Liz Pelly - 'Mood Machine: The Rise of Spotify and the Costs of the Perfect Playlist'

Liz Pelly - 'Mood Machine: The Rise of Spotify and the Costs of the Perfect Playlist'

in

The cover of 'Mood Machine'.

I remember the day when a musician friend came over from the USA. He’d just finished a tour and stood in my wife’s and my apartment. I asked him if he’d heard of Spotify, and he said ‘What’s that?’ Today, my friend has largely left the music business. He works in a completely different field. Partly, this is by his own choice, but partly, Spotify makes musicians have to get other jobs or stop making music.

In some societies, musicians were seen as people to whom we looked for inspiration, guidance, comfort, love, safety, and the unknown. Musicians not granted aids: they could simply live off making music. They gave, they got. They didn’t have to worry about money.

In surveys, music often turns up as the most cherished pastime for peoples across the globe. Yet, at a time when humanity is at its most technologically advanced, when we can press play and hear nearly anything we want, musicians are at their most dire, either forced to give up what they love doing or to have multiple jobs with all the horrid terrors that entails.

How did we get here and what does Spotify have to do with all of this?

Liz Pelly is a journalist who’s been engaged with the music business and has written about music for a long time. For this book, she’s interviewed over a hundred people, ranging from Spotify employees to musicians.

As of this writing, streaming accounts for 84 percent of recorded music revenues, with Spotify, the largest of all services, capturing 30 percent of the market, with over 615 million users and 239 million paying subscribers.

In other words, Spotify have traction.

The book kicks off in the early 2000s, when big record companies were trying to grapple with filesharing and piracy. Napster, Audiogalaxy, SoulSeek, Kazaa, DC++, all of those services made a dent in the music industry.

Before the Internet hit the mainstream, record companies were old and set in their ways. They were suddenly forced to deal with the reality that kids were sharing music. Shock tactics, as when Metallica sued Napster, didn’t help. Dennis Lyxzén, full-time musician:

“There was an idea that here comes a new sort of way of viewing music, an economy that might change, or at least challenge the capitalist infrastructures,” he said, sitting outside the club between sound check and doors. “That we can share stuff with each other, help each other out, and sort of undermine the basis of capitalism. I thought that was super exciting.”

“In the early days, I was very excited about the idea that this could challenge the rigid structures of the record industry,” Lyxzén continued. “How wrong I turned out to be. I mean, 100 percent wrong. From this idea of ‘let’s challenge these power structures’ came a paradigm shift that just became a new power structure that did not benefit musicians at all. It did not challenge capitalism at all, but became an ultra-capitalist sort of thing.”

Daniel Ek and Martin Lorentzon started Spotify as an advertising tool. Peter Sunde, co-founder of The Pirate Bay:

“Daniel Ek and Martin Lorentzon were advertising dudes who wanted to sell more advertising and realized that music was cheap,” Sunde went on. (In fact, music would eventually prove to be quite expensive.) “That was the starting point for Spotify. It was not we want to save the music industry. It was more like here is a business opportunity. Every time I hear the story of Spotify, it’s this dressed up version. And the record labels want to have that dressed up version, too, because they are storytellers more than anything else. If you start looking at Spotify as an advertising company rather than a culture company, a lot of things make more sense.”

At first, the book makes it clear that Spotify wanted a big slice of cake. They wanted to make money through ads and by forcing the hands of the record business. To do this, they focused a lot on user data. A lot.

A 2016 BuzzFeed report described how the editors would use an app called PUMA, short for Playlist Usage Monitoring and Analysis, to track plays, skips, saves, and demographic data on listeners. “Songs live or die in playlists based on this information,” Ford said, on that same panel. “We live or die, and those playlists live or die, based on what the audience thinks. It’s not us, it’s the audience that drives it.” This was part of a bigger project of the platform era, one that suggested data-driven success was a meritocracy, that virality reflected the will of the people, and that social media had somehow empowered fans to nominate pop stars to power. It’s present, too, in how Spotify’s official “Charts” website claims that “hundreds of millions of listeners shape today’s streaming charts, every day.” Truthfully, though, success was also influenced by contracts, connections, and streaming-friendly music.

