Discussion
The biggest difference between #SpotifyWrapped and other music services’ end of year recaps is that Spotify compares my listening activity to its other users.
It’s computationally easy to pull a single user’s listening history. That data processing and infographic generation could be done entirely on device.
However, comparing 700+ million users’ listening histories is a takeover-entire-datacenters scale processing job. So Spotify has to do a snapshot of data at some point.
[Disclosure: I worked at Spotify 8 years ago. My husband works there now. I do not have inside knowledge on how Wrapped is produced.]
I suspect Spotify does not include my listening history for the time between when it pulls data for the massive computation required for Wrapped (early November?) and the start of a new year.
Albums and songs I loved in November and December 2024 didn’t rank highly in my 2025 Wrapped. Xmas holiday music on repeat in December also didn’t rank.
The biggest difference between #SpotifyWrapped and other music services’ end of year recaps is that Spotify compares my listening activity to its other users.
It’s computationally easy to pull a single user’s listening history. That data processing and infographic generation could be done entirely on device.
However, comparing 700+ million users’ listening histories is a takeover-entire-datacenters scale processing job. So Spotify has to do a snapshot of data at some point.
A interesting experiment would be to create a dedicated Spotify account with a playlist of 31 songs and play only 1 song a day—the song position in the playlist corresponding to the day of the month.
The top song played would indicate the last day of the month in the Wrapped data snapshot and the play count would indicate the month the snapshot was taken.
That experiment could continue into the new year and use a different 31-song playlist for the new year in order to test my hypothesis that a listening history data gap exists and therefore albums released in the fall have less chance of making it into Wrapped, assuming most listening to a new release is done closest to release.
An algorithmic transparency feature I want in #SpotifyWrapped:
Rankings of songs I intentionally listened to vs songs Spotify’s algotorial playlists chose.
I didn’t recognize many of my top songs in 2024 Wrapped. They weren’t songs I intentionally listened to. They were just on Spotify “made for you” playlists that I didn’t skip, but also didn’t like enough to add to my library or a playlist.
Only 9 of my top 100 songs in 2025 were not in my library, but I intentionally listened to them all.
A space for Bonfire maintainers and contributors to communicate
