Switched on Pop - How Spotify hacked our ears (and our data)

Episode Date: February 4, 2025

Behind Spotify's promise of infinite music lies a carefully engineered system that shapes not just what we hear, but how music itself gets made. Journalist Liz Pelly's explosive new book "Mood Machine" rips away the curtain on streaming's biggest player, revealing how its algorithms and backroom deals dictate the soundtrack to our lives. With major labels controlling 70% of streams and Spotify commanding over 600 million users, the stakes couldn't be higher. As artists like Björk decry streaming as "the worst thing that's happened to musicians," Pelly uncovers the true cost of our perfect playlists - and what we're really sacrificing for the illusion of endless choice. MORE Subscribe to our newsletter to receive your own bingo card! Learn more about your ad choices. Visit podcastchoices.com/adchoices

Welcome to Switched-on Pop. I'm songwriter Charlie Harding. There's no denying it. Streaming has completely changed the way that we access and enjoy songs. It gives us an unprecedented library right at our fingertips. And for some independent musicians, it's broken down traditional label barriers, providing a platform to share their art with the whole world. Yet not all that glitters is gold. Just recently, Bjork described Spotify as probably the worst thing that's happened for musicians. Spotify may seem like... like a limitless jukebox, but behind the scenes, it's anything but neutral. Label-owned music still dominates at about 70% of music listened to on Spotify, and the platform

is shaped by tons of hidden business agreements and opaque algorithms. And because Spotify is the biggest player boasting over 600 million users and more than 30% of the global streaming market, its business model doesn't just reflect the industry, it actively reshapes it. And that's why I want to speak to Liz Pelley, a music journalist who has contributed to the Baffler, The Guardian, and NPR, among many other publications. Full disclosure, she's also an adjunct at NYU, where I also teach. Liz has a new book called Mood Machine, the rise of Spotify and the cost of the perfect playlist,

where she suggests that much of what we're listening to isn't actually what we think it is. And that's why I wanted to have Liz Pelly on the show today, to deliberately listen to less music than we usually do in the show in order to understand how the music business, its practices, and algorithms are shaping us as, listeners in ways that we may not suspect. If you're tired of endless scrolling to figure out where to eat, same. I'm Stephanie Wu, editor-in-chief of Eater. We've just launched the new-ish and way better Eater app.

It has all the restaurants we love, gives you personalized picks wherever you are, and serves up smarter search results just for you. You can find my list of the best places for martinis and fries in New York City. And save your favorite spots, share lists, follow everything. editors and book right in the app. Download the eater app at eater app.com. It's free for iOS users. I want to start with a seemingly simple question. What is Spotify? Great question. That's a question that actually has surprisingly come up a lot. And it's really funny the first time someone asked me that question, I did this really literal response where I explained what a streaming service was

and I explained how, you know, the files get stored in a central server and you use the internet to access them. But I think that maybe a more instructive way of explaining what Spotify is is just to think about the history of the company a little bit. You know, I mean, Spotify is the most popular music streaming service in the world, but it also in recent years has become more than just a music streaming service. Where Spotify is now, I think, with the expansion into podcasts and audiobooks and not just specifically being a music-focused platform, in some ways, actually really makes sense when you think about the early history of the company, too. So let's go back to, what, 2005, 6?

Yeah, 2006, I think, is a good place to start. So if you are to believe Daniel Eck and Martin Lawrenceon take them at their word, they decided to start Spotify in April 2006 at Martin Lawrence's birthday dinner in Stockholm. I'm pretty sure it was his 40th birthday. And at that time, you know, Daniel Eck had been pitching Martin Lawrence on an idea for something like Spotify, according to Swedish journalists since the following fall, they both had backgrounds in the advertising industry. Right.

And the idea was to, in some way, you know, build on the popularity of music piracy in Sweden at the time where piracy wasn't only incredibly rampant, but also had more of a political dimension than maybe it did here. There was a pirate party in Sweden. Which is pretty cool. Piracy, not so cool. Pirate party. Interesting.

Yeah. It was very rampant in Sweden. And then also, I think, just sort of politicized in a different way. So, you know, according to people who were in Sweden at the time who were kind of involved in the piracy scene, the quest for a solution to piracy wasn't just something that was being discussed within the music business, but also in the mainstream media amongst politicians, just had sort of a different life. And we should say, like, the larger context here is that we are post-peak CD, Napster has been attacked at every front by the music business. but piracy is rampant and has completely, nearly like halved the music industry revenues. Like that's part of the sort of musical context. Exactly. Yeah.

