Spotify’s Black Box is Darker than I Thought
The explosion of the world wide web and the growth of new digital platforms opened up a sea of data to be discovered. As researchers started to dive in, many were met with strong waves of resistance from big media giants. Companies like Facebook are extremely secretive about their algorithms and the inner workings of their company. This resistance left researchers to fend for themselves to find data. Luckily, digital tools like “scrapers” exist to help retrieve data. Other companies are more forthcoming with their data and collaborate with researchers willingly.
A group of researchers from Umea and Stockholm University found much resistance and secrecy in their attempt to study Spotify, the music streaming giant (though this book revealed that music is only a small component of what Spotify truly is). Initial interviews revealed that Spotify was less than forthcoming when it came to sharing data and information on their algorithms.
The researchers took this challenge head-on and went beyond traditional research approaches. Through mixed-methods, ethnography, and intervention methods, the researchers were able to “crack open the black box” of Spotify. It also helped that they had a million-dollar budget ¯\_(ツ)_/¯
What the researchers found when they lifted the lid was quite alarmingly:
Spotify’s business model was highly skewed toward major artists and record labels, screwing the “little guy” over
Spotify’s sketchy beginnings (beta-versions of the app streaming music without a license) and their relationship to companies like The Pirate Bay
Several privacy issues (like collecting sensor data on whether you are walking or running and the speed of your movements & selling personal data to advertisers)
Researcher’s radio loop experiments to find out why people rarely liked their recommended songs discovered that almost all song loops looked the same regardless of thumbs up or down on songs
Spotify dumping large amounts of “data junk” on people’s SSDs, slowing down storage capacities
Though mood playlists cover pictures were mainly of women (women smiling or exercising), the majority of recommended artists were males – “maintaining male privilege in the music industry”
“Menu driven identifications” that were not inclusive (they didn’t add a nonbinary option to gender till 2016, and it’s only available in a few countries)
“Too much data” intervention revealed very little personalization occurring though they have more than enough data to make it happen
“Algorithmic system bias”
Secretive on the trading of personal data to advertisers
Though the researchers revealed many of Spotify’s secretive and obscure practices, there are some areas where I wish they’d dug a little deeper. Much of the book focuses on the economic side of Spotify. There were a few references and allusions to the low percentage of money that artists make from Spotify. They mentioned that Spotify’s business model favors major artists and mainly the record labels, but I wish they had elaborated on how much in royalties do artists actually make. In one of their interventions, the researchers released an album to infiltrate Spotify’s inner workings. The researchers said that they didn’t cash any of the royalties made. However, they didn’t disclose how much money they made. I wish they’d released this amount, as it would have been telling on how much money artists make off royalties.
I feel like artists, especially smaller ones, are some of the most negatively impacted people from Spotify’s platform. This is an aspect I wish they would have explored more. Though they mentioned how Taylor Swift pulled her music from Spotify because she felt that it de-valued it, I would’ve like to have seen interviews with smaller artists to see how they feel about Spotify and its economic impact.
In conclusion, I knew I was always a Tidal > Spotify person for a reason! (I felt especially proud when the researchers mentioned that Tidal willingly worked with a team of academic researchers and gave them data for their work)
Dr. Porter told us to look at our readings from a dystopian vs utopian perspective. As our weekly blogs come to an end, I have decided that my view of emerging media platforms is mainly dystopian (though I hope in a few years I read some newer blogs from future students that change my mind).












