Mood Machine: The Rise of Spotify and the Costs of the Perfect Playlist
How Streaming Surveillance, Ghost Artists, and Algorithmic Payola Turned Music Into Mood Data While Impoverishing the Musicians Who Make It
Part 1: Chapter-by-Chapter Logical Mapping
Introduction: The Investigation Begins
Core Claim: Spotify’s rise represents not democratization of music but a sophisticated apparatus of extraction—transforming listening into surveillance, artistry into content creation, and culture into commodified mood data.
Supporting Evidence:
Author’s decade-long investigative reporting (2016-2025) from inside DIY music venues
Timeline: Spotify founded 2006 (Stockholm), launched Europe 2008, US 2011, IPO 2018
Market dominance: 615M users, 239M paid subscribers, 30% of streaming market, 84% of recorded music revenue
Corporate structure: Major labels (Universal, Sony, Warner) owned 18% at launch, controlled 70% of recorded music market
Logical Method: Investigative synthesis—connecting insider testimony from 100+ sources (former employees, musicians, label workers) with internal Slack messages, patent filings, and financial documents to expose gap between public narrative (”democratization”) and operational reality (extraction).
Logical Gaps:
Introduction asserts connections between playlist manipulation, ghost artists, and pay-to-play but doesn’t yet prove causal mechanisms
Claims “leveling the playing field” was false but hasn’t demonstrated the specific mechanisms of inequality
States listening became “mechanized” without defining what constitutes genuine listening versus data generation
Methodological Soundness: Strong foundation—author embedded in DIY scene for decade, multilingual research (Swedish sources), access to internal documents, cross-verification through multiple source types. Potential bias acknowledged through author’s stated “impulse toward demystification” and DIY background.
Chapter 1: The Bureau of Piracy
Core Claim: Spotify emerged not from mission to “save music” but as opportunistic advertising platform exploiting Sweden’s unique cultural/legal context around piracy—where file-sharing was normalized, copyright enforcement was weak, and politicians desperately wanted tech-solutionist answer.
Supporting Evidence:
2001 Gothenburg protests context: anti-globalization movement, police brutality → shift to localized activism
Piratbyrån (Bureau of Piracy) founded 2003 as counterweight to anti-piracy lobbying, spawned Pirate Bay
Swedish piracy prevalence: 1.2M of 9M citizens (13%) file-sharing by 2006 census
Political pressure: US threatened trade sanctions via WTO; Swedish police raided Pirate Bay May 2006
Spotify hiring: recruited uTorrent creator Ludwig Strigeus; beta version used Pirate Bay downloads
Logical Method: Historical contextualization—demonstrates how Spotify’s founders (Daniel Ek, Martin Lorentzen) positioned company to benefit from specific moment when Swedish establishment was desperate for “legitimate” alternative to piracy that major labels would accept.
Logical Gaps:
Chapter establishes correlation between piracy culture and Spotify’s emergence but doesn’t prove Ek/Lorentzen consciously exploited this versus simply responded to market conditions
Conflicting ideological positions within Piratbyrån (left activists vs. hackers vs. musicians) inadequately resolved—how did this internal contradiction shape outcomes?
Doesn’t address: if file-sharing was about “culture as public good,” why did independent musicians who supported it eventually embrace streaming’s corporate model?
Methodological Soundness: Strong—primary source interviews (Rasmus Fleischer, Peter Sunde, Dennis Lyxzén), corroboration with Swedish journalism, clear timeline. Weakness: relies on recollections about ideological positions 15+ years later.
Chapter 2: Saving the Music Industry
Core Claim: Spotify’s origin story (”existential search for meaning after selling AdVertigo”) is corporate mythology. Reality: two ad-tech entrepreneurs (Ek, Lorentzen) seeking traffic source for advertising product, with music chosen pragmatically (smaller files than video), not mission-driven.
Supporting Evidence:
Ek’s background: AdVertigo (ad targeting), SEO work, sold to TradeDoubler for $1.38M (March 2006)
Lorentzen’s background: TradeDoubler (automated banner ad sales), went public creating his initial millions
Corporate structure: Luxembourg tax haven registration, Cyprus holding companies
First US patent: described platform for “any kind of digital content”—not music-specific
Lorentzen quote: “Revenue source was ads... traffic source we were debating... we ended up with music”
Consultant hire: Fred Davis (Clive Davis’s son), entertainment lawyer advising on major label approach
Logical Method: Documentary deconstruction—juxtaposes public narrative (Ek’s “Ferrari/cabin/meditation” story) against timeline evidence (companies registered before claimed soul-searching, patent applications for generic content platform) and insider testimony.
Logical Gaps:
Timing question: If companies registered by late 2006 and domain purchased April 2006, does this definitively disprove the “soul-searching” narrative or just compress its timeline?
Chapter doesn’t explain why major labels eventually signed despite Spotify’s ad-focused model conflicting with their preference for per-stream payments
Missing: How did consultant Fred Davis specifically frame Spotify differently than failed predecessors (Pressplay, MusicNet)?
Methodological Soundness: Strong documentary evidence (patent filings, corporate registrations, Swedish journalism). Weakness: relies on inference about motivations from timeline rather than direct testimony about founding decisions.
Chapter 3: Selling Lean-Back Listening
Core Claim: Spotify’s 2012 “Music for Every Moment” strategy wasn’t artist-driven but emerged from marketing research identifying larger market in passive “lean-back” listening—fundamentally reshaping company from search-focused utility to behavior modification machine.
Supporting Evidence:
2012 market research: 40 participants keeping listening diaries revealed music as background experience
Strategic pivot: From “Google of Music” (search-focused) to mood/moment taxonomy
TuneGo acquisition (May 2013): Absorbed ~20 employees, thousands of playlists (Today’s Top Hits, Mood Booster, Your Favorite Coffeehouse)
“Syndicate” playlist hierarchy: flagship playlists → feeder playlists → “early bets” playlists, graduation based on data (skip rate, completion rate)
Sleep playlist success celebrated internally: “proved they’re not music company, they’re time filler for boredom”
Logical Method: Strategic archaeology—traces shift from initial product vision through market research findings to implementation, showing how advertising/marketing objectives (not musical ones) drove curation philosophy.
Logical Gaps:
Research methodology unexplained: 40 participants is tiny sample—how were they selected? What demographics?
Causal chain unclear: Did research reveal passive listening preference or did researchers/Spotify interpret data through lens of existing business model needs?
Chapter doesn’t address: Were users already passive listeners, or did Spotify’s interface design create passive listening behavior?
TuneGo founders’ vision for “single button” perfect playlist predated Spotify acquisition—who influenced whom?
Methodological Soundness: Relies heavily on single anonymous source “close to company” for research details. Multiple sources confirm TuneGo acquisition and syndicate strategy. Philosophical framework (Pauline Oliveros on deep listening vs. hearing) is useful but normative—imposes value judgment that “real listening” requires focused attention.
Chapter 4: The Conquest of Chill
Core Claim: “Chill” playlists weren’t organic user preference but manufactured category weaponizing mood-based organization to (1) target anxious millennials, (2) normalize background music as legitimate use case, (3) create fungible content where artists become interchangeable.
Supporting Evidence:
Historical precedent: Edison’s 1921 “Mood Music” pamphlet (27,000+ mood charts collected), Muzak’s 1940s “stimulus progression” (pseudoscientific productivity claims)
Spotify internal celebration of sleep playlist success as proof of business model
Former editor testimony: “Goal is reducing friction and cognitive work when opening app”
Lofi beats transformation: SoundCloud community discussing J Dilla → YouTube study streams → anonymous playlist fodder
Lofi Girl business model: labels take ownership, artists remain anonymous, 300K monthly listeners = 20 people at LA show
Logical Method: Genealogical analysis—traces lineage of functional/mood music through Edison → Muzak → streaming to show recurring pattern: pseudoscience → marketing gimmick → normalization of background music as legitimate product.