As Pelly says, the leap from serving listeners what they’re looking for was quickly recreated into serve listeners what they won’t skip.

One former Spotify engineer I spoke with described the TikTok feed, and Spotify’s eventual attempts at approximating it, as an ultimate distillation of lean-back listening: “You’re not putting any input in. You’re just being shown things. You’re giving it input based on which things you linger on longer, or what you skip. But it’s not like you’re choosing what you want.”

Spotify quickly grew so big that they could use all of their data to see what worked and didn’t work in a matter of minutes. Tests went live in seconds and affected how Spotify could generate more profit.

Spotify wanted ‘passive listening’, users who just pressed play on a ‘chill playlist’ and went along with the waves. Spotify found that this is exactly what they could achieve: passive users who no longer went looking for new music, to check out band details to see where they could go next.

How could Spotify use this knowledge to their advantage?

There’s a playlist called “Musical Therapy,” made by Spotify, filled with generic ambient tracks by artists that don’t actually exist. In a sea of similar playlists and personalization feeds positioning themselves as musical tinctures, this one felt especially harmful. What were the stakes of streaming services passing this off as “music therapy”? For perspective, I turned to a certified music therapist with a master’s from the SUNY New Paltz Music Therapy graduate program, Dan Goldberg (also an artist I met many years ago through the DIY music scene). “This sort of playlist is belittling to the scope of practice of music therapists and the goals they work to achieve, as they are significantly wider than just ‘soothing your mind’ as the playlist states in the description,” he told me. “There is no prescriptive music that works for everyone. It’s pretty reasonable to expect that the music from these playlists (or the associations that come along with it) might get on some people’s nerves and have the opposite effect of relaxation. Music therapists accept that someone might feel more relaxed from nu metal or hardcore punk than from easy listening.” Goldberg said that music therapy is generally not a “passive experience” and playlists like these are a growing concern for practitioners in the field. “A lot of music therapists receive training on improvisation techniques that welcome and empower the people they work with to play, express themselves, and take the lead. This certainly isn’t possible from an online playlist.”

Pelly keeps level while exploding the innards of Spotify, which since longed ceased to be a musician’s help, a vehicle to help their career, and instead turned into the musician’s worst enemy.

In her seminal article Ghosts in the Machine1, Pelly clearly showed something that was first exposed by Swedish newspaper Dagens Nyheter: Spotify employed a few companies (for example, Swedish companies Epidemic Sound and Firefly Entertainment) that use ghostwriters, so-called PFC artists, who sell songs for scraps. Those songs are then pushed into Spotify playlists that get maximum exposure and generate the most money possible - for Spotify, who don’t have to deal with more expensive (ergo less profitable) artists.

Spotify have for a long time actively been trying to rid themselves of the hassle that is artists and record labels, which they both see as parasites who try and take their profit. In the search for more profit, for music (which is called ‘content’ by Daniel Ek) to quickly sell, they’ve used artificial intelligence, AI. There’s nothing intelligent about the sounds that are generated by AI for Spotify, but it is telling.

Another side-effect of Spotify is that they clearly show how little they care about their users. For example, they found out that a lot of users looked for the word ‘chill’ when searching for anything to which they could listen. So, Spotify started making en-masse playlists with names like ‘Chill Morning Beats Mix’ and ‘Morning Chill Instrumental Mix’, names that don’t mean anything. What ended up in those playlists?

Another anonymous Spotify employee told me that no one could agree on what “chill” meant. “Chill has no definition that is meaningful,” he said. “Everybody has a chill mix, and it’s all music that they would never bother to turn off. It’s good sometimes, but it’s ambiguous. It probably has a strong meaning to every individual user, but those meanings don’t often agree with each other. For me, it could describe at least 30 percent of the music that I listen to. Indie rock that’s not too loud all counts as chill. It’s really vague.”

