A “Hit” for Every Mood —

Spotify’s analysis of our emotional states

190805

Spotify’s business model is about finding not just the music (and now podcasts) we like, but also music to match how we feel. There’s nothing wrong with having a Spotify playlist for every mood, but we should be concerned about relying on an algorithmically-generated “hit” for every moment in our lives.

 

In the past few years, streaming services have convinced us via aggressive marketing that listening to playlists that match our moods and activities is a service to listeners and artists alike: they help users navigate and choose from some 40 million songs and curate our mood. In the process, it’s likely that listeners find new songs and artists they might not have otherwise found, elevating the platform into an “engine of discovery.”

By the numbers: Spotify the new beast

The decision to define audiences by moods and activities was part of Spotify’s advertising strategy leading up to its IPO in 2018.

 

  • 2014: Spotify acquires Echo Nest, a music intelligence firm

 

  • 1.5 billion user-generated playlists at Spotify’s disposal

 

  • 400,000 user-generated barbecue playlists. This becomes the company’s big epiphany

 

  • 2015: The year Spotify starts selling advertisers on the idea of marketing to moods, moments and activities instead of genres. As of May 1st of that year, they could target ads to free users

 

  • 2016: Spotify shares mood data with their partners, including some of the world’s biggest advertising firms.

 

 

  • 207 million: The total number of people using Spotify across 97 different countries. 96 million are direct subscribers with the remaining 111 million relying on the ad-supported version.

 

  • Billions: the number of daily data points Spotfy claims, in their advertising deck, to receive across devices.

From data to dollars

As part of its aforementioned study with Ypulse, Spotify gathered 600 in-depth “day in a life” interviews recorded as “behavioral diaries.” They then paired this data on their streaming habits with data about their interests, lifestyle and shopping behaviors from third-parties to create a comprehensive profile of them.

 

In 2016, Spotify signed a multi-year partnership with WPP, the world’s largest holding company for advertising and PR agencies, giving the conglomerate unprecedented access to its mood data, again part of its plan to ramp up its advertising business ahead of the IPO.

 

The reason why this mood data is so valuable? Millennials absolutely destroy the traditional segmentations of the market: their aversion to being labeled means they don’t all fit predictably into certain categories that can be targeted.

Angel/Devil’s Advocate: the soundtrack to your life

Spotify’s commitment to mining and monetizing our emotional states should come as no surprise: it’s said before that it has ambitions to be one of the biggest platforms alongside Google and Facebook.

 

Where many platforms that use algorithms are good at serving us content based on our interests (or what’s corroborated as such based on our browsing history), Spotify’s trying to match marketing moments to the immediacy of dynamic emotional states. We might not be shopping for clothing online every day , but there’s a good chance we listen to music as we get ready in the morning, while we’re commuting, working and eventually, when we get home and relax.

 

On one hand, we can argue that Spotify might just be doing a good job of being our personal content concierge, offering up music and podcasts that we’re sure to like at the times we want them. These in turn help us stay intellectually and creatively stimulated, or at least they give us the emotional nudge to get over the speed bumps of a stressful day. Music, or more broadly, listening to something, has always been this sweet immaterial drug to us, whether it was when listening to CDs, cassettes, radio or vinyl.

 

Few people still make their own vinyl, that’s for sure, but when processes for burning our own CDs and before that, cassette mixtapes became widespread — curating our own high, if you will— it made a purer product where we no longer had to switch albums and skip the tracks we didn’t like; it was all in one spot. We’d then share these playlists with peers who could appreciate the thought put into them.

 

We can still absolutely make our own playlists today, but perhaps we’re more wary about sharing them (after all, no one wants to be caught sharing a ‘wack’ personal playlist at a gathering, or others overusing songs that are closer to gems, unknown to a wider audience). The issue arises if we eliminate ourselves from the curation process entirely and outsource it to algorithms that are not just vulnerable to the biases of those that create them, but also to the interests of influential companies and their partners.

 

It becomes less about finding and listening to content that addresses how we really feel and what we need, but more about what the data suggests.