Logical Gaps:
Comparison to Muzak effective but incomplete: Muzak targeted workplaces (employers paid), Spotify targets individuals (users pay)—does this difference matter?
Brian Eno’s “Ambient One: Music for Airports” discussion doesn’t resolve whether ambient music can be both artistically valid AND functionally useful
Chapter doesn’t prove users were deceived about chill playlists versus simply choosing background music because they genuinely wanted it
Missing data: What percentage of Spotify listening is actually passive vs. active? Anonymous engineer claims “tiny percent” are lean-in listeners but provides no numbers.
Methodological Soundness: Strong on historical documentation. Weaker on proving causation (Spotify created passive listening vs. responded to existing preference). Relies on artist testimony about playlist recontextualization but doesn’t quantify scope—how many artists affected? What percentage of playlists?
Chapter 5: Ghost Artists for Hire
Core Claim: Spotify developed “Perfect Fit Content” (PFC)—secret program licensing stock music at reduced royalty rates to replace real artists on mood playlists, prioritizing profit margins over artist livelihoods while deceiving users about content origins.
Supporting Evidence:
Internal program name: “Perfect Fit Content” or PFC
Spotify definition: “Music commissioned to fit playlist/mood with improved margins”
2016 Music Business Worldwide first reporting; 2022 Swedish DN investigation proved ~20 songwriters behind 500+ artist names
Internal Slack reviewed (2023): 100+ playlists over 90% PFC; monitoring dashboard tracked “PFC %” for each playlist
Firefly Entertainment connection: founder Fredrik Holte childhood friend of Nick Homestein (Spotify Global Head of Music); both played in 90s band Apple Brown Betty
Strategic Programming team (10 employees) managed “strategic genres” (ambient, jazz, lofi, classical) vs. “editorial genres” (popular music)
Logical Method: Investigative exposure—combines insider testimony, internal communications, corporate registration data, and Swedish collection society (STIM) records to prove systematic program replacing artists with cheaper alternatives.
Logical Gaps:
Scale ambiguity: “Over 100 playlists” with 90%+ PFC = what percentage of total ~10,000 playlists? What share of overall streams?
Causation question: Did PFC cause streaming’s low artist payments or merely reflect business model already requiring cost reduction?
Chapter doesn’t address whether some PFC tracks are musically indistinguishable from “real” artists—if users genuinely can’t tell difference, does deception matter beyond principle?
Missing: Did major labels have similar programs? Chapter implies Spotify uniquely guilty but later mentions Epidemic Sound also on Apple Music, Amazon Music.
Methodological Soundness: Strongest evidence chapter—internal Slack messages, DN investigation corroborating with STIM records, multiple PFC musician testimonies. Weakness: no Spotify official response incorporated; relies entirely on critical sources.
Chapter 6: The Background Music Makers
Core Claim: PFC musicians aren’t “scammers” but precarious workers accepting exploitative buyout deals—revealing how streaming economy pressures musicians into devaluing their own labor through stock music production that strips away artistic agency and long-term earning potential.
Supporting Evidence:
Jazz musician testimony: one-year contract, anonymous tracks, 15 tracks/hour recording sessions, “play simpler” as primary feedback
Epidemic Sound business model: $1,700 flat buyout, owns master, 50-50 royalty split on reduced streaming rate, requires composers resign from PROs (Performance Rights Organizations)
Epidemic financial growth: 40M streams/day, $450M funding 2021 (Blackstone Growth, EQT Growth), $1.4B valuation
Advertising library musician: required to release under real name on pre-existing Spotify page, tracks based on Epidemic-curated playlists, “95% had very little to do with my artistic vision”
Ivors Academy UK opposition: “buying out composers’ luck,” racing to bottom, eliminates long-term performance royalty rights
Logical Method: Labor analysis—examines working conditions, compensation structures, and power dynamics to show PFC musicians as exploited workers rather than autonomous “fake artists,” contextualizing within broader gig economy exploitation.
Logical Gaps:
Chapter doesn’t resolve whether flat buyouts are inherently exploitative or only exploitative in context of reduced streaming rates + mandatory playlist placement
Missing comparison: How does $1,700 Epidemic buyout compare to traditional sync licensing fees? To label advances?
Doesn’t address: Could musicians negotiate better terms collectively, or does precarity make this impossible?
Ambiguity: Chapter says Epidemic tracks get “reduced streaming rate” but doesn’t specify whether this is Spotify’s doing or Epidemic’s negotiating position
Methodological Soundness: Strong firsthand testimony from PFC creators. Weakness: small sample (3 musicians detailed), may not represent full range of PFC working conditions. David Bowie quote about “never playing to the gallery” is rhetorically effective but normative—imposes judgment that artistic integrity requires rejecting commercial considerations.
Chapter 7: Streambait Pop
Core Claim: Streaming financial model (royalties only after 30 seconds, playlist-driven discovery, data optimization) systematically shaped pop music itself—producing “Spotify Core” aesthetic (whispery vocals, minimal production, chorus-first structure, emotional monotone) optimized for passive consumption and playlist compatibility.
Supporting Evidence:
Spotify for Artists “How to Read Your Data” video: explicitly encourages artists to “lean into” what algorithms surface
Post Malone “Rockstar” (2017): opens with chorus; label released 3.5-minute looped chorus version for stream optimization
Khalid “Location” example: playlist-adaptable (workout, chill, background), 9 tracks with 1B+ plays
Songwriter testimony: “Let’s make one of those sad girl Spotify songs” became normal studio session prompt
TikTok shift (2020): “Re-engagement triggers every few seconds,” snippet-based songwriting, A/B testing hooks before finishing songs
Warner Music Group Data Science: processing 4.5B streams/day to “forecast,” “build propensity models,” power algorithms determining future signings
Logical Method: Platform effects analysis—documents how technical affordances (30-second monetization threshold, skip rates, playlist sequencing requirements) created specific selection pressures on music production, analogous to how vinyl’s physical limitations shaped classical music (vibrato, dynamics).
Logical Gaps:
Correlation ≠ causation: Did Spotify cause Spotify Core sound, or did artists who naturally made playlist-friendly music simply succeed more on platform?
Chapter doesn’t address: How much is artists following data vs. younger generation genuinely preferring this aesthetic because they grew up with it?
Missing control: What happened to artists who didn’t optimize for playlists? Did they fail, or just succeed differently/elsewhere?
Temporal confusion: TikTok emerged 2020, but Billie Eilish/Khalid success was 2017-2018—which direction did influence flow?
Methodological Soundness: Strong on documenting what happened (platform features, music characteristics, artist testimony). Weaker on proving why it happened (multiple causal explanations possible). Jeremy Wade Morris’s “platform effects” concept borrowed from Mark Katz’s “phonograph effects” provides useful framework but doesn’t eliminate alternative explanations.
Chapter 8: Listen to Yourself
Core Claim: Algorithmic personalization (Discover Weekly, Daily Mix, etc.) reframed music discovery from connecting users to world of music → selling users their own data back to them, creating silos of self-referential listening optimized for session extension, not musical exploration.
Supporting Evidence:
The Echo Nest acquisition (2014): $49.7M for MIT-derived content analysis (acoustic metadata) + context analysis (cultural metadata from web crawling)
Discover Weekly launch (2015): 1.7B streams by year-end, shifted company direction toward “make everything Discover Weekly”
Personalization team growth: 30 US editorial employees vs. 700+ personalization employees by 2023
Success metrics: “extending listening session length,” “growing engagement,” “new user retention”—not musical diversity
Metadata categories: valence (0-1 scale, “happy” to “sad”), energy (0-1 scale, intensity), 287 mood terms from user playlist titles
2018 “consumption shifting” initiative: move users from search/library → homepage feed for “fully programmed surface”
Logical Method: Technical deconstruction—reverse-engineers algorithmic systems from patent applications, API documentation, employee testimony, and product feature analysis to expose how “perfect playlist” optimization served retention goals, not discovery goals.
