Why You Must Read ‘Spotify Teardown’
Hello, hello, short intro this week. I forgot to properly update the Patreon tiers before I sent out last week’s newsletter, but that’s fixed now. My bad! This will be my last reminder that the Penny Fractions survey is still up for those who haven’t responded. Beyond that, let’s dive into this written word text.
I love transparency, so I’ll offer a bit of my own for this newsletter. Back in 2017, when I was blogging for the now-defunct website TrackRecord, I worked on a story about the Spotify Teardown. I came across the story of Spotify trying to shut down this academic book and couldn’t help but write a blog post about it after exchanging a couple of emails with one of the authors. Fast forward to 2019 and while reading the book, I cracked a smile when a reader pointed out that a different TrackRecord article I wrote was cited in it. I say this because I loved the book and I highly recommend it, but I wanted to make sure to disclose my bias upfront.
Spotify's Investment History
Spotify Teardown, written by researchers Maria Eriksson, Rasmus Fleischer, Anna Johansson, Pelle Snickars, and Patrick Vonderau spooked Spotify. I’ll save the book’s publication history for the close but within the first full chapter of reading Spotify Teardown, it’s easy to understand why Spotify tried so hard to hold the production of this book. The opening chapter (“Where Is Spotify”) peels back every press release, outside investment news, and marketing moment to re-chronicle Spotify’s rise without a public relations sheen. Spotify is portrayed not as a company that was interested in saving the music industry but as one that was created by a couple of bored advertising bros. By not immediately accepting Spotify’s narrative of reviving the music industry, the narrators illustrate Spotify’s initial product as simply ‘piracy with a nicer face’. This is corroborated when the concept of subscriptions is introduced in the late 2000s by Spotify’s earliest investors—major labels:
The idea of selling subscriptions was rather forced upon Spotify by the holders of music licenses, and the economic crisis exerted an even stronger pressure in that direction. It was expedient at this time, however, for Spotify to modify its own history—and to distance itself from the pirate heritage. In the same blog post, Ek also promised that Spotify would not only offer music streaming but would also sell individual downloads on a pay-per-song basis.
Ek’s indifference towards subscriptions is understandable given his advertising background. There can be a needless battle over how to describe the company but I’ll just restate this: Spotify isn’t a music company, Spotify isn’t a tech company, Spotify is an advertising company. Chapter 4 (“What Is the Value of Free”) devotes a number of pages towards exploring the ecosystem within Spotify in which advertisements are bought and sold. The meticulous record of investments into Spotify, while contrasted with the lack of direct artist statements, helps illuminate where Spotify puts its real energy.
Cracking into the Black Box
Outside of the meta-media critique, the authors attempt their own deep dives into how Spotify’s actual platform works. Before talking about their findings, the authors present an internal monologue explaining how even their own research is impacted by Spotify’s marketing, which I found quite unexpected. Chapter 3 (“How Does Spotify Package Music”) shows the researchers testing exactly how personalization works on Spotify’s platform. The researchers drill down deep into Spotify’s radio feature and eventually figure out its looping pattern, which corresponds to online listener complaints about why the recommendations are so poor. Still, as stated in the book, “Such assumptions reveal a normative claim that the radio algorithm should produce apt recommendations.” In other words, why should consumers believe that any company could perfect radio curation? The question I’d now like to raise is: If Spotify didn’t tout its own ability to personalize content, would there be a user outcry against it or even researchers interrogating it?
The researchers arrived at a similar conclusion about Spotify’s Discover tab:
In fact, our interrogation of parts of the Genres & Moods and Discover features showed that the apparent personalization of music delivery, brought about by a multitude of actors and algorithmic processes and materialized on the surface by Spotify’s interface, was not very personal at all.
This conclusion doesn’t feel too surprising given that public relations copy drives most stories about Spotify’s personal recommendation tools. This isn’t to say that Spotify’s app isn’t recommending enjoyable music, but the idea that a single company has managed to perfect music recommendations at a global scale is fairly farcical. Instead, all of this personalization is just helping to form better user profiles to target advertising. The fact that certain users are connecting with music is simply a pleasant side effect. Spotify Teardown highlighted a series of annoyed online posts expressing dissatisfaction with Spotify's music recommendations, which helped remove the company's glossy veneer of personalized music discovery.
All About that Sweet Sweet Coca-Cola
I first read about Spotify Teardown from a TorrentFreak story back in 2017, which centered on one of the book’s claims about how Spotify was built on pirated music. Prior to getting proper licensing deals, Spotify simply pulled music available on the Pirate Bay, which understandably isn’t the type of headline a company wants to be thrown around before going public. Spotify contacted the Swedish Research Council (who was funding the research in an effort to shutter the book), which speaks to the company’s concerns with its own self-deluded image. Even still, I don’t think that was the most damaging part of the book.
