How TikTok recommends videos on the For You tab

How does TikTok recommend videos on the #ForYou (For You) tab? After the controversy that exposed the methods of the social network that avoided promoting videos of “ugly” and “poor” users, the platform went into more detail about how “TikTok For You” works. The feed on the home screen acts as a discovery area, is critical to the in-app experience and is also where most users spend their time.

TikTok ban

How For You works on TikTok

When you open TikTok, the first thing you see is the For You feed. There, you receive a stream of videos curated by TikTok based on your interests. That is, algorithm. This feed is powered by a recommendation system that delivers content to each user (including only what matters, excluding what does not). 

“Part of the magic of TikTok is that there is no single For You feed – different people can find some of the same featured videos, but each person’s feed is unique and personalized for the specific individual,” explains a post on the official blog.

ByteDance, owner of the app, says that TikTok’s mission is to inspire creativity and bring joy. But it was not always so.

The “For You” defines the experience in TikTok, and it is natural to question how the recommendations are chosen. In early 2020, leaked documents showed that TikTok censored faces and bodies considered ugly, people with disabilities and poor environments, such as slums, in its strategy to attract new users. In its defense, the social network said that such policies are no longer used.

In the eye of the hurricane, the social network decided to give more details on how everything works.

TikTok also said that it works to tackle problems that all referral services can deal with and promised to share tips on how you can personalize your TikTok discovery experience and have more control in the feed.

Recommendation systems

The recommendation systems are on Facebook, Twitter and also on TikTok. As the social network itself points out, from stores like Amazon to streaming like Netflix, recommendation systems help to create a personalized experience.

In general, these systems suggest content after taking into account user preferences, expressed through interactions with the application, such as posting a comment, cutting a link or following an account with a specific theme. These signals help the recommendation system to rate the content you like best.

TikTok For You

Therefore, For You should also consider your interactions on TikTok. However, those who started using the application – without an extensive history so that artificial intelligence could deliberate – noticed that there is a tendency for subjects and people there.

“The feed reflects the user’s unique preferences. The system recommends content by rating videos based on a combination of factors – starting with the interests you express as a new user and adjusting for things you’re not interested in as well – to form your personalized feed”, it says.

The recommendations are based on a few obvious factors:

  • User interactions, such as videos you liked or shared, accounts you follow, comments you post and content you create;
  • Video information, which may include details of subtitles, sounds and hashtags.
  • Device and account settings, such as language preference, country setting, and device type. 

And, also, in other more subjective factors:

These factors help the recommendation system, but they are given less weight compared to other data, considered “stronger”. Videos are rated to determine the likelihood of interest in the content and delivered to the exclusive feed.

“A strong indicator of interest, such as if a user finishes watching a longer video from beginning to end, would receive more weight than a weak indicator, as if the viewer and the creator of the video are in the same country,” explains platform text. 

How the TikTok algorithm works

Also according to TikTok, although a video is likely to receive more views if posted by an account with more followers, because that account has built a larger follower base, neither the follower count nor if the account had previous videos from high performance are a “direct factor” in the recommendation.

That is, if you’re watching and enjoying videos you don’t like, just to help an account’s statistics or because it’s a friend, this is shaping your For You experience. You should see more videos like these in some aspect. To change that, you need to be more selective and not get carried away, watching everything that appears in the home screen feed until the end, although it is almost impossible to stop.

Leave a Comment