Logical Gaps:
Chapter critiques algorithms for optimizing engagement but doesn’t prove engagement optimization necessarily produces worse recommendations—could high engagement reflect genuine satisfaction?
Missing: What’s the counterfactual? If Spotify didn’t personalize, would users discover more music or just stop using service?
Echo Nest acquisition assessment contradictory: one employee calls it “smartest acquisition in music business history,” another says “couldn’t integrate most interesting aspects, mostly PR move”—which is true?
Doesn’t resolve: Is collaborative filtering (finding users with similar taste) inherently problematic, or only when combined with engagement optimization?
Methodological Soundness: Strong technical documentation (patents, API specs, internal metrics). Philosophical framework (Nick Seaver’s “Computing Taste”) useful. Weakness: normative assumptions about “real” music discovery versus “selling taste back to you”—distinction may be less clear than chapter asserts. Former ML engineer’s critique that “reducing to data flattens 3D to 2D” is evocative but doesn’t specify what information is lost.
Chapter 9: Self-Driving Music
Core Claim: AI DJ, Daylist, and other “self-driving music” features represent endpoint of lean-back logic—eliminating user choice entirely under guise of personalization, contextualizing music solely in relation to user’s data profile rather than cultural/historical meaning.
Supporting Evidence:
AI DJ (2023): generative voice chains Spotify employee’s voice, announces “next up: songs that took over your life in 2022”—music as mirror of self
Daylist (2023): changes throughout day, uses “dayparting” (broadcast radio concept), creates hyper-specific titles like “Indie Tronica 2020s Late Night”
Glenn McDonald’s Every Noise At Once: 6,000+ microgenres, “helping musical knowledge self-organize”
Hyperpop case study: Spotify renamed “Neon Party” → “Hyperpop” (August 2019) after 100 gecs TikTok virality, claimed to “discover” scene that existed since 2015
Microgenre creation process: McDonald “made up” Escape Room name for data cluster, “watch and see if they turn into a thing”
Logical Method: Taxonomic critique—examines how Spotify’s classification systems (microgenres, mood tags, personalization algorithms) don’t neutrally describe music but actively reshape culture by determining what’s “real enough” to exist on platform.
Logical Gaps:
Hyperpop controversy shows real harm (trans/POC erasure, scene commodification) but doesn’t prove all microgenres equally harmful—some might genuinely help users discover music
Chapter doesn’t distinguish between: (a) descriptive taxonomies reflecting existing listening patterns vs. (b) prescriptive taxonomies creating new categories
Missing: What’s better alternative? How should music be organized if not by genre/mood/listening context?
Microgenre jokes (social media mockery of “escape room,” “metropopolis”) cited as evidence of harm but doesn’t prove users are actually confused versus amused
Methodological Soundness: Strong on documenting classification systems. Weaker on harm assessment—confuses aesthetic complaints with material impact. Hyperpop case study is compelling (Noah Simon documentary, artist testimony about erasure) but represents single example. Maria Eriksson quote (”you do not exist unless you are data”) is provocative but doesn’t prove Spotify caused this condition versus participating in broader digital culture shift.
Chapter 10: Fandom Is Data
Core Claim: Music fandom has been transformed into data processing labor—fans tagging, sorting, describing music in increasingly granular “vibes” and “aesthetics” that serve algorithmic legibility, surveillance capitalism, and metadata-as-service business models.
Supporting Evidence:
“Oddly Specific Playlists” Facebook group: 400K members, 800+ posts/month requesting songs for niche aesthetics
Aesthetic proliferation: Cottagecore, Dark Academia, Coastal Grandma—each becoming Spotify editorial playlists
Spotify’s “strategic programming” team tagging tracks for “strongly seeded candidate pools”
AI mood data startup at 2022 conference: uploads tracks → AI tags with “proprietary taxonomy of genres, moods, emotions”
Robin James theory: “vibes” = language algorithms use to perceive us; “pre-packaging yourself as data subject”
Logical Method: Labor theory application—reframes fan activity (playlist creation, music description, microgenre naming) as unpaid metadata work benefiting corporations through improved algorithmic efficiency and targeted advertising.
Logical Gaps:
Chapter doesn’t prove fans are unwitting laborers versus willing participants enjoying creative taxonomic play
Oddly Specific Playlists group described as “meme community” and “therapy space”—if users find it fulfilling, is it still exploitation?
Causation unclear: Did algorithmic culture create niche aesthetics obsession, or did existing internet culture (Tumblr, TikTok aesthetics) simply migrate to music?
Doesn’t address: Has music fandom always involved sorting/categorizing (record collector culture, crate digging, genre debates)? What’s genuinely new?
Methodological Soundness: Conceptually sophisticated (auto-surveillance from Fredric Jameson via Jacques Attali). Evidence is observational rather than systematic—author’s impressions of Facebook group, TikTok trends, Spotify playlists. Doesn’t quantify: How many users engage in microgenre/aesthetic tagging vs. passive consumption? Jacques Attali framework (music industry produces demand, not supply) is theoretically useful but unfalsifiable.
Chapter 11: Sounds for Self-Optimization
Core Claim: Generative AI music (Boomi, Endel, etc.) represents logical conclusion of streaming’s functional music paradigm—automated mood regulation positioned as creative empowerment while actually replacing musicians with algorithmic outputs optimized for passive consumption.
Supporting Evidence:
Boomi statistics: 14.5M songs (14% of world’s recorded music), banned for artificial streaming not AI generation
Endel business model: generative soundscapes using time/weather/biometric data; Warner Music Group partnership 2018; universal music group deal; $10M+ revenue 2022
Pseudoscientific validation: Arctop “brain-decoding” study (participants wore headbands); published in Frontiers (journal with retraction controversies)
SleepScore Labs partnership: joint venture with Dr. Oz (known for misinformation), ResMed, Pegasus Capital
UMG “Music + Health Summit” (September 2023): CEO Lucien Grainge promoting AI-accelerated wellness products
Spotify’s internal “Soundscape” product (2022): abandoned project for PFC-only endless ambient streams
Logical Method: Technological determinism critique—shows how each “innovation” (AI music generation, personalized soundscapes, mood optimization) extends rather than disrupts existing extraction logic, now with scientific veneer.
Logical Gaps:
Chapter conflates multiple AI categories: generative music (Boomi, Suno, Udio) vs. generative soundscapes (Endel) vs. remix tools—each has different implications
Doesn’t distinguish between AI trained on copyrighted works without permission (Suno, Udio—sued by majors 2024) vs. AI using licensed stems (Endel remixes)
Pseudoscience critique is compelling but doesn’t address: Even if Endel’s studies are methodologically weak, could functional music still work for some users pragmatically?
Missing: What’s scale of AI music infiltration? Chapter provides revenue numbers but not stream counts or playlist penetration rates
Methodological Soundness: Strong on documenting corporate partnerships and marketing claims. Weaker on proving actual harm—revenue growth for Epidemic/Endel doesn’t prove displacement of human musicians (markets could expand rather than replace). Critique of “neuromarketing” as pseudoscience is valid but somewhat tangential—even if brain-scanning studies are junk science, targeted advertising still works through other mechanisms.
Chapter 12: Streaming as Surveillance
Core Claim: Spotify’s data collection isn’t just for music recommendations but feeds surveillance capitalism apparatus—selling user data to advertisers/data brokers, developing emotion detection technologies, and normalizing invasive tracking under guise of personalization.
