How TikTok Chooses Its Recommendations #ForYou
TikTok owes much of its massive success to its expertly designed, AI-based algorithm. Here, we analyze some new information from TikTok itself on how exactly it comes with videos that appear in its #ForYou feed.
Guest post by Bobby Owsinski of Music 3.0
TikTok is probably the hottest social service in the world right now and much of its success is because of its excellent AI-based algorithm. While we’ve discussed how this algorithm works in the past, TikTok itself has provided information about how it comes up with videos for your #ForYou feed.
The Recommendation System
Just about every social network uses a recommendation system to serve up content that it thinks you’ll enjoy. In the words of TikTok:
“In general, these systems suggest content after taking into account user preferences as expressed through interactions with the app, like posting a comment or following an account. These signals help the recommendation system gauge the content you like as well as the content you’d prefer to skip.”
The TikTok Feed
As you’ve probably experienced, some social services do this better than others. So what makes the TikTok recommendation system that supplies #ForYou work so well? Once again, from TikTok itself:
“On TikTok, the For You feed reflects preferences unique to each user. The system recommends content by ranking videos based on a combination of factors – starting from interests you express as a new user and adjusting for things you indicate you’re not interested in, too – to form your personalized For You feed.
Recommendations are based on a number of factors, including things like:
• User interactions such as the videos you like or share, accounts you follow, comments you post, and content you create.
• Video information, which might include details like captions, sounds, and hashtags.
• Device and account settings like your language preference, country setting, and device type. These factors are included to make sure the system is optimized for performance, but they receive lower weight in the recommendation system relative to other data points we measure since users don’t actively express these as preferences.”
The Filter Bubble
No matter how smart the algorithm is, it can’t overcome complacency. That means that if you consume the same time of content all the time, you’ll never see anything that might stretch your interests. This places you in what’s known as a “filter bubble.”
TikTok claims that you can overcome the bubble by just using the algorithm more. For instance, it will automatically make sure that you don’t see back-to-back videos with the same song, or by the same creator, or duplicated content. It also recommends content that is directly outside of your “sweet spot” as a way of diversifying your #ForYou feed just enough so you get a taste of something new.
All this means that TikTok’s algorithm is truly its “secret sauce” and clearly the one to beat for any up and coming services with aspirations of replacing it as the next big thing.