So like at this point, you know, the music business has spent years trying to stamp out file sharing and music piracy. So anyway, Daniel Eck and Martin Lawrenceon come up with this idea for a streaming platform that would sort of in some ways replicate the user experience of a fully stocked iTunes. library that maybe is similar to the experience someone had become used to in the post-piracy, post-file-sharing era where anything that you could possibly want to listen to, you could listen to right away. And, you know, if you hear them talk about it, they had a really specific focus on trying to make it seem like all the music was already downloaded on your computer. And then the original idea was that this would be supported through ads. And the original idea was for an advertising platform, like an ad-supported. streaming service. But in the original patents, in the early history of the company, it wasn't specifically going to be a music platform if you listen to early interviews with Martin Lawrenson. And they're also their first employee, Andres N, who they hired in the summer of 2006. They both talked about how at first they considered doing a video platform. Maybe, you know, Martin Lawrenson has said maybe they were going to do audio books, maybe product search. And they

ended up, you know, settling on music. And there's lots of different reasons why that might have been. Maybe the music industry's weak status in their home country at the time, maybe because music files were smaller and easier to launch a beta product with. Especially, this is like Web 2.0. Like the internet is still a lot slower. It makes sense to have smaller files. Your book reports that they begin the service using files that they got from the pirate bay. The actual like legal service eventually was hosting pirated. files, that's what you were streaming. That's how they first got access to that information. Yeah, and that's something that other Swedish journalists have reported on over the years two, apparently in 2009, you know, after they'd been launched in Europe for over a year, they did something called the Big Clean, where they wiped all of the unlicensed music off of the platform finally. But yeah, the early beta product was made using whatever music they had on their computers, basically. And part of the way we know this is that some artists had release their music and it was only available on Pirate Bay and no other platforms, but they were also on

Spotify. So it's like, how did you get that music? Yeah, exactly. Okay, so we begin as a bunch of advertising executives looking to build a technology platform around some kind of media. And given the cultural conditions of piracy in Sweden, it's sort of like ripe for some kind of solution to break the stronghold that Pirate Bay has on the country's listener habits. Next steps, how does it legitimate itself and become the predominant musical platform. Yeah. Part of the reason that I think they were able to be so successful in Sweden was because of this way in which piracy existed there. Talking to people who were in the music business in Sweden at the time, they talk about

how the music business had started to think of Sweden as a lost market, how no one was ever going to buy music there. So even though the music industry at the time was like super allergic to anything that had the word free in it because they were trying to get away from people thinking of music as something that you would access for free. They were willing to try a service with a free tier there. So that's something that, you know, at the time made Spotify really different was that they were proposing this ad-supported free tier. But at first, that was the whole idea for the platform was ad-supported free music. And it was through negotiating with the major labels that they came up with this idea of, you know, freemium model. But even to date, you know,

the freemium model, other services now have experimented with something similar, but it is something that I think has made Spotify sort of unique in its history is this reality that they bring so many of their users on through an ad-supported free tier and then try to convert them to the paid tier. I think that's something that was an advantage to them when they launched. That speaks a little bit to who their users are. I mean, obviously they represent so many different kinds of people. Who is Spotify really built for, though? like for fans, is it the music snob, is it the casual listener?

Is there a specific music listener that they're really building for, either in its origins or today? Yeah, that's a really interesting way to think about it. You know, if you look back at the history of the company, after the launch in Europe, 2008, and then the launch in the United States in 2011, you know, you'll read stories about how it wasn't super hard for them to get those really enthusiastic. music people onto the platform. You know, in the United States, when Spotify launched, there had already been so much hype in the music media about, you know, this Swedish music app that's going to be coming to the United States at some point. And Rolling Stone had, like, written a series of article as sort of like hyping up how great Spotify was. So it's easy for them to

get those dedicated music listeners or music enthusiasts at first. But in 2012, towards the end of 2012, when they started to kind of announce changes to the platform, forum, you know, at first, the experience of using a streaming app was really more like a search bar. You'd have to know specific artists or albums that you're going to be searching for or you would have to, you know, know of a user-generated playlist or something, maybe a friend's playlist or some sort of like curator who you were a fan of. After they launched it in the United States, perhaps owing to the fact that there were more streaming services, more digital music services in the United States. Also in the U.S., the iTunes Library had taken off in a way that it had in a country

like Sweden. Steve Jobs was running around on stage, promoting his music, and so excited that he finally scores a deal with the Beatles at some point, who had a contrasting legal issue at one point, given their Apple music. You know, iTunes was doing great. Right. So there was, like, you know, there was that to compete with and other services like Pandora. And just, you know, in my book, I talk about this moment in 2008 or 2009 where there were like dozens and dozens of other digital music services in the United States that people could access in order to listen to music online. So there's more competition. And in order to grow their user base, they couldn't just think about those enthusiastic music fans.

They had to also think about how to attract more of, you know, what the music industry now calls the lean back listener or the passive listener in order to reach a more mainstream audience and grow. It's really interesting. Like when I was researching my book, I spent a lot of time on the internet archive, just like searching the front page of Spotify.com to see how the homepage changed over the years. And there was this moment sort of towards the end of 2012 where the homepage was radically reimagined from focusing on things like instant, simple, free. All of a sudden, there's more of this emphasis on the different moments and moods within which people listen to music and kind of trying to get the Spotify user to imagine themselves as a person who is listening to Spotify all day throughout all of these sort of discrete

moments in their day, you know, like when they're waking up, doing yoga or exercising or meditating or like on their daily commute, driving in a car, hanging out with their kids or whatever. So throughout 2013, there was sort of, you know, what some researchers have called the curatorial turn or the shift away from being a search bar and more towards these mood and moment playlists. That's around the time that they bought this startup called Tuna Go and introduced their editorial team for the first time, created the concept of the official editorial Spotify playlist when they had a team of editors for the first time and really started to sort of like prioritize these editorial playlists like your favorite coffee

shop or today's top hits or morning commute, things like that. Okay, so one of the things that's interesting about these different user models is that you have in our background, the music industry is still struggling and trying to catch up to where it was in the CD era. And moving into streaming where streaming isn't a replacement for purchasing albums, even if maybe streams are calculated in some album equivalent. It's also a auxiliary for radio and different modes of consuming music. You just described in a number of playlist. Everything from top hits, that's a radio format, but also like coffee shop, like a context, a mood.

