YouTube gives users a “don't recommend channel” option and a way to pick topics of interest for videos showing in Up Next or homepage, now on iOS and Android
No matter how advanced algorithms get, none will understand your YouTube viewing preferences better than you.
Context & Ripple Effects
YouTube's mobile homepage has been machine-learning-driven since its 2016 redesign built recommendations around larger thumbnails, and the platform has since layered on more algorithmic surfaces — Explore, and later a personalized For You section for creator channels. The 2019 controls are the first direct counterweight in that arc: instead of tuning the algorithm from behind the glass, users get explicit levers to remove a channel or steer topics in Up Next and on the homepage.
The move foreshadows the direction the product has taken since — suppressing recommendations entirely when watch history is off, limiting repeated body-image and fitness recommendations for teens, and filterable Subscriptions tabs that hand sorting to the viewer rather than the feed.
First-order effects
- Users on iOS and Android can immediately prune unwanted channels from their recommendations and declare topic preferences, directly overriding what the Up Next and homepage algorithms would otherwise serve.
- Creators whose videos get blocked via 'don't recommend channel' lose a recommendation-driven impression stream, and the signal feeds back into YouTube's ranking for every other viewer.
Second-order effects
- Channels that rely on broad algorithmic distribution face pressure to keep recommendations on-target, since a spike in 'don't recommend' signals now carries a direct, user-initiated cost.
- Explicit topic picks give YouTube cleaner preference data than watch-time inference alone, tightening the feedback loop that determines which channels the homepage promotes.
Third-order effects
- The pattern across these releases — user controls in 2019, history-off suppression and teen recommendation limits later — points toward recommendation feeds becoming user-adjustable and regulator-sensitive rather than purely algorithmic, with platforms accountable for what the feed surfaces.
- If explicit user signals keep gaining weight, curation power on video platforms shifts incrementally from the ranking model to the viewer, changing how channels compete for distribution.
The trend: Video recommendation systems are evolving from opaque algorithmic feeds into user-tunable surfaces, as platforms trade some ranking authority for user control and content-safety constraints.