The goals of art and commerce are basically opposite. Art fills our soul; it gives us emotional life. I have rarely heard anyone describe commerce similarly.
At its most ideal, the business of art enables more of it in greater variety, while allowing those who create it a reasonable living. This has rarely been the case. There are hundreds of years of unbelievably wealthy patrons building their cultural cachet by supporting artists of their particular taste and at their behest. More recently, recording contracts are routinely described as “brutal”, “a raw deal”, “predatory”, and “exploitative”. That has been generally true for all artists, but has been particularly pronounced for marginalized — and, to be more even specific, black — artists since the recording industry’s origins.
In “Mood Machine”, released earlier this year, Liz Pelly adds an additional complicating question: what is the relationship between art and commerce when massive data collection becomes routine?
The origins of streaming music may be found first in piracy and later in Rhapsody, but Spotify is where modern streaming platforms truly began. While Spotify’s founders tend to describe a noble birth, Perry points to a 2015 interview with co-founder Martin Lorentzon in which he describes the idea to build a targeted advertising platform first. How it would acquire users was an open question — “[s]hould it be product search? Should it be movies, [or ‘Godfather’], or audiobooks? And then we ended up with music”. That is not necessarily a bad thing. What is bad, though, is that Spotify reportedly began with an unlicensed library and made money on the back of it. That combination does not sound to me like the result of a love of music.
Sadly, the interesting storytelling does not reliably continue. Admittedly, part of the reason for this is my personal distaste for Pelly’s style of writing, something which I would not normally mention — surely not everyone is a fan of my writing style, either — but feel compelled to do so for how intrusive I found it. Far too many sections and chapters in this book end in a tedious turn of phrase: “Under the gaze of streaming surveillance, the exchange is never truly one-to-one;”; “It makes sense that as the digital world has grown to feel more like a shopping mall, it is also sometimes the very companies making music for shopping malls that are flooding its soundtrack”. Another part of the problem is the way this book is organized. Each chapter reads like an individual essay dedicated to a specific problem with Spotify — algorithmic suggestions, vibe-based playlists, and changing business terms, to name a few. What that looks like in practice is a great deal of repetition. I count at least seven chapters, of eighteen, dedicated to background and unfocused listening.
Part of the problem, however, is that Pelly has been documenting these phenomena for years in articles published at the Baffler. I am familiar with the extraordinary amount of “chill” music, trendy sound palettes, the relationship between mood-based music and targeted advertising, and the comparisons to Uber because these articles were all published a minimum of six years ago. That I remember these articles is sometimes a testament to Pelly’s reporting; at other times, it reminds me of things I previously found questionable but could not quite articulate why.
One thing I remember from one article, for example, is its attempt to define a “Spotify sound”. This was revisited in the book in the “Streambait Pop” chapter (page 82):
By the time of Spotify’s IPO in 2018, it seemed that the peak playlist era had produced an aesthetic of its own. That year, one pop songwriter and producer told me that [Billie] Eilish had become a type of poster child for what was being called a “Spotify sound,” a deluge of platform-optimized pop that was muted, mid-tempo, and melancholy. He told me it had become normal for him to go into a session and hear someone say they wanted to write a “Spotify song” to pitch around to artists: “I’ve definitely been in circumstances where people are saying, ‘Let’s make one of those sad girl Spotify songs.’ You give yourself a target,” he said. It was a formula. “It has this soft, emo-y, cutesy thing to it. These days it’s often really minimal and based around just a few simple elements in verses. Often a snap in the verses. And then the choruses sometimes employ vocal samples. It’s usually kind of emo in lyrical nature.”
Pelly’s argument is built primarily around the works of Charlotte Lawrence, Sasha Sloan, and Nina Nesbitt, none of which I am familiar with. But their music — “Normal” and “Psychopath” are both named in the article — sound like a lot of pop music trends of the time: a blend of genres that emerges kind of beach-clubby, pretty breathy, and electronics-heavy but not particularly danceable. Pelly quotes an indie rock label owner calling it “emotional wallpaper”.