And I feel like this is one of the big shifts that happens. There's an era really up until the TikTokization of music where Spotify playlists, they will make or break your career. If you could get on the top of a playlist, it could change your life, right? Spotify likes to say that it democratizes listening. Like the way that this happens is just like what people like gets on the playlist and that it's just about like, yeah, what music is good. But your book reports and we've reported on how Spotify also informs the kind of music that does well on Spotify. people have claimed that there's a Spotify sound, Spotify core. Could you speak to the ways in which Spotify is actually changing the sound of music in this era

and how its playlist model maybe isn't the open democratic model that it purported it to be? Yeah, when I first started writing about Spotify and streaming, one of the things that was really going on at that time was that independent musicians were being sold on streaming as this level playing field that was going to democratized curation that was going to diminish the gatekeeper power of commercial radio stations and major record labels. But what's ended up happening over the years is that streaming services in some ways have emerged as new types of gatekeepers. And in this period of time following, you know, the introduction of the Spotify editorial playlist as this sort of new kind of cultural

phenomenon or as one person I interviewed for my book described it to me kind of like the top of the discovery funnel during what this former Spotify employee called the peak playlist era, which this person estimated was roughly from like 2016 to 2019. Where if you ended up on rap caviar or on today's top hits or even a smaller playlist, there was a whole system of tiered playlists. Even a smaller playlist could get you on bigger playlist and you could be, there were some independent musicians that all of a sudden did really well because they did get place on some of these spots.

Yeah. But something that's important to keep in mind about like the whole playlist system is that first with these editorial playlist and then even as there has been this shift towards way more personalization and algorithmic recommendation in the peak playlist era, in the more highly personalized era of streaming curation, all of these playlists and discovery tools exist in order to boost user engagement. And when you have systems of curation that are built in, order to boost user engagement, to boost streams, there's certain ways of thinking about curation and context and like meaning and then also value and music. There's also certain types of music that is going to do well in these environments that has, I think, impacted for some musicians the way that they're able to, you know, do well in the streaming era or not do well in the streaming era, especially in this peak playlist era. It was also a time when it wasn't just that like the Spotify playlist as a concept was coming more influential, but there are also

certain types of playlists like, you know, what the music industry calls functional music, you know, term also used by Musak in the 1930s to describe this idea of music as not just, you know, music, but as a means of mood stabilization or a way that industries and technology companies sort of sell the promise of mood stabilization. So the rise of the curated playlist didn't just sort of like put forth this idea of Spotify employees as people with a lot of cultural influence. It also put forth this idea of music as utility or they started sort of normalizing the idea of you go to a streaming service and it's not necessarily what you're looking for is music, but what you're looking for is a certain mood or a certain vibe or, you know, like a certain activity playlist or something like

that. I really like in your book you document that there is a long history of this 100 years long. Eric Sette, the French composer, made what he called furniture music. Thomas Edison had his own music company. Was it called mood music? So yeah, you know, like in the early recorded music era, when the Edison company was selling these things called Diamond Disc, were sort of like a signature type of vinyl record. They also, they had this catalog that they would, you know, send to potential customers also that was sort of categorizing their offerings of recorded music according to various moods. And even then they had commissioned a possibly like pseudoscientific sort of research study on their own material in order to create this catalog where they were sort of like pitching

the music that they were selling to potential customers based on different moods. And they also would host these things called mood change parties, you know, where they would kind of like have people host parties where people would listen to certain types of mood music and then like document how their moods change. And I think it's so interesting to look back and see examples of this throughout the history of recorded music, whether it be like the Edison Corporation

encouraging people to host mood change parties or, you know, later on Musak claiming that the really like boring background music that they were selling to workplaces was going to have this effective increasing product amongst their workers, down to now Spotify, commissioning neural marketing agencies to have a bunch of listeners put on brain scanning headbands and track how their sentiment changes when they're listening to different mood playlists. You know, it's like there's so much

history for all of these sorts of like claims of music as a mood stabilization agent and different ways in which various different types of tech companies have like tried to capitalize on that. It's kind of like the snake oil salesman claiming music as Prozac when often using pseudoscience or discredited scientific journals. There's a whole field of music therapy that really acknowledges the intersubjectivity of music that you can't just play. There's not a universal dose of some kind of music that will solve all emotional issues or change everyone's mood. but rather it is a highly interpersonal relationship that one can have with music that can have therapeutic benefits, but generally not on mass is my understanding. Yeah. I mean, one of my favorite interviews that I did for the book was I interviewed a music therapist for my chapter called The Conquest of Chill, who directly was like, in the field of music therapy, we fully acknowledge that listening to metal might make someone more calm than listening to easy listening because it's so subjective and how so much also of music therapy.

involves people participating in the process of making music. It's not like something that can be really boiled down to you just hit play on a playlist called music therapy, which is, you know, these are the sorts of claims that different playlists in not just on Spotify and not just in like the realm of Spotify curation, but there's sort of this whole like cottage industry of self-proclaimed like wellness soundscapes created by various like AI processes and stuff. And yeah, I think that that whole kind of cottage industry is something to be deeply interrogated. Maria, you have a podcast now and you need to start acting like it. What's the first step as a podcaster? Well, you have to ask lots of questions. I'm Maria Sharpova and I'm hosting a new podcast called Pretty Tough. Every week, I'm sitting down with trailblazing women at the top of their game to discuss ambition, work ethic, and the ups and downs that come on the path to achieving greatness.

