Tiktok has released some insight into how the tool decides which videos to show in your feed. Although there is no real ground breaking insight it is worth a read to understand the fundamentals of how the system works.
TikTok has revealed the fundamentals of how it ranks the videos that appear in your feed. It does not seem like rocket science but it uses recommendation data.
- Personalisation. The tool will use your preference data to decide which videos you are more likely to enjoy. It uses data such as the preferences that you organically specific when you setup. They were called interest categories, it uses this data along with videos that you have voted that you liked as well as disliked to personally rank videos for each user.
- Video Quality – TikTok calls them User interactions and they are metrics such as likes, shares or comments.
- Length of Views – If a longer video is viewed from beginning to end this is a strong factor in the platform seeing this as a positive weighting for the video being used.
- There is some device optimisation but the platform says this is minimal. The platform mentions that it is concerned with “filter bubbles” which mean you just keep seeing exactly the same types of content due to over optimization of your preferences. TikTok has said that every so often it will throw a video into your feed that is outside your preferences just to keep your feed data fresh.
If you are interested in the TikTok ranking data they are opening up a transparency Center in Los Angeles where experts will be allowed to view the source data.