Announcements

Help Wizard

Step 1

NEXT STEP

[Discover Weekly] Prevent Duplicates in Discover Weekly Playlists

For the past month or so, I've had so many repeats of songs on my Discover Weekly playlist.  Now, I can count on 5-10 songs that I've had before on the playlist. 

Is there a way to mix up the algorithm?  Or maybe an option to eleminate songs that have been on the playlist before?

Updated on 2017-08-25

Hey @earthquakepanda you'll be happy to know the team behind Discover Weekly recently rolled out a fix for seeing duplicates in your DW. We're marking this as implemented, but if you continue to see duplicates please let us know. Thanks!

Comments
bigkitty74

Please fix this asap. Very anoyning and seems like an easy fix. Still really like discover weekly but this issue is starting to ruin it.

speedfranklin

I don't love that I'm getting repeated songs, and would love, as @AoS1 suggested, to hear other songs by those artists that are being suggested more than once. I've been shown the same Shawn Mullins song twice, and it's not bad, but he has other songs that I would probably enjoy since I saved the first one I got from the playlist.

 

On a separate note, I would LOVE the ability to manually influence the genre of music or at least weight the songs that generate the list. I love classical music (my saved songs are at least 50% music from before 1900) but I've never been suggested a song in that vein.

 

Thanks for listening to our feedback!

jackedgar

This has recently become very bad for me. Discover weekly is one of my favorite parts of monday and spotify. The last two weeks have been almost identical to previous playlists. Strange because this used to never happen.

Please get it together, spotify.

FigaroPinto

It seems over the past months the Discover Weekly algorithm has changed. 

I now not only get repeated songs, week-on-week, but am getting about 30-50% of my DW playlist composed of artists I already know and like, and often from albums I've previously saved. 

 

Whereas previously most of the music recommended was from the past 5 years, there are now much older things creeping in. Nothing against that per se. But, for example, I've pretty wide and deep taste in electronic, indie, rock, etc. Putting a Beatles or a Depeche Mode track on Discover Weekly isn't going to cut it.

DJAdHoc

Couldn't agree with this more.  So frustrating that "Discover" Weekly plays SO many songs from my existing playlists and/or songs over and over from week to week.  Where's the "Discovery" in that?

b4liberty
endurefort

Really hope a 'remove from DW playlists' feature gets added in. There isn't a single DW playlist I get that doesn't have some duplicates of past DW playlists, and sometimes (like this week's) more than half of the playlist is stuff I already passed over previously. I'm starting to move back to Pandora for new song & artist ideas. 

MichaelTapp

^ What they said. 

 

Disover Weekly used to turn me on to so many good new songs. It feels like the recs are much more conservative now. 

harujion

Yes, please!

 

Don't make it based on what I've listened to. Make it based on what the world has to offer.

The new and the old from all over the globe!

 

NatalieRenee

I'm having the same problem! It's very frustrating, especially given the fact that these are songs that I also didn't like enough to save. This week, no joke, my DW only had six songs that were different from last week's DW. All the rest were the same. What's up?? I usually LOVE the DW and can really count on it for great new tunes.

 

I'm wondering: does this have anything to do with our tastes in genre/bands/artists? I've noticed a few trends on the threads I'm looking at, so I'm wondreing if it's a possibility that, after awhile, the algorithim might have a hard time generating new songs within a specific taste of genres and tunes, esp if it's a more niche taste. As in, it just doesn't know what else to suggest that would be in-line with what we've been listening to? Clearly I'm not a techy by any stretch, but maybe this is something a programmer could tackle?