I have a few pretty tough questions for you. Okay. Ready? Do not sugarcoat something for me. No, no. We'll dive into their stories and get valuable insights from top executives, actors, entrepreneurs, and other individuals who have inspired me so much in my own journey. Pretty tough is your front row seat to the women who have demonstrated the power in being unapologetic in their pursuits. I hope you'll join us.

New episodes drop Wednesdays on YouTube or in your favorite podcast app. Immigration may be Donald Trump's signature issue. President Trump is now targeting predominantly Democratic cities for ice raids and deportations. Dozens of protesters clashing with immigration and customs enforcement agents in Minneapolis Tuesday. We will begin the process of returning millions and millions of criminal aliens back to the places from which they came. But what we want to do in this space is talk about America and politics beyond the current president. So what do most Americans think about deportation and border security? period.

I think that Americans are definitely against the kind of violent displays that we've seen in the street from ICE. When it comes to the question of deportation, the answer is more complicated. My sense is that people want border at the border. They don't like the idea of having no idea who's coming into the United States at any given time. The view on immigration from the bottom up instead of the top down. That's this week on America Actually.

Every Saturday in your audio and video feeds. So if mood-based functional music is where it's at on Spotify, should I just go and commit my career to making lo-fi beats music? Because this afternoon, I did make you a lo-fi beat. How do you make it? Oh, I made that in Ableton. I made that. I'm a musician. Yeah, I didn't even use a AI.

You didn't even use like boomy. Wow. There are platforms you can do. So should I be making that kind of music if I want to be succeeding in the world of music today? I mean, if you talk to people who are in that scene, you know, one of my other favorite parts of reporting for this book was interviewing a handful of producers and the sort of lo-fi hip-hop beats to study and relax to scene. And something that came up that I thought was really telling was a couple of the different producers who I interviewed talked about how it's maybe easy to break through and reach, you know, a large number of listeners on streaming. but none of those listeners are really connecting with individual artists. So, like, maybe your streaming consumption numbers will be high, but, like, the number of people actually connecting with you as an artist will be really low.

If you're 17th in the playlist, I don't know who your name is because it's on the background while I'm studying and working. Yeah, totally. And then also I would say that, yeah, chances of making that music and ending up on an official Spotify playlist seems pretty low these days. But there are ways to end up on the official Spotify playlists. There has been a lot of reporting over the years dedicated to issues of everything from, you know, fake streams, but to fake artists, ghost songs. I kind of felt like that reporting had kind of been put to bed and maybe that issue had been solved. But your book really opens that back up and suggests that maybe Charlottetan is my producer handle, trying to make my own low-fi beats music on my own, just trying to pitch it to a Spotify playlist, might not work. Yeah, maybe you could send it to Epidemic Sound and maybe you have a chance of getting it, you know, on a Spotify playlist. Break that down from it. Epidemic Sound and this whole, like, what is, there's these sort of like not real artists all over the platform, especially in these functional playlists. Yeah, it's super interesting because I also maybe thought that the reporting on that subject was over as well until I started working on my book.

So some of the reporting on this supposed issue around a fake artist goes back to 20. I think it was the first time music business worldwide reported on it. Apparently, a source had told them that Spotify was filling its functional music playlist with fake artists that it was creating. This was like some of the original reporting claimed. And that playlists for studying and sleeping, focus, chill out playlists, we were being filled with this like low royalty stock music or music commissioned by Spotify. Spotify really quickly in response to some of the initial allegations said that we're not creating this music. Like we don't we don't create this music. But they sort of never fully denied that there was some sort of broader arrangement

going on, which I think is partially why music journalists and people on social media, people on Reddit threads kind of like continued sort of digging into this over the years, trying to get to the bottom of it. Because if you looked closely at these playlists, especially focused playlists or sleep playlists, it was really obvious that they were totally filled up with music by artists where there'd be maybe like a weird generative AI image for the photo, there'd be no information about these artists anywhere else on the internet. It'd be hard to find social media profiles for them or maybe there'd be no bio, this kind of thing. And kind of in my mind, I was always just like, oh, this is probably just like teenagers trying to scam the music

business in their bedrooms or whatever. I thought it was maybe like individual scammers at first. Then it became pretty clear over the years that there like seemed to be some. sort of direct relationship between Spotify and this handful of companies like Epidemic Sound, which is a service that offers production music, royalty-free production music. So it's not just that their music ends up on Spotify. You know, like if you're a YouTuber and you want access to royalty-free beats to play in the background of your YouTube video, like, while you're giving a monologue or something. As far as I know from talking to different musicians for a long time, this has sort of been something that maybe musicians will do on the side or like as a day job.