Supporting Evidence:
User data download contents: playlists, search queries, streaming history, “inferences” (market segments like “heartbreak playlist listeners,” “Campbell’s soup buyers US”)
67 companies on Spotify’s cookies vendor list using tracking technologies
Axiom partnership: one of world’s largest data brokers (2.5B people, 62 countries)
2018 patents: emotion detection from voice intonation/stress/rhythm; personality traits modeling (openness, conscientiousness, extroversion, agreeableness, neuroticism)
Advertising pitch: “Streaming intelligence,” “We know if you’re listening to chill playlist in morning, you may be doing yoga”
WPP partnership (2016, reaffirmed 2023): selling first-party mood data to global marketing firm
Neuromarketing partnership with NeuroInsight: participants wearing brain-scanning headbands
Logical Method: Surveillance capitalism analysis (Shoshana Zuboff framework)—documents data collection, third-party sharing, and secondary uses to prove Spotify participates in broader apparatus of behavioral prediction and control.
Logical Gaps:
Chapter doesn’t quantify: What percentage of collected data is actually used for recommendations vs. stored/sold? Patents ≠ implementation.
Emotion detection patents (2018, granted 2021) presented as threat but no evidence they’ve been deployed
“Inferences” file includes obvious errors (Verizon users when user isn’t Verizon customer)—if data is inaccurate, is surveillance still harmful?
Doesn’t prove Spotify uniquely invasive versus standard for digital platforms—comparison to Facebook/Google/Amazon would clarify whether this is Spotify-specific problem or systemic
Methodological Soundness: Strong documentation (patents, partnership announcements, GDPR case won by NOYB). Weaker on proving specific harms beyond general surveillance critique. Stefano Rosetti (NOYB lawyer) provides crucial context: “asymmetry of power” = not knowing what they know. Meredith Whitaker quote linking AI to surveillance advertising is accurate but doesn’t prove music streaming specifically enables unique harms.
Chapter 13: The First 0.0035 Is the Hardest
Core Claim: Pro-rata royalty system is deliberately opaque, mathematically complex, and systematically designed to benefit major labels while keeping artists (even successful ones) unable to understand or challenge their payments.
Supporting Evidence:
Penny fractions: ~$0.0035/stream (technically meaningless but symbolically accurate)
Rep. Rashida Tlaib calculation: 800K streams/month = $15/hour minimum wage equivalent
Pro-rata mechanics: labels paid percentage of total stream share, not per-stream rate
Opacity factors: NDAs prevent artists from seeing label-DSP contracts; promotional rates vary; per-stream vs. per-user minimums; free vs. paid tier differences
Damon Krukowski (Galaxie 500) 2012 example: 5,960 quarterly streams = $1.05 publishing + $9.18 recording = $3.17 per member
2023 UK Musicians Census: median income £20,700 ($26K), 44% under £14,000 ($18K), 50%+ sustained by non-music income
Logical Method: Economic deconstruction—breaks down royalty calculation to expose how complexity serves as “smoke screen” (Hunter Giles, Infinite Catalog) preventing artists from identifying where money goes.
Logical Gaps:
Chapter advocates user-centric model (your $10/month → artists you listen to) but doesn’t address major labels would never accept this (threatens their cross-subsidization from popular to developing artists)
Pro-rata described as “definitive flaw” but doesn’t explain why major labels negotiated for this system—what were their incentives?
Missing comparison: What do musicians earn from radio, sync licensing, live performance? Is streaming uniquely exploitative or just latest iteration of music industry’s traditional model?
2018 Princeton/MusiCares survey shows <25% of musicians earn from streaming, but 61% say music income insufficient—implies problem predates streaming
Methodological Soundness: Strong on documenting system complexity and artist confusion (UMAW testimony, UK Musicians Union census). Weaker on proving alternative systems would work better—user-centric model assumed superior without addressing label opposition or implementation challenges. UN WIPO 2021 report cited but not interrogated—does their pro-rata critique have limitations?
Chapter 14: An App for a Boss
Core Claim: Spotify for Artists (S4A) transforms musicians into customers buying promotional services while positioning platform as “boss” through data dashboards that encourage optimization, self-exploitation, and alignment with streaming-friendly aesthetics.
Supporting Evidence:
S4A mission: “Make them feel like they can grow” (employee quote)—feeling, not reality
App features: stats dashboard, playlist pitching tool, paid ads (Marquee, Showcase), Discovery Mode enrollment
Content strategy: 70%+ of featured artists in first 3 years were solo acts
Times Square billboards: “illusion of success,” costs hundreds of thousands, actual exposure “questionable”
Daniel Ek 2020 quote: Artists who “used to do well may not do well in future landscape where you can’t record music once every 3-4 years”—demanding continuous content generation
Internal tracking: S4A team segments artists into “archetypes” based on app behavior to sell them ads
Logical Method: Platform labor analysis—examines how “creator tools” actually function as disciplinary technologies enforcing platform-friendly behavior through combination of data visibility, norm-setting, and paid promotional gatekeeping.
Logical Gaps:
Chapter doesn’t address: Are tools like streaming analytics inherently harmful or only harmful in context of exploitative payment model?
Billboard assessment contradictory: label managers say “questionable exposure” but also mention fans recognizing artists on street—some value exists even if unquantifiable
Missing: What percentage of artists actually use paid promotional tools (Discovery Mode, Marquee, Showcase) vs. ignore them?
Doesn’t prove S4A causes artists to make streaming-friendly music versus simply rewards those who already do
Methodological Soundness: Strong on documenting S4A features and messaging. Weaker on proving causation—correlation between S4A content and solo artist focus doesn’t prove platform caused individualism versus reflected existing industry trend. “Creator economy” critique draws on Taylor Lorenz, Astra Taylor (solid sources) but doesn’t fully develop how musician experience differs from YouTuber/influencer experience beyond superficial similarities.
Chapter 15: Indie Vibes
Core Claim: “Independent” has been redefined from artist-controlled production/distribution → streaming-optimized solo entrepreneurs, with Spotify’s “indie” playlists dominated by major label acts and chill pop while actual independent labels are systematically deprioritized.
Supporting Evidence:
Lance Allen case study: instrumental guitarist, added to “Acoustic Concentration” → “Peaceful Guitar,” paid mortgage from Spotify, bought Subaru Outback, pitched by S4A as model independent artist
AWAL (Artists Without A Label) model: takes 15-30% of royalties for distribution + playlist pitching, sold to Sony for $430M (2022)
Spotify “indie” hub analysis (spring 2024): ~25% of “Front Page Indie” playlist was major label music, another 25% independent labels with major distribution
Pollen/Laura playlists: “lifestyle brand for youth,” “no genre just vibes,” editorial strategy targeting Gen Z consumer segments
Independent label testimony: “Their indie playlists have turned into pop playlists,” “Lord and Taylor Swift on indie playlists constantly”
POV Indie microgenre: appears to be pop music hyper-targeted to Gen Z users, not actual musical distinction
Logical Method: Definitional analysis—traces how “independent” transformed from production model (labels, distribution, values) to marketing category (vibe, aesthetic, consumer segment) serving platform’s need for granular audience segmentation.
Logical Gaps:
Chapter doesn’t prove major label presence on “indie” playlists is intentional manipulation versus result of genre ambiguity (what is “indie” in 2024?)
Sony buying AWAL presented as evidence of consolidation, but doesn’t address whether this improved/worsened terms for artists using service
Missing: Do independent labels want to be on Spotify’s indie playlists, or do they prefer alternative discovery mechanisms?
Lance Allen follow-up (December 2023 tweet about being replaced by Epidemic/Firefly) is poignant but represents single anecdote—how widespread is this displacement?
Methodological Soundness: Strong on documenting playlist contents and label testimony. Weaker on proving systematic bias—playlist analysis appears to be author’s manual review, not comprehensive data scraping. Independent label managers’ complaints about playlist access difficulties are credible but can’t distinguish between (a) deliberate exclusion vs. (b) shift toward algorithmic curation reducing all editorial placements.
