For all of its talk about prioritizing “discovery” and “knowing your tastes” ("I just want to meet a girl who knows and loves me like my Spotify Discover Weekly playlist," read a Spotify ad last year), what Spotify feeds to Browse and pushes to Discover is influenced largely by whether an artist already has a massive marketing campaign and corporate push behind them...

Playlist culture is introducing an unprecedented dependence on data. We hear about the stacked human playlisting teams, with “genre leads” and “junior and senior curators” building thousands and thousands of playlists. (Though we never see their faces or names on the platforms—Spotify’s way of building trust in the mystified Oz-like “magic” of Spotify, rather than human intelligence needed to program playlists.)

These human curators are responding to data to such an extent that they’re practically just facilitating the machine process.

As BuzzFeed reported last year, Spotify uses a performance tracking application titled PUMA, or Playlist Usage Monitoring and Analysis, which “breaks down each song on a playlist by things like the number of plays, number of skips, and number of saves.” PUMA also tracks “the overall performance of the playlist as a whole, with colourful charts and graphs illustrating listeners’ age range, gender, geographical region, time of day, subscription tier, and more.” In the curated playlist factories, human beings essentially reproduce the work of the algorithm....

— Liz Pelly, Cashmusic.org