It's just kind of like make music for advertisements. And, you know, so in 2022, the Swedish newspaper, DN, did a big report on it where they were able to use copyright documents and pieced together that there is actually this really small handful of songwriters, affiliated with not just epidemic sound, but this company called Firefly Entertainment A.B. And that this small handful of songwriters were responsible for like hundreds of artists monikers and hundreds of tracks that were like all over these really popular first party playlists. and that there was also like a clearly documented direct relationship between a former senior music programming executive at Spotify and the founder of one of these supposed fake artist companies. They had played in a band together previously. And they had played in a band together previously, which is something that I dug up while I was reporting my book. So, you know, a pretty clear connection.

So just to be clear, you're going to your stress relief playlist. The first artist is a real artist. The second artist might be a real artist. The third artist, not a real artist. literally like one of maybe a couple dozen people that are making hundreds of songs under hundreds of artist names and you can't really trace back to them. Yeah, totally. And, you know, some of these playlists are just fully made of this music too. Like, you know, I interviewed sources, including musicians, like people who worked in the music business that time, former Spotify

employees and really pieced together more information about this specific program that exists internally at Spotify where there's this small group of editors who are specifically tasked with overseeing these sort of more lean-back mood playlists and that there's this specific group of licensers that commission this music from musicians and then Spotify gets this music at, as they say, with improved margins to them, which seems to suggest some sort of different royalty arrangement or different rate. When we listen to a song that counts toward some larger pool of money, that gets paid out to artists eventually,

or gets paid to, we should say, rights holders, which then pay their artists. And potentially with these, with epidemic sounds and others, there is a diminished royalty rate that they have negotiated. That's what seems to be true. And, you know, something that's super interesting about this is that Spotify, in response to it, you know, one of the only things they have really said is that,

well, this music is just a small percentage of total streams on the platform. And I think that's super interesting because even a small percentage of total streams on the platform is still a percentage of total streams on the platform. And I was reading this article just, you know, the other day about how, for example, like in 2013, I think it was, you know, like Taylor Swift's, like total stream share on the streaming service is like just under 2% of like total stream share. And there's this music business worldwide headline from last year that's like Taylor Swift's total stream share is more than the entire. genres of jazz and classical. So you think about like the entire genres of jazz and classical don't even comprise 2% of total stream share, then for this type of music to be, oh, it's a very small percentage of total stream share. It's like, well... Small percent. That could be 7%. That could be more than Taylor. You know, it's still small, speculative, yeah. But I think that

people who make jazz, classical, ambient music, lofi hip-hop beats, you know, this isn't a small deal, like especially to musicians who have experienced being replaced on some of these bigger first-party playlist or musicians who have been trying for years to sort of break through on streaming services and have been mystified by these systems or unable to kind of like reach playlist curators. It's interesting, like talking to people who work at independent record labels on one hand, a lot of them will talk about how, well, the impact of official Spotify editorial playlists isn't as impactful as it was a few years ago. Like, it's all about personalized recommendations now.

But while I was reporting my book in 2022, I interviewed two people who run an independent record label who talked about how, you know, even then they really were relying on editorial playlist placements to drive their streaming revenue. So, like, that was, you know, a few years ago at this point. But I still do think that there is an extent to which this matters. Well, there's a, there's a euphemism for this better content. Better content, meaning better for Spotify. They have to pay less. There's an incentive for them to promote more of this. What do they call it?

Perfect fit content. Perfect fit content. You report that there is pressure from management to get playlist editors to put more perfect fit content on their playlist. There are internal tools that tell them how much perfect fit content is on a playlist, because then that playlist basically pays out less as you listen to it. As you reported, there are mood and ambient playlist that are entirely perfect fit content, which means you're listening does not count as much to the royal level. multi-pool, as it would if it were other artists. And, you know, for some people might think, oh, yeah, yeah, okay, well, yeah, this is background music, whatever. And I think that that does definitely diminish people who make ambient music, people who make jazz and classical. You reported

on some artists who actually are asked to make jazz music for epidemic sounds, but it's not quite jazz music, right? Yeah. And I should say that, like, you know, Spotify does deny those internal pressure. They deny that they're trying to grow the percentage of stream share, but that does conflict with some of the things that I came across in my reporting, but I just figured I'd mention it. Yeah, interviewing musicians who have made music for these production companies was super interesting. You know, there's a quote in the book of one of them saying that definitely the directions that they were given were to play as simply as possible. The idea is that this is music that's supposed to kind of like play out in the background and not be too distracting.

No challenging harmony. No jazz. Yeah. Play jazz, but not real. Yeah, yeah, totally. And what was really interesting talking to the musicians who'd actually made this stuff is that these musicians would say things like, I just make the tracks and submit them, and I don't know what happens after that. And they get paid, obviously, but not know what happens after that. So there's this secrecy around it, both for the musicians who are participating, listeners, people who work in the music industry, part of what makes it so concerning, you know, like there's not really any transparency around it. And it's really created a situation. not just with PFC and that's how the spot. Perfect content. Yeah, that's how Spotify refers to it. But, you know, there also has reason to believe that other streaming services are likely engaging in similar practices as well. And we don't know how those companies refer to it internally. They might have their own names for these practices as well.