Chapter 16: This Is Payola
Core Claim: Discovery Mode constitutes digital payola—artists accepting 30% royalty cuts for algorithmic promotion without disclosure to listeners, creating race-to-the-bottom where even ethical objectors must participate to remain competitive.
Supporting Evidence:
Discovery Mode launch (November 2020): 30% royalty reduction on enrolled tracks when surfaced through radio, autoplay, algorithmic mixes
Internal success metrics (May 2023): €61.4M gross profit over 12 months; €6.6M in May 2023 alone; 50%+ of tier 2-3 artists opted in
Top spenders: Believe (€1.8M), Merlin (€1.9M), indies (€1.6M), Warner (€0.6M)—notably NOT Universal or Sony
House Judiciary Committee letter (June 2021): warned of “race to the bottom,” demanded answers on safeguards, calculation transparency
Internal “Ethics Club” Slack: employees acknowledged “moves money from some artists to others, keeping more for ourselves”
No disclosure: tracks enrolled in Discovery Mode not labeled for users
Logical Method: Regulatory comparison—analogizes to 1950s radio payola (outlawed by FTC for deception + artificial popularity inflation) to argue Discovery Mode violates same principles: undisclosed pay-for-play warping perceived popularity.
Logical Gaps:
Chapter doesn’t address Discovery Mode vs. traditional radio promotion budgets—major labels spend millions on radio promotion teams; is 30% royalty cut cheaper/more expensive?
Internal Slack shows UMG had separate “Repertoire Discount Program” (RDP) for promoting catalog—suggests Discovery Mode not sole algorithmic promotional tool, but chapter doesn’t explain RDP mechanics
Missing: Did Discovery Mode work? Independent label managers report mixed results—some got more streams, some didn’t. What determines success?
Doesn’t prove Discovery Mode worsened overall artist payments versus redistributed within existing low-payment system
Methodological Soundness: Strongest evidence: internal Slack messages showing gross profit tracking, employee ethical concerns, spender breakdowns. Weakness: no Spotify rebuttal incorporated. Kevin Erickson (Future of Music Coalition) provides credible regulatory framework (FTC Section 5 could ban digital payola) but doesn’t address whether FTC has jurisdiction/will.
Chapter 17: The Lobbyists
Core Claim: Spotify spends millions on lobbying and political influence (€8.72M total 2015-2023) to shape data privacy laws, copyright policy, and platform regulation while executives extract billions, revealing whose interests streaming system actually serves.
Supporting Evidence:
Obama playlists (2015): “Welcome to Spotify, Mr. President”; eventual podcast deal (2019); job offer publicity stunt
Lobbying expenditure: $1.58M in 2023 (9th among internet companies); 40+ lobbyists over years; firms include Peck Madigan Jones (rebranded Tiber Creek Group)
Former government hires: Jonathan Prince (Clinton/Obama White Houses), Tom Manitose (Pelosi staffer), Dusty Jenkins (George W. Bush appointee, $860K base salary 2023)
Executive compensation: Ek $4B net worth, Lorentzen $7.7B, CFO Paul Vogel cashed out $9.377M day after laying off 1,500 employees
SEC filings: data privacy laws listed as “risk to business”—GDPR, CCPA could “impact ability to collect user information and provide personalized content”
Apple fight: complains about 30% App Store fee while charging artists 30% royalty cut for Discovery Mode
Logical Method: Follow the money—traces lobbying expenditures, executive compensation, and political connections to demonstrate whose interests are protected when Spotify lobbies against privacy laws and for weaker copyright enforcement.
Logical Gaps:
Chapter doesn’t prove lobbying was effective—politicians launched podcasts but did Spotify actually kill/weaken privacy legislation?
Comparison to Big Tech (Amazon $19.2M, Meta $19.3M lobbying) suggests Spotify is relatively small player—undermines claim about outsized influence
Executive compensation presented as evidence of exploitation but doesn’t compare to other $67B companies—is this typical or exceptional?
Lorentzen “I don’t want to die” / 120-year-old longevity quote is bizarre but irrelevant to structural analysis
Methodological Soundness: Strong documentation (OpenSecrets.org lobbying database, SEC filings, Billboard money makers list). Weakness: conflates lobbying expenditure with lobbying success—no evidence provided that Spotify’s €8.72M actually changed laws. Missing: What specific bills did Spotify oppose/support? Generic “data privacy laws” doesn’t identify particular legislation.
Chapter 18: The New Music Labor Movement
Core Claim: United Musicians and Allied Workers (UMAW) organized against streaming exploitation, culminating in Living Wage for Musicians Act (March 2024)—creating new royalty stream bypassing labels, paid directly to artists from streaming subscriber fees + 10% non-subscription revenue.
Supporting Evidence:
UMAW formation (April 2020): pandemic-era Zoom meetings, Downtown Boys members (Joey DeFrancesco—labor historian), grew from initial email thread
Justice at Spotify campaign (spring 2021): protests in 32 cities; demands included 1¢/stream, transparency, user-centric payments, stop fighting songwriters in court
Legislative path: Rep. Rashida Tlaib (Detroit) partnership, Andy Gaudiris (senior policy counsel) facilitation
Living Wage for Musicians Act provisions: new Artist Compensation Royalty Fund, nonprofit administrator, max payout cap (after 1M streams/month → back to pool), opens door for state/federal contributions
Legal precedents: Audio Home Recording Act 1992 (2% fee on recordable media), Digital Performance in Sound Recordings Act 1995 (direct payments from satellite radio)
Swedish engineers organizing: Spotify Workers Union formed, collective bargaining agreement fight (70% Swedish workers unionized vs. 10% US)
Logical Method: Organizing strategy documentation—chronicles transformation from decentralized complaints → coordinated campaign → legislative proposal, showing how collective action can challenge corporate power.
Logical Gaps:
Bill’s financial viability unexamined: New subscriber fee + 10% non-subscription revenue = how much total? Would this actually provide “living wage”?
Chapter doesn’t address major label opposition: Why would Universal/Sony/Warner allow direct-to-artist payments when current system lets them capture 70% of revenue?
Missing implementation details: Who would administer nonprofit? How would “artist” be defined legally? What prevents fraud?
Swedish engineers’ struggle for collective agreement presented but outcome unreported—did they win?
Methodological Soundness: Strong on documenting organizing process (author attended protests, interviewed organizers). Weakness: uncritical of Living Wage Act—presents as solution without interrogating whether it’s politically feasible or whether major labels would withdraw from streaming entirely rather than accept. Historical precedents (AFM strike 1942-1944, AHRA, DPRA) are useful but limited—those faced different political/economic contexts.
Bridge: Synthesizing the Logical Architecture
Pelly constructs her argument through concentric circles of extraction:
Layer 1: Founding Deception (Chapters 1-2) Spotify’s origin myth (”save music from piracy”) obscures founding reality: ad-tech entrepreneurs seeking traffic source, exploiting Sweden’s unique political moment (desperate politicians + weak labels + privacy laws preventing enforcement). The “better than piracy” framing becomes permanent justification for low payments.
Layer 2: Behavioral Engineering (Chapters 3-4)
“Music for Every Moment” strategy converts discovery platform → behavior modification machine. Key insight: 2012 market research identified larger market in passive listening, leading to mood/moment taxonomy optimizing for “lean-back” consumption. “Chill” becomes weaponized category—anxious millennials sold mood stabilization through playlist selection.
Layer 3: Content Manipulation (Chapters 5-6)
Perfect Fit Content program proves Spotify willing to systematically deceive users by replacing artists with cheaper stock music on mood playlists. Reveals what “democratization” actually means: anyone can participate in being exploited. Ghost musicians working under buyout contracts expose gig-economization of creative labor.
Layer 4: Cultural Flattening (Chapters 7-9)
Streaming creates self-replicating cycle: platforms reward playlist-friendly music → artists optimize for playlists → algorithms detect patterns → recommend more of same → culture homogenizes. “Streambait pop,” microgenres, and “self-driving music” represent progressive elimination of human curation and cultural context.
