For all data points that are growing rapidly, we might track them more often than the below.
- We track all Hashtags with more than 500 videos at least once every 24 hours.
- We track all Creators with more than 1,000 followers at least once every 24 hours.
- We track all Sounds with more than 500 videos created at least once every 24 hours.
- We discover new Sounds by looking at our existing database of creators and tracking when they post a new sound.
- We track all Favorites (if a user “stars” something) at least once every 24 hours.
- How does Trendpop gather demographics data?
- Some text signals (bio, video descriptions)
- Some visual signals (profile picture)
- We take a look at every single creator on the platform, and try to estimate their location, age, and gender by using a variety of signals:
- Are Trendpop’s demographics data accurate?
- We try to prioritize accuracy over data breadth.
- In a lot of cases, if we can’t collect enough signals, we won’t display anything.
How is your engagement calculated?
- Video Engagement Score
- [Likes + Comments + Shares] / Total Views
- Creator Engagement Score
- [Likes + Comments + Shares] / Total Views (for the most recent 30 videos)
- We run most sounds that are in our database through an audio reconciliation algorithm to try to match the audio to a track release.
- We won’t always be able to successfully match – especially if the audio qualities are excessively warped – but we’re constantly evolving our pipeline to improve our precision.
- Our systems are more likely to catch audio of songs that are quickly growing / viral as opposed to lower rates.
- We often / usually match the original sounds with the actual recording and display the sound recording that is used in the original sound.
- Trendpop does a better job of audio recognition than TikTok.
- One of our customers let us know that there was a discrepancy between TikTok’s data report and Trendpop’s data report (in our favor):
- TikTok: 5B+ views across 12 total sounds
- Trendpop: 10B+ views across 23 total sounds
Article is closed for comments.