I think that it creates this situation where when people go to streaming services and they're being recommended music, sometimes that music is being shown to you because of some sort of commercial deal that's in place that is totally, not made clear to the listener at all. And it's not just this. It's also similar with like discovery mode, which is this, you know, program where artists are given the opportunity to accept a lower royalty rate in exchange for algorithmic promotion. And that program only influences music that is served on daily mixes, radio, and autoplay. But those are places that a lot of people go to to listen to music, especially the daily mixes and niche mixes and, you know, genre mixes and things like that.

And you said there's been sort of a move away from playlisting. You report that there has been a shifting of hiring resources away from editorial and towards programmatic music. Certainly, when TikTok took over mainstream consciousness in the pandemic, it really resorted priorities as to how we would discover things rather than curation, things that were for you. Can you talk about how Spotify shifted from the curated playlist era into this hyper-personalize AI-driven music feed that we now have? Yeah, for sure. And it seems like Spotify itself, it seems to be really doubling down on the fact that they still have human curators and like an

editorial team, but just, you know, through reporting, it's very clear that there is much more energy and like more resources that go into their efforts around personalization, which makes sense because if you listen to you, for example, one of the co-presidents of Spotify, Gustav Soderstrom, there's an interview that he did where he talks about how when it comes to Spotify, the AI, is the product. And what he means is that all of the data that Spotify collects on you, its knowledge of your listening history, its ability to make personalized recommendations on music or podcasts or audiobooks, like, you know, is sort of in this framing the products that they're selling to users. And they always double down on this that like the thing that sets them apart from other streaming services is that their users love

their personalization and they love like the playlists and things like you know a i dj or daylist or whatever but yeah it's super interesting thinking about like the influence that ticot might have had on that um trajectory and it's also super interesting thinking about you know like i know that you guys have covered in the past the way in which streaming has impacted the sound of pop music and it was pretty interesting going back and looking at this sort of like broader history of how you know We had this playlisting era where it seemed like the popularity of the Spotify playlist as sort of like a cultural force was sort of influencing the way in which people were thinking about songwriting and the way pop songwriters might have been thinking about playlists and skip rates and sort of different playlist categories and writing songs that maybe would be like adaptable across playlist categories. You have a great moment in your book where Spotify hosts a songwriting camp and ask people to basically write songs pitched to specific playlist. that then will end up not just on that playlist, but lots of a different editorial playlists.

It's interesting because at first I was really interested in writing about the Spotify playlist camps because it just seemed so telling that they had these songwriting camps that revolved around playlists. You know, like they'd have like a songwriting camp that they had for this playlist called the Most Beautiful Songs in the World, which seems really in line with the sort of Spotify sounds that you mentioned earlier. Uplifting, not too challenging. Yeah, yeah, just like kind of soft, you know, coffee shop music maybe or sort of muted, a little sad. That was interesting. And then it was also interesting thinking about not just these songwriting camps as a way of generating streamable content or playlist-friendly content, but also like thinking about it in the context of Spotify's like really oftentimes strange relationship with the songwriting community in general.

Right. And, you know, there's a part in the book contextualizing this as maybe an attempt to be like, yeah, look, we're creating resources for songwriters. Like, we don't hate you. We're definitely not suing you in court trying to, you know, reduce the royalty rates that we're paying songwriters. And at every turn, trying to make sure that, like, songwriting royalty rates don't go up. And, you know. Sounds like a songwriting camp as crisis management. Yeah, yeah, exactly. Yeah.

Like, it seems like it was sort of like a way in which to kind of try to smooth over the relationship with pop songwriters. and tracing this shifting moment where it was like the desire to land on Spotify playlist that was impacting maybe the way some songwriters were thinking about what they were doing into this next era where it really became more about like TikTok or short form video as this other sort of influence on it's never really like solely about what types of ways in which platform pressures are like causing people to change how they think about songwriting but also how platform pressures or shaping what type of music becomes popular. Right.

Like long intros don't work. Aside from Blinding Lights, which has like a 30-second long, amazing intro. But if you're Max Martin, you can write a very compelling instrumental intro. That aside, short intros, songs that just jump in right on the chorus, post-malone songs, things like that. But let's go back to sort of the algorithmic side of this. I hadn't realized until reading your book how much my homepage had changed. That now when I go to my homepage, the first thing I see right now is, I'm just looking at it right now.

Lola Young, Messy, listen on Spotify. That's an ad. The big banner that I'm looking at is an advertisement that the rights holder to that song is currently paying for. If I click on it, they'll have to pay Spotify. As I scroll down, I say music, podcasts, audiobooks. Audio books part of a ploy to pay less of the subscription fee to music holders, also to expand the offering to be generous to Spotify. I just expand the audio offering into audiobooks. And as then I scroll down further, daily mix, one, two, three, four, five, six, AIDJ, new releases for you, new music from our editors.