Layer 5: Surveillance Infrastructure (Chapters 10-12)
Fandom becomes data processing labor. Personalization isn’t service to users but mechanism for behavioral prediction and targeted advertising. Emotion detection patents, neuromarketing partnerships, data broker collaborations show music streaming as entry point for normalizing invasive surveillance.
Layer 6: Economic Exploitation (Chapters 13-14)
Pro-rata system’s deliberate opacity prevents artists from challenging payments. Spotify for Artists converts musicians into customers buying promotional services (Discovery Mode, Marquee, Showcase) while “creator economy” rhetoric masks gig-economization. Platform becomes boss, algorithms become disciplinary tools.
Layer 7: Market Consolidation (Chapters 15-16)
“Independent” redefined from artist-controlled production → streaming-optimized solo entrepreneurs. Major labels consolidate distribution (Sony buys AWAL), ghost music companies (Epidemic, Firefly), and even AI music generators (Warner + Boomi). Discovery Mode functions as payola, extracting payments from those with least power.
Layer 8: Political Capture (Chapter 17)
Lobbying expenditure ($8.72M) and executive compensation (Ek $4B, Lorentzen $7.7B) demonstrate system’s actual beneficiaries. Obama playlist partnerships, former White House staffers hired, podcast infrastructure built for politicians—all serving to prevent privacy regulation and maintain exploitative status quo.
Layer 9: Resistance & Alternatives (Chapter 18 + Conclusion)
UMAW organizing, Living Wage for Musicians Act, cooperative models (Catalytic Sound, Resonate), library streaming projects, public funding examples (Ireland, France, Norway) demonstrate alternatives exist—but require collective action, regulatory intervention, and reimagining digital infrastructure as public good.
The Argumentative Architecture’s Logical Coherence:
Pelly builds from foundational deception (origins) → technical implementation (playlists, algorithms) → cultural consequences (homogenization, surveillance) → economic extraction (royalties, paid tools) → political protection (lobbying) → organized resistance (labor movement, cooperatives, public alternatives).
Each chapter provides evidence for claims made in previous chapters:
Chapter 1-2’s founding story explains why extraction was baked in from start (ad-tech DNA)
Chapter 3-4’s lean-back strategy explains why PFC program (Chapters 5-6) targets those playlists
Chapter 7-9’s algorithmic systems explain how exploitation scales beyond human curation
Chapter 10-12’s surveillance apparatus explains what makes personalization profitable (selling data)
Chapter 13-16’s payment/promotion mechanisms explain who benefits (majors, executives, platform)
Chapter 17’s lobbying explains how system is protected from regulation
Chapter 18’s organizing shows what resistance looks like and why it’s necessary
The Central Logical Tension:
Pelly never fully resolves the determinism question: Does streaming technology inevitably lead to exploitation, or could similar technologies serve different values under different ownership/governance?
Her evidence proves Spotify as implemented is exploitative. But the book’s ending—advocating for cooperatives, library streaming, public funding, Living Wage Act—suggests technology itself isn’t the problem, governance/ownership is.
This creates productive ambiguity: The book is simultaneously technological critique (streaming affords certain behaviors) AND political economy critique (capitalism determines how technologies are deployed).
Part 2: Comprehensive Literary Review Essay
Opening: The Extraction Machine Disguised as Discovery Platform
You face a choice that isn’t a choice. Press play on Spotify and surrender to a feed of “personalized” recommendations algorithmically tuned to extend your listening session while harvesting behavioral data for advertisers. Or press play on Spotify and manually search for music, fighting against an interface redesigned to push you toward the “fully programmed surface” where the platform controls what you hear. Or don’t use Spotify at all and accept that 84% of recorded music revenue flows through streaming services you’ve now opted out of—making your boycott materially meaningless to the musicians you’re trying to support.
This is the trap Liz Pelly exposes in Mood Machine: The Rise of Spotify and the Costs of the Perfect Playlist. Over 400 pages of investigative journalism, she demolishes the decade-long mythology that streaming “democratized” music. What Spotify actually built was a sophisticated apparatus for converting listening into labor, culture into data, and art into mood-optimized content engineered for passive consumption. The extraction occurs at every level: from musicians forced to sell their royalties to private equity firms, to playlist editors coerced into replacing jazz artists with Swedish stock music producers, to users whose “chill vibes” selections train algorithms later sold to data brokers tagging them as “Campbell’s soup buyers.”
Pelly’s argument proceeds through concentric circles of complicity, starting from Spotify’s founding deception and spiraling outward to implicate major labels, venture capitalists, politicians, and the surveillance capitalism apparatus. But her most devastating insight appears midway through: what makes culture less interesting for listeners is also what makes it less sustainable for artists. The same lean-back listening environment that destroys musicians’ livelihoods also destroys the listening experience itself—reducing music to background utility, mood to preset categories, discovery to algorithmic regurgitation of your own data. When your competitor is silence, as Daniel Ek reportedly said, you’re no longer in the music business. You’re in the attention extraction business.
The Genesis Lie and Its Consequences
Pelly begins where all investigative work must: by interrogating the official story. Spotify’s founding mythology positions Daniel Ek as existentially adrift 23-year-old who sold his ad targeting company AdVertigo for $1.38 million, bought a Ferrari, descended into millionaire malaise, retreated to a cabin for meditation, and emerged with a mission to “save the music industry from piracy.” This narrative—repeated in The New Yorker, at Stanford talks, across tech media—serves crucial functions. It casts Spotify as mission-driven (not profit-driven), positions Ek as industry outsider (not establishment insider), and frames streaming as artist salvation (not corporate extraction scheme).
The documentary record tells a different story. Companies were registered in Cyprus tax havens and Luxembourg by late 2006. The domain was purchased in April 2006—meaning the “soul-searching” period, if it occurred at all, lasted weeks. Spotify’s first US patent application described a platform for “any kind of digital content such as music video, digital films, or images”—not a music-specific mission. Co-founder Martin Lorentzen, whose TradeDoubler fortune funded Spotify’s first two years, has stated plainly: “The revenue source was ads... the traffic source we were debating, should it be product search, should it be movies or audiobooks, and then we ended up with music.” CTO Andreas Ehn confirmed: “It wasn’t even clear back then that we were going to do music at all.”
This matters because the founding deception establishes a pattern that recurs throughout Spotify’s history: using the language of artistic salvation to obscure commercial extraction. When Ek hires entertainment lawyer Fred Davis (son of music mogul Clive Davis) to negotiate with major labels, he’s not disrupting the industry—he’s learning its rules. When Spotify’s beta version runs on music pirated from The Pirate Bay, it’s not ideological alignment with free culture—it’s cost savings until licensing deals close. When the company positions itself alternately with pirates (to attract Swedish engineers and users) and against pirates (to appease labels), it’s branding flexibility, not principle.
The major labels weren’t victims here. Universal, Sony, and Warner secured for themselves: 18% collective equity stake, $25 million advances (Sony’s first US deal, leaked 2015), $9 million in free advertising (which they could sell for cash without sharing with artists), guaranteed minimum per-stream and per-user payments, and “most favored nation” clauses ensuring terms at least as good as competitors got. As Pelly notes, “These were the people for whom streaming was made: major label execs, consultants, ad-men, and venture capitalists, all working to get their own share of the pie.”
Independent labels, negotiating through Merlin Network, got equity too (later cashed out and distributed to members). But here the divergence begins. While majors celebrated streaming as their “single biggest source of income” by 2010, independents were already noticing rate disparities—up to 6x difference reported by Swedish newspaper Dagens Nyheter. The Prorada system (paying percentage of total revenue based on usage share rather than per-stream rates) was designed for the major label catalog’s benefit. As one independent label manager told Pelly: “Looking back, that’s one of the bigger regrets of my career” not explaining to artists what they were accepting.
