapter without becoming a paying customer. A person who pays once may never return. Different questions require different denominators.
A completion rate should name what was completed and which starters were included. Retention needs a cohort and a time window. Revenue per paying user excludes people who never pay. None of these is inherently the best metric; the mistake is substituting whichever number looks largest for the one the argument needs.
Fandom needs evidence of its own
Fandom describes a relationship with a story, performer or community. Repeat discussion, fan translation, close reading and organised recommendations can be evidence of that relationship. A screenshot of a large view counter cannot show it by itself.
An interview-based fandom feature would need identified participants, consent, context and a clear account of how they were found. This edition does not invent fans or infer their identity from platform marketing. Until direct reporting supports a demographic claim, our interpretation stays with the documented viewing interface and the observable story premise.
What a good audience claim includes
Keep the source, date, territory, unit and denominator together. Distinguish a platform-wide estimate from title-level evidence. If two sources measure different populations, show both scopes rather than averaging their totals. That discipline makes comparisons smaller but more useful.
Our research room preserves the earlier audience dossier, including its disputed measurements. Those entries remain reporting leads. The reader-facing figures used here have their own source trail and are not upgraded merely because they appeared in the original spreadsheet.
Sources & evidence
Sensor Tower · Vendor estimates
Source date: 2026-06 · Checked: 15 Sept 2026
- Q1 2026 estimated downloads exceeded 850 million; IAP revenue roughly US$750 million.
- Southeast Asia 32%, Latin America 23%, India 22% of Q1 2026 downloads.
- Download scope is App Store and Google Play, with stated exclusions.