So editors, so there, audiobooks for you. And then finally, yeah, discovery picks for you. So a lot of customized material. And one of the big ones that has gotten a lot of attention recently is the day list. What is the day list? So day list is this playlist that sort of just like changes throughout the day. It's a personalized algorithmic playlist that the title has this sort of like word cloud of work. Like what is your say right now?

Mine is power ballad female pop afternoon. Yeah, you know, stuff like that. It's I think in some ways maybe capitalizing on how over the years it become kind of like popular on social media for people to sort of make jokes about Spotify's weird genre names around rap season. and maybe they just sort of like saw the way that people were reacting to that on social media. It also oftentimes, like as yours right now, have this sort of like gradient background of colors. Yeah, it's kind of like an aura. Yeah, like an aura reading, it sort of looks like. So I feel like there's some sort of capitalizing on Gen Z vibiness.

It's all lowercase letters, usually the title. It is all lowercase letters. So this is, I feel like this is the combination of the like discovery mode, stuff that is made for me paired with the like, mood functional music at the same time. Yeah. It's like contextualized. It's in the afternoon. It's like the power ballad. Like the gradient looks like I'm supposed to get inspired right now because I'm sleeping. It's at the end of the day and I need to like have my last burst of energy before the work day ends. Yeah. So one thing that I trace in the book is sort of the evolution of this long term goal that different music programming executives talked about at Spotify for years and years, which was this pursuit of a button where a user could open the app and basically just be met.

with the perfect recommendation and not have to make any choices and just, you know, you open the app, you press play. It's the perfect music at the perfect moment. And back to 2013, one of the early interviews that I could find with one of the co-founders of Tuna Go, which was the playlisting company that Spotify bought in 2013, talked about this as the goal. You know, there's a direct quote where he says, our goal is to just have a play button. And then in 2018, when Spotify went public on the stock exchange. Gustav Sordf from talked about this as how Spotify's goal was to eventually reach self-driving music as how you described it. You know, like you don't have to make any decisions. You just open the app and you'll get the perfect recommendation at the perfect moment.

And something that was super interesting was that around the same time, you know, maybe around two years ago, at the same time they unveiled Daylist and AIDJ. And in some ways, they both sort of fulfill this like long-term goal that the company seemed to have been after where you just just one button that you can press. But the way that in interviews, different Spotify executives had characterized it is that Gen Z is more comfortable with like the personalized playlist metaphor,

which is be DayList, whereas like older listeners want more of like a radio function. So the AIDJ in some ways sort of provide something really similar to DayList, which is just one button you click it. It's some music you can listen to. But AIDJ is these little slices of algorithmic, recommendations based on different aspects of your taste profile or like different playlist.

You know, it'll be like, here's some stuff from your release radar and the voice that is sort of introducing it is a generative AI voice trained on one of Spotify's employees. So it'll be like, you know, here's some songs from your release radar. Here's some songs you were really into you in 2021. Hey, you're an elder millennial. You've totally fallen out of the culture. You're really into something from 1996. Yeah, yeah, totally.

You know, like stuff like that. And then, you know, sometimes Spotify's like self-created genre names will pop up in AIDJ2, you know, where it'll be like, all right, next up, we're going to play some songs. We think you might like that fall under the category of metropopolis or something like that. And you're like, okay, that's definitely not real, but like, go ahead. And, you know, those genre terms also are like part of the component of the daylist, you know, recipe as well, it seems like. So, okay, here's the thing. In my day list, I really like the day list because sometimes I am lazy and I just like, I need the thing to listen to. Because I have both lean forward and lean back listening. I actually think there is really room for both kinds of listening. What's strange, though, sometimes it's really good at getting my like Sunday morning, right? But it's just like, I want my, it's not coffee shop music.

I want, like, Ethiopian jazz. It's like, I really like that on my Sunday morning. I don't know why. And it figured that out about me because I had maybe already selected it. And then it kind of just like give me more of the same thing that I already wanted. It's not actually always awesome at like suggesting like, you liked this last time, but here's something slightly different in the same vibe. What's weird though is that playlist almost feels like it starts to hallucinate like an AI where like it just like goes to some like lowest common denominator as it like, especially once the playlist maybe ends and it goes into like just radio mode. Eventually I'm in like bad piano versions of like Beatles covers all the sudden.

Wow. And I've found, it seems like, at least in my own listening, I have found that this kind of music, which sounds like perfect fit content, isn't just ending up on our mood playlists. It's also ending up in like just whatever genre you're into. That's super interesting. You know, from my reporting, I have like no reason to believe that there is any sort of like algorithmic boosting that is happening with this music. But I do think that any time tracks are added into first party playlist by editors that creates a situation where, those tracks are made visible to users. So then, you know, someone might be listening to their deep focus playlist. And I assume that, I mean, it seems a little bit hard to believe that this would happen, but like maybe it could. Some say, oh, this is a cool track.

Like, let me add this to my own deep focus playlist. And like, you know, it could happen. There's a feedback loop potentially. You know, like, so if some of these songs are being, you know, boosted in editorial playlist, you know, maybe a user hears it and likes it and copy some of it over to their own playlist. And then by nature of it being added to their own playlist, that could boost the likelihood that maybe it could end up in some sort of algorithmic playlist. Because the playlist that users make on Spotify, you know, that is part of the training data that algorithmic playlists are based on.

