Clearing Like/Dislike Cache and General Radio Improvements

First of all, I'm not here to bash Spotify's radio. In fact I am always recommending Spotify to people. I have found some of my new favorite artists because of the Spotify radio.

That being said, there are some things that could be better. Generally, the songs and artists that come across on my playlist radio stations are very similar to the songs in the playlist. Problems seem to occur when the genre is not popular, new, or a specific sub-genre of music that is not widely listened to. Some specific examples: trap, drum and bass, IDM, remixes of any kind, and decade playlists.

A few examples:

If I make a playlist of nothing but alternative music from the 90's, or new-wave from 80's, Spotify doesn't seem to realize that everything in the playlist is from a specific decade. To me, this seems like simple logic that could be added to the playlist.

When you talk about the sub-genres however, understandably, things get a little more complicated. My guess is that the people that tag genres in songs (or however songs are calculated to be "similar") are not familiar the unique attributes of that genre. The cheapest option to fix this for Spotify would be to let the community help you. Obviously there are inherent problem with letting people tag genres for everyone, but there are things that could be done to mediate those problems. My recommendation is to provide some type of gamification to Spotify for tagging songs. It's been proven time and time again that people like earning points, showing up on leaderboards, and getting other incentives for being contributors that provide accurate information. I have many ideas on how to achieve better tagging results using gamification; if you would like to hear them let me know (I'm trying to keep this as short as possible).

The 2nd option would be for Spotify to hire musicians and other sorts of professionals from the industry to help tag music. Obviously this costs money in the sense that Spotify is paying professionals vs letting the community sort things out.

My last gripe is around like/dislike in the radio stations. There are many instances of my playlist stations that are essentially unlistenable now, because I have liked one song that was either a remix, or had some other similarity to an artist or song in a different genre that is way more popular than the music in my playlist. Once you like that one song, you are pretty much screwed, because it seems like Spotify is giving priority to that song since it knows more music like the song you liked than the songs in the playlist. This is again a result of the poor tagging for these sub-genres, but there should be a way to undo the damage you caused by clicking "like".

Giving users the ability to clear the like/dislike cache for a particular station would be super helpful. At least then you don't have to rebuild the station to start over.

Also, your forum spellchecker thinks I'm misspelling the words "Spotify", "unlistenable", and "proven". Just sayin'.

Hi and thanks for your suggestions! A similar idea has  been suggested here regarding changing 'likes' and 'dislikes':
https://community.spotify.com/t5/Spotify-Ideas/Modifiable-Thumbs-up-down/idi-p/212314

Regarding 'tagging' music, please see this idea: https://community.spotify.com/t5/Spotify-Ideas/Tag-Music/idi-p/57185

Additionally, earlier this year Spotify acquired The Echo Nest, the world's leading company for music information retrieval, in part to improve the aspects of identifying genre and so forth. 

 

Status: Duplicate
0 Likes
Comments
Rock Star 21
Rock Star 21
Status changed to: Duplicate

Hi and thanks for your suggestions! A similar idea has  been suggested here regarding changing 'likes' and 'dislikes':
https://community.spotify.com/t5/Spotify-Ideas/Modifiable-Thumbs-up-down/idi-p/212314

Regarding 'tagging' music, please see this idea: https://community.spotify.com/t5/Spotify-Ideas/Tag-Music/idi-p/57185

Additionally, earlier this year Spotify acquired The Echo Nest, the world's leading company for music information retrieval, in part to improve the aspects of identifying genre and so forth.