On June 18, 2020, TikTok published a newsroom post titled “How TikTok recommends videos #ForYou,” giving an unusually direct official account of how its central feature works. The For You feed is the screen a user sees on opening the app: an endless, personalized stream of short videos chosen by a recommender system rather than by who the user follows.
TikTok said the system ranks videos using a combination of signals, grouped into three kinds. User interactions include the videos you like or share, accounts you follow, comments you post, and content you create. Video information includes details such as captions, sounds, and hashtags. Device and account settings include language preference, country setting, and device type. The post explained that a strong indicator of interest, such as whether a user watches a longer video all the way to the end, carries more weight than a weak one, such as the viewer and creator being in the same country. It also stated that neither follower count nor a creator’s history of high-performing videos is a direct factor - which is why a brand-new account’s video can go viral.
The For You feed is widely credited with making interest-based recommendation, rather than a social graph of friends, the organizing principle of a major platform. Its success pushed competitors including Instagram and YouTube to lean harder on algorithmic recommendation of content from accounts a user does not follow.
Why business readers should care: TikTok demonstrated that a recommender tuned on a few seconds of behavior can build a deeply engaging feed for a brand-new user almost immediately. That capability reset competitive expectations across the entire social-media industry.