Recently, some variation of this ambiguous data-tuning process led a very unchill track by the singer and songwriter Anohni called “Why Am I Alive Now?” to appear near the top of my “Chill Vibes” playlist on Spotify. The playlist is “algotorial,” meaning that it was created using a mix of editorial and algorithmic input. When I delivered the news to Anohni on a video call—that her song had been absorbed into this chill-vibes playlist—she appeared visibly perplexed. “That song is so despairing,” she said, deadpan, before reflecting on the dangers of the current climate of decontextualization.

This type of derogatory filth, an attempt at bad en-masse destruction of language, is made clearer the more Spotify use AI:

On a different day, while listening to the AI DJ, the app described an upcoming block as a dive into the world of “metropopolis,” which immediately stuck out to me as a Spotify-created genre name. For years, Spotify had been coming up with these bizarre microgenres, a popular subject of conversation on social media especially each year during “Wrapped.” Spotify had increasingly come to use the year-end campaign for showing users not just their top artists, but also their top genres, which would often include strangely specific names that most users had never heard of before. And in turn, each December, music fans would take to Twitter to express their confusion that Spotify had called them a top fan of “solipsynthm,” “braindance,” “traprun,” “otacore,” or “escape room.” Apparently Spotify had been beating the “metropopolis” drum for a while; back in 2014, Slate ran a piece under the headline “Stop Trying to Make Metropopolis Happen.” I guess no one at the company got the memo.

Spotify make up words and terms that don’t mean anything, just to make more of a profit. This is standard fare in the world of advertising. But do Spotify try and change users?

Music journalists began taking note of how streaming was changing the sound of pop music around 2017. In terms of song structure, because streams were only monetized after thirty seconds, there was a particular emphasis placed on perfecting song intros.

Not only does companies like Spotify not care about the people who pay them, but they change how musicians think. So, Spotify are philistine. They are both indifferent and hostile to artists and music. What about profits?

Pelly’s access to Spotify employees has shown to be crucial in her details. A lot of Spotify messaging conversations show how employees, over time, went less happy about music and more greedy, more into squeezing as much money as possible from the artists.

Epidemic Sound, one of the PFC companies, claim to ‘simplify’ things for musicians. What this actually means, is they take away musicans’ money before handing over their product to Spotify:

In claiming to “simplify” the mechanics of the background music industry, Epidemic has championed a flat-fee buyout system, where it pays a set amount to retain all rights to the song. One Epidemic composer I spoke with said his payments were routinely around $1,700. “All tracks I made for Epidemic are purchased by them as a complete buyout,” he told me. “They own the master.” But he still received royalties from every streaming service that played those tracks. Epidemic’s selling point is that the music is royalty-free, but it does collect royalties from streaming services, which it splits with artists fifty-fifty. The royalties collected are, of course, smaller than what non-PFC tracks would generate, due to Epidemic’s discount deal with Spotify. But that deal also means the tracks are boosted on the platform: the company says its catalog gets 40 million streams per day. Still, the composer I spoke with said his streaming royalties from Epidemic tracks were always smaller than royalties from non-Epidemic tracks. And there are certain royalties artists are not entitled to at all: to refine its exploitative model, Epidemic also mandates that its composers resign from their performance rights organizations (PROs), the groups that collect performance royalties for songwriters when compositions are played on TV, radio, online, or even publicly, in a cafe or a retail shop.

Is this how musicians are supposed to act? They make the product, the be-all, and hand it over to a PFC company that takes not only a significant cut but most of the money that an artist could make if, for example, one of their songs becomes popular or added to a TV advertisement. The PFC company then licenses the track to Spotify. Because the track is made by a throwaway company that brings Spotify the biggest profit, that song gets pushed to one of their more successful playlists.

This is what happens at Spotify: they profit from artist misery.

The biggest scam in all of this, is that Spotify make it look as though artists should be grateful to use Spotify, the biggest marketing tool in the world, to be able to use their platform.

Remember, Spotify have actively lobbied to lower songwriter royalties2. At the start of the year 2024, Spotify stopped paying for around 86% of all streamed music on their platform3. Let that sentence sink into your head for a little bit.

