Skip to main content

TikTok is adding new ways to fine-tune your For You Page algorithm

TikTok is adding new ways to fine-tune your For You Page algorithm

/

A new slider tool allows users to signal they want more content based on certain topics, like ‘dance’ or ‘current affairs.’

Share this story

Vector art of the TikTok logo.
Image: The Verge

TikTok is giving users a way to more precisely shape what kind of content they see on their feeds — or at least signal to the algorithm what they’re interested in.

Under Content preferences > Manage topics (which can be found in “Settings and Privacy”), users can adjust sliders to signal they want more or less of certain topics in their For You Page. Topics include “creative arts,” “current affairs,” “humor” and more. Sliders are set to a default middle ground at first, and users can tweak it from there.

TikTok’s hyper-targeted For You Page is famously a bit of a black box, built on a trove of user data that sometimes makes it feel like the algorithm is reading your mind. This new slider feature gives us a peak behind the curtain, especially regarding how TikTok categorizes content. It’s interesting, for example, that “dance” is its own distinct topic separate from other arts — not totally surprising given TikTok’s early role as the kids dance app.

“Manage topics” settings that include sliders for categories like “humor” and “Fashion & beauty.”
Image: TikTok

Over the years, TikTok has rolled out different features to give users a sense of control over their own algorithm. The “not interested” button on videos is meant to signal to the system that you want to see less of something, and users can also block certain keywords or hashtags from their feeds entirely. Last year, TikTok also began testing a way to reset your For You Page to start fresh.

It’s not yet clear how effective adjusting the sliders will be — historically this kind of tool isn’t necessarily a hard and fast rule for what kind of content recommendation systems return. Prior research into YouTube’s “dislike” button, for example, has found that the algorithm kept recommending similar content regardless.