In re-reading Pelly’s “streambait” article for this piece, I found this paragraph a good distillation of many arguments made throughout the book:
Musical trends produced in the streaming era are inherently connected to attention, whether it’s hard-and-fast attention-grabbing hooks, pop drops and chorus-loops engineered for the pleasure centers of our brains, or music that strategically requires no attention at all—the background music, the emotional wallpaper, the chill-pop-sad-vibe playlist fodder. These sounds and strategies all have streambait tricks embedded within them, whether they aim to wedge bits of a song into our skulls or just angle toward the inoffensive and mood-specific-enough to prevent users from clicking away. All of this caters to an economy of clicks and completions, where the most precious commodity is polarized human attention — either amped up or zoned out—and where success is determined, almost in advance, by data.
Much like the similar essays in “Mood Machine”, very little of this feels like it is directly traceable to Spotify or streaming generally. There has long been pop music that is earwormy, and pop music that is kind of a silence-filler. When radio was more influential, the former could be found on the contemporary hit radio station and the latter on any number of adult contemporary variants.
Coalescing around a particular sound is also not a Spotify phenomenon. The early 1990s brought an onslaught of Nirvana imitators, and the late 1990s polished the sound so hard it removed any trace of edge and intrigue it once held. The early-2000s dominance of Coldplay made way for the mid-2000s Timbaland production craze, which led to a wave of late-2000s dance and club pop, which was followed in the early-2010s by Americana revival. This is the power of radio. Or, it was the power of radio, at least. You could describe any of the chart-topping songs in similar terms as Pelly uses for “streambait”: “attention-grabbing hooks”, “chorus-loops engineered for the pleasure centers of our brains”, and “chill-pop-sad-vibe playlist fodder”. Should this be blamed on the precise listener analytics dashboard available to artists on Spotify? I am not sufficiently convinced.
Pelly describes a discussion she had with two teenagers outside an all-ages venue as they struggled to describe the “aesthetic rap” show they were attending, a genre which seems to be a slower and spacier take on rage (page 118):
The kids I spoke to outside Market seemed genuinely enthused. But as I headed home, I was struck by how palpably it seemed that most of those conversations were more concerned with a niche vibe fandom — which no one could even really explain — than the artists themselves.
I cannot imagine this is a new phenomenon. Some people develop a deep fascination with music and seek releases from specific artists. But plenty are only looking for a sound and a vibe. It is why retailers and magazines gave away sampler CDs in the mid-2000s scratching a generic indie rock itch.
What Pelly keeps describing in these chapters is a kind of commodification of cool, none of which is new or unique to Spotify. It is the story of popular culture writ large: things begin as cool for a small group, are absorbed into broader society, and are sold back to us by industry. This most often happens to marginalized communities who find community in vocabulary, music, visual art, and dance, and then it gets diluted as it becomes mainstreamed.
As noted, Pelly dedicates considerable space in the book to chill playlists — “‘Chilled Dance Hits,’ ‘Chilled R&B,’ and ‘Chilled Reggae’ were all among Spotify’s official playlist offerings, alongside collections like ‘Chillin’ on a Dirt Road,’ ‘lofi chill,’ and ‘Calm.'” (page 45). This is partly not the fault of Spotify; YouTube expanded the availability of live streams in 2013 and it resulted in plenty of samey chill hop stations. In a 2018 New York Times article about these nonstop live-streams, Jonah E. Bromwich writes:
Channels like College Music, ChilledCow, Chillhop Music and others are unlikely to have a broad impact on the music industry. But they represent an underground alternative to the streaming hegemony of Spotify and Apple Music. The industry commentator Bob Lefsetz said that while the stations were not likely to become a lucrative endeavor, they were a way for members of the public to seize power back from cultural gatekeepers.
Instead of being predominantly inspired or encouraged by Spotify, it is possible the growth of the background music genre is something the company is instead taking advantage of — and take advantage it did.
Whether on YouTube or Spotify, the beat-makers behind these tracks are loosely inspired by artists like J Dilla, but they have coalesced around a distinctly hazy and slowed-down instrumental palette. That these tracks are deliberately indistinct has led to Spotify commissioning low-royalty generic tracks and, as Pelly writes, this passivity is an area “where the imminent A.I. infiltration of music was most feasible: the replacement of lean-back mood music” (page 132).
All told, it sure sounds like Spotify aligned its recommendations to compel users into filling silence with music featureless enough it could be replicated by what are, in effect, stock tracks. If this was a deliberate strategy to allow Spotify to have lower royalty expenses, it has had a mixed effect. Setting aside the ongoing popularity of big pop stars like Taylor Swift and Justin Bieber — I would love to know how much of Spotify’s revenue is sent to those two artists alone — the rise of streaming also coincided with the explosion of in-your-face K-pop groups, renewed interest in rock music, and revivals of funk and disco. These are not the kinds of passive listening genres Pelly seems so concerned with.