The Manufacture of Lean-Back Listening
The transformation from “Google of Music” (search-focused, 2009) to “Music for Every Moment” (mood-focused, 2012) represents Spotify’s most consequential strategic pivot. Pelly traces this shift not to user demand but to marketing research commissioned because Spotify “needed to grow, to get past the early adopters and reach a more mainstream audience.” Forty participants in major cities kept listening diaries. The revelation: “Active listening was a smaller part of the experience. There were way more listening hours using music as a background experience.”
This insight—that passive listening represented a larger market than active discovery—inverted Spotify’s business logic. If users wanted to “lean back and let Spotify choose things,” then success meant optimizing for extended session times, not musical diversity. The May 2013 acquisition of TuneGo, a Swedish playlist app, imported its mood/moment taxonomy wholesale: Today’s Top Hits, Mood Booster, Your Favorite Coffeehouse. TuneGo’s founders had backgrounds in pop songwriting (Nick Homestein) and hit prediction (Doug Ford, previously at “Hit Predictor” testing songs for commercial radio potential). Their north star: “a single button users could hit to get the perfect playlist based on real-time data.”
What Pelly documents next is a systematic program to normalize passive listening as legitimate—even superior—mode of music consumption. Sleep playlists became internal success metrics. A former employee recalls an all-hands meeting celebrating sleep playlist numbers: “They were very proud of this. It proved to them that they’re not a music company. They’re a time filler for boredom.” Another employee, present at a different meeting where Ek allegedly said “Our only competitor is silence,” reflects: “I definitely think people are afraid of silence, and Spotify has capitalized on that pretty well.”
The philosophical stakes are profound. Composer Pauline Oliveros spent her career teaching the distinction between hearing (involuntary) and listening (requiring consciousness). Streaming’s lean-back paradigm collapses this distinction, treating music as ambient utility. Pelly asks: “To what degree does this constitute listening?” If you’re streaming “Deep Focus” while writing emails, checking “Chill Vibes” boxes to soundtrack your commute, pressing play on “Peaceful Piano” to fall asleep—are you listening to music or tolerating its presence? And if a population pays so little conscious attention to music, why would they believe it deserves more than fractions of pennies per stream?
The lofi beats transformation crystallizes the broader pattern. In early 2010s, it was teenagers in SoundCloud chat rooms discussing J Dilla’s time-shifting drums, flipping samples on SP404s, creating community around technical experimentation. By late 2010s, after YouTube’s 24/7 study streams and Spotify’s algorithmic playlist empire took hold, lofi became “sad piano ballads with weird drums”—anonymous producers submitting tracks to curator-labels like Lofi Girl, which owns the recordings, controls the playlists (7M+ followers), and renders artists functionally invisible. One producer told Pelly: “There’s no such thing as a lofi artist anymore. There’s only such thing as a person who is part of the lofi machine.”
Graham Johnson (quickly, quickly) described his exit from the scene: After releasing music with actual singing, “the streams just completely plummeted, like kind of half. But it was weirdly freeing because I didn’t have to check my monthly listeners or wonder, did my song get put on the playlist?” The money was “kind of crazy,” he admits, but the cost was creative autonomy. This is streaming’s devil’s bargain: artists can pay rent from playlist placements while simultaneously destroying what made their music worth making.
Perfect Fit Content: The Ghost in the Machine
If lean-back listening created the market, Perfect Fit Content (PFC) exploited it. Pelly’s investigation into Spotify’s ghost artists program—combining leaked internal Slack messages, Swedish journalism (Dagens Nyheter exposé), and testimony from PFC musicians—exposes the most brazen manifestation of streaming’s exploitation logic.
The mechanics: Spotify licenses instrumental tracks from production music companies (Firefly Entertainment, Epidemic Sound, Hush Hush LLC, Cat Farm Music AB, Queen Street Content AB, Industrial Works/Mood Works, Mind Stream, Slumber Group LLC) at reduced royalty rates in exchange for guaranteed placement on mood playlists with millions of followers. These tracks appear under fabricated artist names (Ekvatt—”classically trained Icelandic beatmaker” with completely invented biography) on editorial playlists for ambient, jazz, classical, lofi. By 2023, internal communications showed 100+ official playlists over 90% PFC, with dedicated “Strategic Programming” team (10 employees) managing the program.
The financial scale: Between May 2022-May 2023, PFC generated €61.4 million in gross profit. In May 2023 alone: €6.6 million. The playlist monitoring tool showed each playlist’s “PFC %” in real time, with editors encouraged—then pressured—to increase the percentage. One former editor: “Initially they would give us links like, oh, it’s no pressure, but if you can, that would be great. But then that column came up. And after that... it became more aggressive.”
Pelly interviews three PFC musicians, and their testimony reveals not scammers but precarious workers. The jazz musician in Brooklyn signs one-year contract, records 15 tracks/hour in single takes, gets paid upfront fee but production company owns master and most publishing rights. The creative process: “They send links to target playlists as reference points. You write charts for new songs that could stream well alongside ones already on the reference playlists. Honestly, for most of this stuff, I just write out charts lying on my back on the couch.” Primary feedback: “Play simpler.” Goal: “Be as milk-toast as possible.”
Another musician, making ambient tracks for different PFC licensor, produced tracks under aliases, made couple thousand dollars, then watched one track get millions of streams—generating far more revenue for Spotify and ghost label than he’d ever see. “I’m selling my intellectual property for essentially peanuts. Whoever can get you generating that amount of plays, they hold the power. If that entity also owns your master or owns the streaming service or owns the means of distribution, that’s some antitrust levels of collusion.”
The program’s racial dimension: When stock music started replacing tracks on playlists historically dominated by Black and brown jazz/lofi artists, one source noted: “Spots for Black and brown artists making this music started getting cut down to make room for a few white Swedish guys in a studio.” The connection between Nick Homestein (Spotify’s Global Head of Music, TuneGo founder) and Fredrik Holte (Firefly Entertainment founder)—childhood friends from same Swedish town, played together in 90s power-pop band Apple Brown Betty—makes the nepotism explicit.
Pelly’s most damning evidence: internal justifications for PFC centered on “low supply” of ambient/jazz/lofi music. This is transparently false—these genres have no shortage on Spotify. What they mean is low supply of cheap ambient/jazz/lofi willing to accept reduced royalties. One source: “When you’re a DSP and you have that much power and influence over people’s education about music, it’s such a great responsibility. If I have a kid and I’m trying to teach them about the history of ambient music and go to Spotify, more often than not, what you’ll find is PFC artists.”
The Algorithmic Transformation of Listening
The Echo Nest acquisition (2014, $49.7M) marks Spotify’s full commitment to algorithmic personalization. The MIT-spinoff company brought “cultural metadata” (scraped from blogs, press, user playlists) and “acoustic metadata” (automated audio analysis: key, tempo, valence, energy, danceability). But as Pelly demonstrates through former engineers’ testimony, the goal wasn’t understanding music—it was understanding users.
Discover Weekly (2015) proved the model: 1.7 billion streams by year-end, described as “your best friend making you a mixtape every week.” But the algorithmic “friend” had specific priorities. Success metrics included “extending listening session length,” “growing engagement,” “new user retention”—never “expanding musical horizons” or “discovering challenging art.” One former ML engineer told Pelly why he left: “I became pretty disillusioned with this myopic focus on just the amount of minutes listened. Some of the records I would consider really life-changing wouldn’t even show in my top hundred because they’re really challenging records.”
The personalization explosion followed: Daily Mix (2016), Release Radar (2016), Wrapped (2016), algotorial playlists (2017), redesigned ultra-personalized homepage (2018), Smart Shuffle (2023), Daylist (2023), AI DJ (2023). Each iteration reduced user agency further. The 2018 “consumption shifting” initiative explicitly aimed to “move people out of other places of the app and into the home page, so there is a single place where you would listen unless you knew exactly what you were looking for.”