As independent artists have come to terms with the state of streaming, some have pondered this question: Is $0.0035 really better than nothing? It’s one thing to make nothing from the free circulation of your work, but it’s another to make nothing while Daniel Ek and Martin Lorentzon become billionaires, and the majors earn higher and higher revenues. Another reckoning came in late 2023 when Spotify announced a substantial update to its royalty system. In a supposed attempt at fighting fraud, it would be demonetizing various types of white noise and “non-music content.” In addition, the new model also would demonetize any track garnering fewer than one thousand streams annually—an estimated 86% of the tracks on the platform, to which Spotify now seemed to say, Actually, we do think this is worth nothing.

Yes, lobbyism plays a big part of what Spotify do; submerge the work of the majority (the musicians) into the profit for the minority (the top Spotify shareholders):

And part of Spotify’s strategy for gaining political influence is indeed getting lawmakers and DC insiders onto the platform. Case in point: in 2019, Spotify paid $240,000 to the DC lobbying firm Peck Madigan Jones, which has since rebranded to the Tiber Creek Group, hiring four of its lobbyists to meet with politicians and sway them on issues related to licensing, copyright, and music industry competition. Additionally, according to lobbying reports, they were to “inform Members of Congress how to use Anchor,” referring to the podcast creation software Spotify had purchased that year, “to create podcasts to communicate with constituents and others.” This was the year that Spotify made its big company-wide bets into podcasting, reimagining itself as not just the world’s biggest music streaming service, but the world’s biggest “audio first” service, paying hundreds of millions to bring Joe Rogan and other big names exclusively to its platform. And while it is difficult to prove, for sure, whether all of this lobbying was successful, within a couple of years so many lawmakers had launched podcasts that the Capitol Hill publication Roll Call published a blog post about how “Congress wants to get into your earholes,” and in 2022, Politico ran with the headline “Wow, Politicians Are Really Bad at Podcasting”—actually, not just bad, but “characterized by a crushing, mind-melting boredom.” It makes sense that Spotify would want politicians to become playlisters and podcasters: by offering support, it could earn the company goodwill, and forge connections in DC.

Pelly has written a book that for music does what Shoshanna Zuboff did for the tech industry as a whole. Zuboff wrote Surveillance Capitalism, a major work that showed how tech companies scavenged, used, abused, and sold user data, most often without the users knowing about it. Some of these practices have been curtailed by regulation, never by companies having changed their ways.

This book goes deep into far more than what I will bring up in this simple rant-cum-review. Her words on how much a stream actually costs is on par with what Steve Albini produced in his 1993 monograph, The Problem with Music4, which was basically an attack on big music companies, long before streaming came into the picture. Pelly is able to explain part of the consciously-made complicated streaming-payout model that Spotify use, for which I laud her. There are many disgusting things revealed in the book:

There is so much more about the nature of streaming royalties that remains unknown to the public. But some additional details have come to light over the years. In 2015, just a few years into Spotify’s launch in the U.S., its contract with Sony was leaked to The Verge, revealing that the label had received a $25 million advance for its first two years, with no clarification of whether this had to be shared with artists; it received $9 million in free ad space, which it could either use or sell for cash, with no stipulations about whether that income had to be shared with artists. Sony had guaranteed for itself a series of complicated negotiations regarding per-stream minimums on the free tier and per-user minimums on the subscription tier. And, of course, they had secured themselves equity in the company.

One of the things I like the most with this book, is Pelly’s ability to keep a level head, carry forth expert understanding of the music business and how it’s being curtailed by big tech, and what can be done about it. She looks into what data Spotify collect about their users (spoiler: everything) and how they consciously wronged all user-data compliance, to the extent that a Swedish court fined Spotify five million Euro in June, 20235, for not being transparent with users on what data they keep.

Pelly is no pessimist. She speaks with representatives of musician union United Musicians and Allied Workers (UMAW), people who have made platforms that pay musicians reasonable, simple-to-understand, and fair fees (Resonate and Catalytic Sound, for example), and talks about the future.

