Smart Shuffle can really be a gem with a little bit of execution tweaks. I am an avid Spotify user of 10 years and my favorite thing about the app had always been its ability to bring me music and artists that I’ve grown to love - it used to prioritize genuine, personalized musical discovery over trends and popular streaming and thats what made Spotify stand out. Smart Shuffle is not reflective if that.
However, I think that the larger issue a lot of users are having with Smart Shuffle feature is that unlike a lot of the platform’s other features that user’s love like Discover Weekly, Playlist Suggestions, etc - this one seems to be entirely impersonal. All of my Smart Shuffle recommendations have been “TikTok” or social media popular songs, which is not my listening style at all. And regardless of how many times I vote negatively for these recommendations (which I ensure to do in hopes to alter my recommendations) there is no change to my suggestions and it has been this way since the feature’s release.
Smart shuffle algos should be more specific to user listening styles, genres, most listened to artists, and more responsive to the installed voting system to ensure a more enjoyable experience and match the reputation Spotify has created. I think that the engineering team needs more support or to restructure the ways that they are creating algo responsiveness based on user input and past data as it respects to people’s Smart Shuffle to have a better response to the feature.
Don’t mean to sound like a know it all, just love the product and hope to see improvements on the feature, there is a strong opportunity for something great!