Spotify Algorithm
The Spotify algorithm is a layered recommendation system combining collaborative filtering, natural language processing of track metadata, and audio-feature analysis. It powers Discover Weekly, Release Radar, the autoplay queue, personalised search, and the 'similar artists' surface.
Reviewed by FollowerStar Editorial · Updated April 2026 · Part of the Social Growth Glossary
The single highest-weighted signal across every Spotify algorithm surface is save rate - the percentage of unique listeners who save a track to their library after hearing it. Tracks with save rates above 8% routinely break out to algorithmic playlists without any editorial help.
The second-strongest signal is full play-through rate (≥30s plays / total plays). Tracks with weak hooks fail this threshold and get throttled from algorithmic recommendation - even with high stream volume.
Spotify's recommendation pipeline tests new releases on small listener pools first (similar to TikTok's FYP tiers) and progressively widens distribution as engagement signals clear thresholds. Early signals in the first 48 hours weigh far more than late ones.