So what you're listening to and putting in a playlist is affecting what I'm listening to. If we're into the same kind of music. Yeah. In theory, yeah. You know, like when you make a playlist and put a title on it, like if I make a playlist and I call it, you know, Sunday morning chill music, you know, that could potentially impact how Spotify would then algorithmically recommend Sunday morning chill music to someone who maybe has like a similar tasted music to you. How I always describe it as like, you know, these platforms really rely on the lean back listeners who are going to be consumers of the day lists and the AI DJs. And I think that they hope that people will take advantage of or use these algorithmically generated personalized discovery tools that

they're making. But they also rely on people making their own playlist because that's a, you know, useful form of training data in order to make recommendations of the future. And I think the bigger theme here that it's not, it's not always clear what you're listening to. Like if you're listening to these discovery playlists, you're listening to music that likely some tracks on there are getting a lower royalty rate because artists are joining that program. And one of the challenges is that, according to some of your reporting, generated millions of dollars or saved millions of dollars in royalties. And so even if you aren't participating in that program as an artist, you're competing against other artists who are. And there is an internal incentive to have a bunch of perfect fit content in these playlists or discovery mode content in those playlists. And as listeners, we don't know which is a track that's being fed to us because it's something we like. It's something that has a special side deal. With the company, it's something that an artist has said to, yes, I would like to promote this more than other stuff. Or it's coming from someone else's playlist.

Like, this is all happening and it's not always clear what those signals are when it's saying this is music made for you. Yeah, totally. And the interesting thing about Discovery Mode is that it was pitched to artists as a way to have more input over the signals that go into algorithmic discovery. Like, you know, we take in X amount of data points every day to inform our personalized recommendations. And this is just giving artists the ability to, you know, add another data point to the mix. But if you pay us. Right. I was just trying to pull it up.

You report that 50% of Tier 2 to 3 artists, Spotify tiers, it's artists. Tier 2, 3 artists, which are a lot, have tried discovery mode. And then it has helped bring in a total of $61.4 million in gross profit. Yeah, that was over the course of a year, you know. That's in 20203. So it's a big program, and that's influencing what we're listening to. Okay. Spotify has had a lot of evolutions in its business. We haven't even covered all of them, right?

It begins as an ad marketplace. There's a whole moment where it thinks it's going to be an audio app marketplace. It had an editorial playlist phase or in an what do you call it, allegorial phase, the 4U algorithm phase. We've had podcasts, audiobooks, all these things. Given all these evolutions and where we are today, do you have any insight to where you think Spotify will be heading? That's a good question. I mean, it seems it seems, it seems, really obvious that, you know, what they've been in pursuit of for a really long time is just figuring out, like, how to reduce the cost of content. So now that has sort of, like, led them in the direction of an industry like podcasting, for example, where, you know, there's all of this audio content that they don't even need to pay royalties on.

People are listening to this podcast on Spotify right now, if they don't understand. Spotify does not pay when you, when you were hearing this show. Yeah, it's great for them. It's great for them. It's great for them if you're listening to the show on Spotify. So, you know, I think that's telling, yeah, it's kind of hard to, like, you know, make predictions about what might happen. What about on the listener side? Is this the year that we're going to finally see a backlash to this business model? Well, according to the Switched-on-Pop bingo card, it might be. Thanks for the plug.

I thought that was really funny. But it does seem from your reporting overall that there is more exposure to business practices that are materially affecting musicians every single day. and that are informing our listening, and they're not always clear. So I really appreciate you sharing Mood Machine with us. I think it is an essential read for anyone who's interested in how we listen to music, how it affects us, and if we're listening to the things that we think we're listening to. Thank you. Thanks for being here. Appreciate it.

Thanks for having you on. Switched on Pop is produced by Rihanna Cruz, edited by Art Chung, engineered by Brandon McFarlane. Our theme music is by Zach Tenario and Jossie Adams of Arc Iris, illustrations by Iris Gottel. we're a member of the Vox Media Podcast Network and a production of Vulture, which is part of New York Magazine. You can subscribe at NYMag.com slash pod. As a reminder, this year,

we are playing 2025 bingo. We've made 24 predictions that you can follow along throughout the year. You can get your own unique bingo card to play along by subscribing to our newsletter in our show notes

or going to our website Switchedonpop.com. We're also on social media at Switched on Pop. We'd love to hear what you think about this discussion. We'll be back again next week talking about the weekends.

final installment in his trilogy of albums. And until then, thanks for listening. Attention Spotify. Has given the new Good Girl Jasmine Absolute of Caroline Herrera. A fragrance intense with character gourmet and addictive. Imagine a jasmine emvolventy, caramelized, and tonka-tosted. A combination that seduce from the first instant and doesn't know away. Good Girl Jasmine Absolute, hypnotic, irresistible.

Discoveringlao-o-o-o-o-y-dove-involver for its essence. You know, Conviety You know, and you're and you're records of with the form

of the way of the market with the world. The Merelde The Miserion of the

world. The incredible system of Shopify facilita the on your website, in the

services, and in any place. That's music for your ears. No,

you're more your business is a super-exit with Shopify.

Empe your period of per at all more in shopify.com.s bar records.