In 2014 an indie band named Vulfpeck put out an album that was a slight departure from its usual instrumental funk style: It was entirely silent. Called Sleepify, it featured 10 tracks, all roughly 30 seconds long, all filled with cold, dead nothingness.
The band wasn’t veering into experimental music. Rather, the album was a brazen attempt to game Spotify’s royalty system. The band asked fans to play the silent album on loop while they slept. Every time someone listened to one of its songs, the group would earn a fraction of a penny, about $0.0030 to $0.0038. Almost immediately, the album became a—heh—sleeper hit, snagging an impressive 5.5 million listens. It netted the band about $20,000 before Spotify removed the album from the platform for violating the company’s rules, or terms of service.
Spotify is worth about $25 billion, with more than 100 million paid subscribers (out of 217 million monthly active users), and has fundamentally altered how the world listens to music. And yet apparently the best way for a small indie band to make money off it was with a viral stunt.
For the few who don’t use it, Spotify is a subscription-based music-streaming app. It’s also an ad-tech data broker, selling information on its users. Plus it has all the trappings—a newsfeed, a “community”—of a social network. And its music recommendation system functions like an Internet radio station. With all that in mind, it’s easy to get confused: What even is Spotify?
Spotify Teardown: Inside the Black Box of Streaming Music, a new book written by a group of professors and researchers at Stockholm and Umeå universities (Maria Eriksson, Rasmus Fleischer, Anna Johansson, Pelle Snickars, and Patrick Vonderau), attempts to answer this question. It’s trickier than you’d think.
The book begins by tracing the history of the company, using as a timeline Spotify’s many funding rounds. Typically, start-ups receive early funding from investors in exchange for equity in the company. It’s a fascinating look at the financial realities of running a tech company and the many ways it contorts its mission in its quest for scale.
Spotify was founded in Stockholm in 2006 by two multimillionaires, Daniel Ek and Martin Lorentzon, who met after cashing out at the ad-tech companies they founded, Advertigo and TradeDoubler, respectively. The idea was, as Ek explained in 2009, to legally give users “access to all music in the world, for free.” Like Facebook and Google, it would make its money primarily from advertising.
Somewhat counter to Ek’s framing, Spotify began as peer-to-peer network software for sharing not just music but all kinds of files, including videos and images. During beta testing, many of the files on Spotify’s servers were originally downloaded from the Pirate Bay, the torrent search engine popularly used for illegal file sharing, which means “Spotify began as a de facto pirate service.”
Popular
"swipe left below to view more authors"Swipe →
As the company moved beyond its initial invite-only beta phase, it needed to pay for music licenses in order to become legal, necessitating a constant, aggressive hunt for funding. Over the past decade, a long list of investors jumped on board, including Goldman Sachs, the Coca-Cola Company, Digital Sky Technologies (a Russian firm that also invested $200 million in Facebook), and Sean Parker (a cofounder of Napster).
The need for funding, then, helps explain the many changes the company underwent. To continue operating through the financial crisis of 2008 and ’09, which largely gutted the ad market, Spotify supplemented its primarily free, ad-based business model with paid subscriptions. As it expanded into the United States, it briefly became a social platform, partnering with Facebook, broadcasting users’ listening habits to their friends and enemies. To compete against other music-streaming companies like Pandora, it began prioritizing its recommendation algorithms, integrating features of Internet radio services. Spotify was no longer just a massive database of music but also a curator of musical experiences. Along the way, it became a data broker, collecting information about its users and their listening habits and selling that data to its partners.
In subsequent chapters, the authors investigate how the service functions from a technical standpoint—how a file travels through its ecosystem and how and where data is collected. They then explore the different ways music is packaged in the service, which includes everything from Spotify’s black-and-green aesthetic to the company’s categorization of songs through genre and mood. They conclude the book with a rumination on the notion of what a free service means in practice. As the now classic Web aphorism goes: If you’re not paying for it, you’re the product.
The word “teardown” in the title refers to a process in which a given product—typically something shrouded in corporate occultist mystery—is carefully dissected and all its component parts noted and observed so that its functionality, design, cost, true purpose, etc. can be more fully understood, its demons exposed and exorcised. A teardown typically is unsanctioned and violates a company’s terms of service. There’s a long tradition of this kind of activity in maker/hacker culture, in which users figure out how to repair or modify their electronics and jailbreak software to remove corporate restrictions. It happens in consumer advocacy, too, like when journalists discovered that the Juicero—a hyped $400 juice machine with Wi-Fi connectivity—was essentially useless. Teardowns even take place in the corporate world, like when Compaq successfully reverse-engineered the IBM BIOS—that is, the firmware that manages input-output operations in computers—in the early 1980s, which blasted open the PC market.
In keeping with this tradition, the authors of Spotify Teardown conducted a series of “interventions” (or experiments unsanctioned by the company) that poked and prodded at specific aspects of Spotify’s tech and business model. To learn more about how music files are absorbed into the platform, the researchers created their own record label and began uploading “obscure sound materials” to the service. This included warped recordings of advertisements that originally appeared on Spotify, in a sense feeding it its own waste. The garbled ads were allowed on the service without any issues, unlike that silent Vulfpeck album. (Interestingly, John Cage’s silent experimental composition 4’33” remains streamable on Spotify.)