None of this is new. Neoliberalism, capitalism, feudalism, fascism; it’s all the same, to a large extent. I’m reminded of what Ursula Le Guin once said:

We live in capitalism. Its power seems inescapable. So did the divine right of kings. Any human power can be resisted and changed by human beings. Resistance and change often begin in art, and very often in our art, the art of words.

According to UMAW, in 2024, Daniel Ek earned a salary that was 2.5 billion (!) times that of the average musician. His personal wealth is valued at billions of U. S. dollars. What does he do with a small part of that money?

It matters that Daniel Ek has used his fortunes from Spotify, in part, to invest in AI militarism, through a 100 million euro investment in the German military-AI company Helsing, putting that company’s valuation at 400 million euro. Helsing alleges to use its software to “assist battlefield operations” by using live data to “identify and assess multiple collected forms of data via sensors in order to assemble a picturesque viewpoint which military agents could then use at their discretion.” Ek sits on the board of Helsing, and has publicly advocated for the expansion of investment in its technology. In 2022, he coauthored an op-ed for Politico arguing in favor of what he dubbed “New Defense,” urging for a new wave of AI military companies. The piece was riddled with tech-culture buzzwords, calling out the need for defense contractors to be “digital natives” with “agile and iterative development practices.” It should go without saying that many musicians would prefer the profits made from circulating their work did not fuel the war machine, as was evidenced most recently by the United Musicians and Allied Workers’ 2024 “War Profiteers Out of Music” campaign.

‘New defense’. Funny, that. How oligarchs use words like ‘defense’ to really mean ‘offense’. Just as ‘Spotify’ actually means ‘against musicians’.

Sweden was once a country that generated musicians by leaps and bounds. This was mainly built by a government that gave kids no-cost access to training, instruments, rehearsal spaces, and freedom to experimentation. This is collapsing if the current right-wing government gets their way. In contrast, Ireland has experimented with a three-year Basic Infcome for the Arts pilot, where 2000 artists (not only musicans), got payments to suppleant what they tried to do for a living. The payments were minimal, and the effects were gargantuan: mental health went up, musicians had time to rehearse and create music, to collaborate, to create what…what’s that again? Oh yes, create what brings most human beings the most happiness of all leisurely pursuits.

At the end of the day, Liz Pelly has written a monumental book, one that should be thrown at politicians and then engage in meaningful conversation. No longer can a handful of persons rule millions. Remember what our ancestors fought against during all wars? This is just one of them, but it’s happening now, and Pelly has put together a carefully crafted blueprint for what we should rail against with all of our might.

Mood Machine: The Rise of Spotify and the Costs of the Perfect Playlist is published by Atria/One Signal Publishers on 2025-01-07.

  1. Pelly, Liz. “The Ghosts in the Machine.” Harper’s Magazine, December 18, 2024. Accessed December 20, 2024. https://harpers.org/archive/2025/01/the-ghosts-in-the-machine-liz-pelly-spotify-musicians/

  2. “Songwriters of North America Hit Back at Spotify CRB Appeal.” Music Business Worldwide. Last modified March 11, 2019. Accessed November 27, 2023. https://www.musicbusinessworldwide.com/sona-hits-back-at-spotify-crb-appeal/

  3. “Songs on Spotify Will Need 1,000 Streams to Make Royalties from 2024.” DJMag.Com. Last modified November 7, 2023. Accessed November 27, 2023. https://djmag.com/news/songs-spotify-will-need-1000-streams-make-royalties-2024

  4. Albini, Steve. “The Problem with Music Steve Albini.” The Baffler. Last modified December 30, 1993. Accessed May 9, 2024. https://thebaffler.com/salvos/the-problem-with-music.

  5. European Data Protection Board. “IMY Issues an Administrative Fine against Spotify for Shortcomings Regarding Transparency.” Last modified June 12, 2023. Accessed December 23, 2024. https://www.edpb.europa.eu/news/national-news/2023/imy-issues-administrative-fine-against-spotify-shortcomings-regarding_en