From Spotify’s Daily Mixes playlists to its “Songs to Sing in the Shower!” playlists, each is curated to make your listening session as seamless and enjoyable as possible. The song selection is undoubtedly important, but this research focuses on taking that a step further and gauging how the placement and ordering of the songs in the curated playlist could impact whether or not listeners skip songs during their session.
Researchers found that acoustic coherence (similarity of genre/rhythm/key/etc. between neighboring songs) increased the number of completed tracks and decreased the number of skips in each sesison. These results were especially prevalent for Free users compared to Premium users. Given the limited number of skips Free users have, the optimization-based sequencing could increase consumption and make the listening experience more enjoyable.
Another point brought up in this research includes the importance of the familiarity of the first song in the playlist and the positioning of some songs that come later in the playlist.
The conclusions from this research pushed me to briefly analyze my personal habits when exploring a Spotify curated playlist. Here are some general observations I made about the way I approach it:
I’m less likely to give a playlist a try if too many of the songs are familiar or unfamiliar. If I’m exploring the Spotify curated playlists, I’m likely looking for exposure to new music but with some familiar music also there to keep me hooked. Basically, my brain recognizes a “Goldilocks” playlist with just the right amount of familiar and unfamiliar in the first few seconds of scrolling through the playlist.
Familiar artists are important, but familiar album covers is what really catches my eye. If I’m scrolling through a playlist and notice an album cover of a song I listen to a lot, I feel like I’m more inclined to give the playlist a shot just because of how much I like the other song(s) in that album. It kind of triggers a “damn, I should know all the songs in the album if I like that one song so much” feeling.
I feel kind of unsettled when I see two languages of the “same vibe”. For example, sometimes Spotify will curate me something like a “2000s Hype” playlist and put Kanye and Pritam in the same playlist. Sometimes that’ll pique my interest, but mostly it just throws me off.
In terms of songs positioning, I feel like it’s crucial to have that Goldilocks of familiarity specifically within the first ~20 songs of the playlist. I’m someone who almost always has the Shuffle feature on (literally all the time unless I’m listening to a new album that dropped). I believe that looking into shuffle/smart shuffle trends would be interesting while approaching this topic of sequential music listening. It’d be interesting to see whether it’s more effective to physically place two acoustically coherent songs next to each other in a curated playlist or to optimize the shuffle algorithm to achieve similar (I don’t know how shuffle works so not sure if that’s even possible or if the shuffle feature is truly completely random).
Ultimately, based on the research and my personal listening experience, I think that the placement of songs is extremely crucial when a listener has to make the decision of whether or not a specific playlist is worth listening to. In the first couple of seconds of scrolling, the playlist should be able to catch the listener’s eye, whether it’s because of a familiar album cover or just the right amount of new music to discover.
Interesting questions I’d explore next:
Impact of positioning of songs familiar/unfamiliar to the user
Analyze data regarding the shuffle and smart shuffle feature
This research was done under the assumption that the user is listening sequentially, but it’d be interesting to see the trends around how users’ habits vary around that.
How can shuffle/smart shuffle be optimized to provide greater acoustic coherence