Through this process, the researchers reveal that music files aren’t uploaded directly to Spotify but through third-party aggregator services that act as gatekeepers for the platform. Companies like Record Union, RouteNote, CD Baby, TuneCore, and Ditto Music do the dirty work of rights management, uploading and categorizing music, and collecting and paying royalties. For the convenience, musicians must give these services an upfront fee or sign over a percentage of their future royalties.
To figure out how royalty payments are determined (e.g., what counts as a listen), the researchers trained about 300 bots to repeatedly play their weird songs in Spotify. This “SpotiBot experiment,” as they call it, found that the platform didn’t really scrutinize their bot traffic, which earned the authors $6.28. (Don’t worry; they never cashed the check.) Not surprisingly, the interventions eventually drew the ire of Spotify. The authors received a letter from the company’s legal counsel, demanding that they cease any actions “in violation of Spotify’s Terms of Use.”
Though the book is incredibly thorough, if there’s one area left unexplored, it’s the worrisome environmental consequences of streaming. A recent study by the University of Glasgow and the University of Oslo found that while streaming music has resulted in a drop in plastic waste, it has increased carbon emissions. The data centers that streaming services rely on require a tremendous amount of energy to operate. (It’s not a glaring omission, though, since there’s a whole book coming out in the fall that expands on the study, Decomposed: The Political Ecology of Music.)
While Teardown seems chiefly aimed at an academic audience—its long explorations of research processes and experiment design aren’t fare for the typical pop nonfiction tech book—it remains highly readable. It’s never really overwhelmed with jargon, and a lot of it is pretty humorous. At one point, for example, the authors invent a joke app called Songblocker. It serves as a reverse ad blocker, muting everything on Spotify except for the “100% awesome ads.”
More so, the rigor and curiosity with which they treat their subject are necessary: Spotify has completely transformed the music industry, and to truly understand how and why requires illuminating each nebulous part of the company. For one, it helped fundamentally alter the economics of music distribution. Put simply, Spotify uses a pro rata payment method, in which all the available money is dumped in one big pile and then split up among artists based on how well their individual songs perform. This method tends to favor popular acts, who bring in huge percentages of listeners.
And that’s before one factors in the percentage of those payments that record companies keep for themselves. While the old business model of the music industry was by no means equitable—as the lovably cantankerous independent musician-producer Steve Albini explained in his famous 1993 essay, “The Problem With Music,”—it was at least still possible for working musicians to make a living. Now they bring in fractions of fractions of fractions of payments. Hence that Vulfpeck Sleepify stunt.
Additionally, Spotify has changed how many people discover music. It largely uses hands-off, data-driven curation powered by proprietary algorithms, and these fully automated systems built for economies of attention are highly gameable. Mystery artists with names like Berngenulo Five and Bratte Night are invading the service. Bots, like those deployed by Teardown’s authors, can easily juke listening stats. And, as recently reported by Pitchfork, the service is extremely prone to streaming fraud.
This isn’t to say that algorithm-based discovery is necessarily bad. Consider the sudden popularity of Japanese singer Mariya Takeuchi’s “Plastic Love” (currently unavailable on Spotify), an ’80s city pop gem that was rediscovered in 2018, thanks to the alien whims of the YouTube algorithm. While YouTube has a ton of its own problems (including rampant music piracy, disturbing computer-generated children’s videos, and extremist political content), it has been integral, in some cases, to helping resurface or give shine to unknown or forgotten artists.
The problem is in the intent of algorithmic discovery. In the case of Spotify, its implementation is driven not out of altruism for giving talented unknown musicians the attention they deserve or to help us find a life-altering album but primarily to turn a profit at scale. It gives us more of what we want, eschewing the potential abrasiveness of the new with the dull, safe comforts of the familiar, serving us the blandest versions of ourselves. It’s designed to keep us listening at all costs. Listening to what, it doesn’t matter.
Ultimately, Spotify is nothing more than a cultural landlord. It owns a space, for which some of us pay rent to access our music and artists pay to have their music heard. It can raise the rent and evict occupants on its whims. Music regularly disappears from the platform because of licensing disputes, making small bits of culture and history suddenly inaccessible.
It has created another means through which advertising can invade our lives. It spies on us—what and when and where we listen to songs—and then sells its notes to brands. Those who don’t shell out for a subscription are subject to blaring ads, beamed directly into their lives between their favorite songs.
When Spotify came out, I distinctly remember attending a party—ironically, thrown by a broke musician—where the playlist was regularly interrupted by loud ads for Pepsi. The ads were unskippable and would pause if anyone tried to turn the volume below a certain level.
It was a shock to hear a brand suddenly insert itself into a formerly intimate space, especially after more than a decade of listening to music ad-free with MP3s and vinyl instead of the radio. We’ve somehow circled back to a warped version of what we had before. All that’s been disrupted is who reaps the profits off the recordings. The musicians are left shaking a tip cup.
