Experiencing mood on Spotify
James Lynden


I have been saying for a while that emotion is the most valuable data metric – especially so for the music industry as it is an emotive medium.

Spotify’s ‘Discover Weekly’ playlists are a great step forward in the recommendation space but there is still a huge untapped opportunity to be had around emotion. At the moment, recommendation algorithms are over reliant on logical data. “OK, you listen to Nirvana’s loudest songs, check out Mudhoney…”

Here’s the thing though – liking one song doesn’t necessarily mean that I’ll like another song, even if I logically should. Our rationale for liking songs is unpredictable. How many times have you heard someone say “Y’know, I *shouldn’t* like this but I really do!” in your lifetime?

Recommendation, audience segmentation, KPIs and analytics are all built around WHAT rather than WHY – but when you think about it, WHY audiences engage is much more interesting and valuable to understand than engagement alone.

Facebook Reactions are a great step forward. In a few years time, it will allow brands to reach certain types of audiences based on their emotional state and tailor their messaging and content creative to suit – as opposed to just blasting out one message to 18–24 year olds in X who like Y and Z.

For the music industry, this is even more important. For example, I might listen to happy songs when I’m sad or sad songs when I’m happy. I may love a certain song because I have a great memory associated with – but if the person associated with it suddenly breaks my heart, tells me that they love Marmite (other food spreads are available) or dresses up as a clown (my biggest fear), I may never want to listen to that song again.

Likes, shares, clicks and plays are ‘important’ metrics for a marketing report but they don’t make music better. If anything, chasing spikes only widens the disconnection people have with music and devalues it (further).

Music is (mostly) made and experienced emotively, yet it is (mostly) marketed (and judged) on logic. If data does not tell the full story, do not let data drive the story. Never underestimate or devalue the importance of emotion, instinct and context.

In short, there is much more that can and should be done in the emotive data space.