Here is the obvious, one more time. I’ve seen other data indicating much the same thing this does. A couple years back, I read an account of an experiment Spotify conducted. Some of those “recommended for you” daily playlists were put into the hands of human curators (known as ‘music directors’ back in the day.) They studied the data and found the human-tooled playlists had much longer TSL; time-spent-listening.
Here’s another affirmation of what’s always been obvious to me. Algorithms make lousy music mixes. This data shows it in a different way. Listeners to human-curated playlists have a much, much lower skip-rate than the algorithm’d playlists of Spotify and its cohort.