That is not to say Spotify plays no influence in what is popular. Just as what was made popular on the radio brought a wave of imitators, so too is the case for an era where streaming is where most people listen to music. None of this is new. A streaming listener’s context is often quite similar to a radio listener’s, too. Pelly’s exploration of the chilled-out Spotify playlist and lean-back listener reads, to me, with considerable disdain. But that kind of passive listening was common in the radio days. People put music on in the car and at work all the time. My parents used to put a C.D. on when they were cooking dinner. I do not think they were captivated in that moment by the sounds of Genesis or the Police. I was listening to music while reading this book and writing this article. Sometimes, an album will get played as a background to other tasks; any musician is surely aware of this.
Perhaps you, too, are now seeing what I began to understand. What Pelly keeps gesturing at throughout this book — but never quite reaches — is that the problems Spotify faces are similar to that of any massive two-sided platform. Pick your case study of any of the large tech companies and you can find parallels in Spotify. It has hundreds of millions of subscribers; everything it does is at vast scale.
It has privacy problems. Pelly dedicates a chapter to “Streaming as Surveillance”, pointing to Spotify’s participation in the data broker and ad tech economy. Spotify suggests its ads can be targeted based on users’ moods correlated with playlist and song data. This, like so many other ad tech sales pitches, is likely an inflated claim with only limited success. Yet it is also a little creepy to consider it is what Spotify aspires its ad product to be.
Spotify, like many others, faces moderation problems. Spotify does not want to put too many limits on what music is accepted. In the best of circumstances, this makes it possible for a nascent artist to rise from obscurity and start a career. But there are financial incentives to gaming the system. There are people who will follow trends, and even commit outright fraud manually or with A.I.-generated material. This is true for other broadly available revenue machines — Google Ads and YouTube are two that immediately spring to mind. In an attempt to disincentivize these behaviours and reduce Spotify’s costs, the company announced in November 2023 it would stop paying royalties for tracks with fewer than one thousand annual streams. Pelly writes this “was part of a campaign waged by Universal Music Group to revamp streaming in its favor” (page 155). When Spotify rolled out this new royalty structure, UMG CEO Lucian Grainge bade good riddance to “merchants of garbage […] no one actually wants to listen to” (page 157). How much it actually hurts low-effort spammers is a good question, but it impacts legitimate indie artists — what Grainge calls “garbage” — for whom Spotify now presents no advantage over piracy.
I would not be so presumptuous as to say that is what this book ought to have been but, as a longtime reader of Pelly’s articles about the subject, I was frustrated by “Mood Machine”. It is the kind of book I wish would be taken apart and reassembled for better flow and a more coherent structure. Spotify and the streaming model have problems. “Mood Machine” identifies many of them. But the money quote — the one that cut through everything for me — is from an anonymous former Spotify employee (page 167):
“If Spotify is guilty of something, they had the opportunity at one point to change the way value was exhanged in the music industry, and they decided not to,” the former employee told me. “So it’s just upholding the way that things have always been.”
We treat art terribly. It is glamorous and alluring, but it is ultimately a plaything of the rich. The people who make money in art, no matter the discipline, are those responsible for its distribution, management, and sale. Those creating the actual work that enriches our lives too often get the scraps, unless they have enough cachet to dictate the terms in their favour.
Spotify is just another layer in the chain; another place that makes far more money than artists ever will. An artist understands they are signing up for a job with unpredictable pay, but Spotify’s particular structure makes it even more opaque.
The four biggest audio streaming platforms — Spotify, YouTube, Tencent, and Apple Music are the top-down force pushing culture in a particular direction, a level of influence Clear Channel’s executives could have only hoped for in its heyday. Streaming can help people learn about music they have never heard before. But it is not very effective as an artistic conduit. It is an ad-supported platform operating at global scale, much like a handful of other large tech companies, and it faces similar problems.
Yet none of them are as essentially tied to the distribution of art as is Spotify. It is a shame it did not create something to upend the status quo and made more artists more money. I guess part of the reason for that could be because its co-founders saw music as one of several interchangeable user acquisition strategies to sell advertisements — you know, for the love of art and music.