TikTok users will soon get a peek under the hood of their For You pages to understand why each video is being recommended.
In a blog post today, the company said it would begin rolling out a new feature on videos that breaks down why they’re being shown to the user. The feature — labeled “Why this video” — is accessed through the share panel and by clicking a question mark icon.
TikTok recommends content based on a variety of signals the company collects from users but has largely kept its exact methods under wraps. A blockbuster Wall Street Journal investigation last year detailed some ways the TikTok algorithm reads users, including how long a person lingers over videos even if they don’t interact with them.
Explanations in the new feature include a video trending in the user’s region, the video originating from an account being recommended, or the recommendation being based on what a user has watched, liked, shared, commented on, or searched for. The feature will be rolling out to users over the coming weeks.
TikTok’s highly personalized feed is part of what keeps users coming back but has also been a topic of criticism, particularly how users can be inundated with harmful content. A Wall Street Journal report last year detailed how young girls were sent down rabbit holes of dieting and weight loss content on the app, continuously serving up related videos. Shortly before the story was published, TikTok announced it was working to diversify recommendations to avoid repetitive content.
The company has gradually added more ways for users to fine-tune their feeds to their liking, at least by telling the app what they don’t want to see. In July, TikTok introduced the ability to filter out certain videos by keywords or hashtags. Users can also “dislike” videos by long pressing on a post and clicking “not interested.”
TikTok’s algorithm is the company’s secret sauce, the thing that keeps users entertained, intrigued, and at times a little creeped out at content’s relevance. Though the new explanation feature likely won’t be as precise as to say “we showed you this video because you watched another related one 10 times yesterday,” it might help users better understand why the For You page looks the way it does.