In my opinion, the most cutting part of Spotify Teardown wasn’t all the original research that went into the book or the claim that Spotify’s business was built on pirated music. Rather, what intrigued me the most was a small section about the company’s early days of investor funding. In 2012, while the company was still in the preliminary stages of entering the United States market, Spotify received a round of funding led by Goldman Sachs with an assist from Coca-Cola (I said it last week, but it obviously makes a ton of sense that Coca-Cola was the first brand to partner with TikTok).
The reason this small bit of news, which was widely reported on, stuck out to me so much while reading the book was because it showed the company’s true interests. Without investment from major music labels, a company like Spotify simply wouldn’t exist, so I’ve always found tech writers that credit Spotify for “saving” the music industry to be stupidly misguided. There is no Spotify without the record industry and its willingness to see this product play itself out. That’s why the Coca-Cola investment shows that Spotify is a company that exists purely to sell advertisements, not to create sustainable careers for artists. Much like TikTok, these platforms are just advertising companies. Spotify Teardown breaks down some companies involved in programmatic advertising:
Supply-Side Platforms (SSPs): Adscale, PubMatic, Rubicon, and YieldLab
Demand-Side Platforms (DSPs): AdRiver and Sociomantic
Ad Exchanges: Admeta, AppNexus, Facebook Exchange (FBX), OpenX, and Yahoo Ad Exchange
Ad Networks: Adkonteskst
Ad Servers: Adtech
Data Suppliers: Seed Scientific and the Echo Nest
Measurement: Moat and Google
Verification: ComScore and Nielsen
The authors compare this system to the stock market and argue it’s intentionally created to be incomprehensible to a normal person. If artists don’t see where they’d fit into this equation, that’s the point. They don’t. This is all information being traded based on user data being collected while playing music on the platform. If one wonders why Spotify is so excited to dive into podcasts, one reason is that the company was never truly invested in music beyond sheer marketing purposes. Spotify Teardown shows a company much like all advertising-support music platforms, where the ultimate customer is Coca-Cola; not music fans nor artists.
6 Links 2 Read
4 songs on “When I Get Home” may earn Solange no money from Spotify - Quartz
The headline and overall framing of this piece capture a big issue I hold with how business and tech media frame stories about music streaming. Instead of investigating why an artist isn’t making money for songs under 30 seconds, Quartz places the burden of responsibility on the artists to figure out how to work within this narrow frame. Journalists often struggle to frame collective industry issues and, instead, choose to push these issues onto individual artists. The issue isn’t Solange’s artistic choices but rather a system that harms such choices.
The Downside of Tech IPOs - Wendy Liu
Wendy Liu, a former startup founder and upcoming author, is writing a blog post every day this year with content that is at the intersection of technology and left politics. I’ve loved a lot of the posts but this one, where she breaks down the problems with IPOs, felt even more relevant after reading Spotify Teardown.
Pandora’s new Stories feature lets artists create playlists interspersed with voice tracks - Music Business Worldwide
*Extreme David Turner Voice* This is just a press release but artists being able to add in little voice tracks to playlists is neat. Certainly not new, but still I appreciate the marketing put behind something so simple.
As Warner and Spotify Face Off Over India, What Comes Next? - Billboard
I wrote my thoughts on this topic last week on Patreon, but this situation reminds me of Spotify’s early usage of pirated music to launch the company. The attempt to skirt Indian law in how they licensed the music of Warner Music Group isn’t bold or interesting, it’s just an unethical. Spotify claims to have had “one million users” within a week, but didn’t clarify if these were paying customers or even how much users were engaging with the platform.
De La Soul’s Digital Album Releases Postponed by Tommy Boy Music (EXCLUSIVE) - Variety
I’d argue this isn’t really a streaming story, as De La Soul’s interviews repeatedly stressed that the group never made much money from record sales. Still, the group is now drawing a line to finally demand more from Tommy Boy. Good for them.
TikTok has been illegally collecting children’s data - Vox / Who Are Online, Recruited by Advertisers and 4 Years Old? Kidfluencers - The New York Times
I now wish that in last week’s newsletter I had been bolder in saying how gross TikTok’s business model clearly is, but the fact that the company received a $5.7 million Federal Trade Commission fine reaffirms that it is just a child exploitation app. Then, on the other hand, is a New York Times piece offering a slightly more human side to exploited child labor. Lovely stuff all around.
The Penny Fractions newsletter arrives every Wednesday morning (EST). If you’d like to support it, check out this Patreon page. The artwork is by graphic designer Kurt Woerpel whose work can be found here. The newsletter is copy edited by Mariana Carvalho. My personal website is davidturner.work. Any comments or concerns can be sent to firstname.lastname@example.org.