But what are you actually hearing in these personalized feeds? Not music contextualized by history, genre, or cultural meaning—but music contextualized by you. The AI DJ announces: “Next up, some songs that took over your life in 2022.” Daylist creates titles like “Optimistic Cinematic Sunday Afternoon.” Algotorial playlists are named “my life is a movie” (tagline: “every main character needs a soundtrack”). The Nigerian Spotify commercial makes the solipsism literal: a woman opens Spotify, clicks “Your Daily Mix,” and everywhere she goes she encounters clones of herself—selling juice, cutting hair, riding the van. “Playlists made just for you.”
Music journalist and scholar Robin James, quoted by Pelly, theorizes this as “pre-packaging yourself as a data subject”—fans learning to describe music in algorithmically legible terms (vibes, aesthetics, moods) so systems can better target them. The “Oddly Specific Playlists” Facebook group (400K members) exemplifies this: users request songs for “jellyfish would listen to,” “rotting and decomposing,” “drunken stumbling tragicomic clown”—hyperspecific aesthetic categories that function as metadata work benefiting platforms.
Former engineer testimony clarifies what’s lost in this self-referential loop: “It’s like taking a three-dimensional picture and flattening it to two dimensions. To say your tastes are really represented by a list of things you’ve listened to, almost anyone would say that’s not exclusively true. It’s decontextualized.” You don’t learn about music’s relationship to history, culture, or other listeners. You learn about algorithms’ interpretation of your own past behavior—then have that interpretation sold back to you in infinitely subdivided boxes.
The Glenn McDonald microgenre project exposes the absurdity. Over 6,000 genre labels on the Every Noise At Once map, from “metropopolis” to “escape room” to “POV indie.” McDonald’s process: identify data cluster of listening patterns, name it himself, “watch and see if it turns into a thing.” This is taxonomy as speculation—observing correlations in user behavior, assigning labels, then watching whether the labels stick. But what makes something “real enough” to qualify as a genre? McDonald’s answer: when enough data accumulates on Spotify. Maria Eriksson’s observation: “We seem to be entering a point in time in which you do not exist unless you are data.”
The Hyperpop case study demonstrates the violence of this process. A sprawling internet scene (SoundCloud, Discord, SPF420 digital venue) since early 2010s—sonically varied, geographically dispersed, united more by creative impulse than sound. Influenced by vaporwave, nightcore, J-pop, Y2K pop, dance music—especially significant was its queer/trans alternative to mainstream electronic culture. Then Spotify rebranded its “Neon Party” playlist to “Hyperpop” (August 2019) following 100 gecs’ TikTok virality. New York Times ran headline: “How Hyperpop, a Small Spotify Playlist, Grew Into a Big Deal.”
Artists responded with fury. Noah Simon’s four-part YouTube documentary “Hyperpop Origins” opens with direct debunking: “Hyperpop was not invented by Spotify in 2019.” Artist Quinn told journalist Kieran Press-Reynolds: “The creation of the Spotify Hyperpop playlist and the invitation of labels led directly to the erasure of trans influence.” Producer Omniboi: “Why does it feel like we were all erased?” Umru, NYC-based producer: “It made it so easy to be a fan of the sound without being really interested in the community or the specific artists.” The playlist became the brand; artists became interchangeable content.
The Economics of Exploitation: Pro-Rata’s Deliberate Opacity
Pelly dedicates Chapter 13 to deconstructing the royalty calculation—and her central point is the opacity itself is a feature, not bug. The “approximately $0.0035 per stream” figure is technically meaningless (Spotify pays percentage of revenue share, not per-stream rates) but spiritually accurate. It communicates what matters: musicians make basically nothing.
The actual calculation: Net revenue (subscriber fees + ad revenue - taxes/fees) × 52% (average recording royalty pool share) × artist’s percentage of total platform streams that month = label’s payout. Then the label takes its cut per artist’s contract. At every stage, complexity obscures value extraction:
Different plans (free, student, duo, family) generate different revenue → different royalty pool sizes
Different countries generate different revenue
Major labels negotiated guaranteed minimums per user other rights holders can’t command
Promotional rates (Discovery Mode’s 30% cut) reduce payments further
NDAs prevent artists from seeing label-DSP contracts
“Pay-all-alike” practices mean accepting reduced rates on behalf of artists
The result: Even successful musicians can’t trace where their money goes. Hunter Giles (Infinite Catalog, royalty accounting for indie labels): “You lose so many people. It’s not made clear by anybody.” The UK Music Managers Forum’s 2020 survey of 50 artist managers working across majors and 100+ independents confirms even insiders struggle to understand calculations due to NDA curtain.
The material reality: Damon Krukowski (Galaxie 500) received $9.18 recording royalty for 5,960 quarterly streams (2012). Split three ways: $3.06 per member. 2023 UK Musicians Census: median annual music income £20,700 ($26K), but nearly half earn under £14,000 ($18K), and 50%+ sustained by non-music work. Princeton/MusiCares 2018 survey: median musician $20-25K/year total (music + non-music jobs); 61% say music income insufficient for living expenses.
Congresswoman Rashida Tlaib’s estimate: 800,000 streams/month = $15/hour minimum wage. Spotify’s internal artist tier system provides the company’s own assessment: Tier 3 (where artists “start to make a living”) ranges $5-49K annually, averaging $13,500. Spotify publicly claims ~200,000 “professional or professionally aspiring” artists made $10K+ in 2022, extrapolating they “must have made $40,000 total from recordings” across all platforms.
But these calculations don’t account for: bands splitting payments, collaborators/producers taking percentages, label cuts, distributor fees, manager commissions. A solo self-releasing artist making $40K is viable. A four-piece band on an independent label splitting $40K after label’s 50% and manager’s 15% is making $4,250 per member before taxes. This isn’t a living wage—it’s supplemental income requiring day jobs.
The pro-rata system’s defenders argue it’s meritocratic: most-played music earns most money. But as UN World Intellectual Property Organization’s 2021 report notes, this means “major label superstars tend to derive the bulk of revenue from streaming platforms”—because the system was designed that way. Not all music is meant for infinite replay. Taja Cheek (L’Rain) distinguishes between music requiring “complex meditative relationship” versus “narcotic relationship.” Darius Van Arman (Secretly Group) reflects on releasing challenging bands like Oneida in pre-streaming era: “The label could sell a few thousand CDs, cover costs, have $25K left to split with the band. People would buy CDs at shows, or because they read a good review, and even if they only listened once and put it on a shelf, it had value.”
Streaming eliminates that model. Van Arman: “It’s based on what gets repeat listens. It’s not sustainable to put out challenging records. To be sustainable, you have to put out records that get repeat listens in coffee shops.” Asked if a label like his could start today by releasing bands like Oneida: “I don’t know. Probably not.”
Discovery Mode and the Institutionalization of Payola
The November 2020 introduction of Discovery Mode represents Spotify’s most brazen monetization of musicians’ desperation. Artists accept 30% royalty cuts on enrolled tracks in exchange for “algorithmic promotion” through Radio, Autoplay, and personalized mixes (Daily Mix, Artist Mix, mood/genre/decade mixes). Spotify framed this as “democratizing”—no upfront cash required, so “no barrier to entry.”
The financial reality: May 2022-May 2023, Discovery Mode generated €61.4 million gross profit. May 2023 alone: €6.6 million. Top spenders: Believe (€1.8M), Merlin (€1.9M), indies (€1.6M), Warner (€0.6M). Glaringly absent: Universal and Sony—the two biggest majors have different promotional tools (internal Slack reveals UMG’s “Repertoire Discount Program”). By 2023, over 50% of tier 2-3 artists had enrolled.
Employees in internal “Ethics Club” Slack channel immediately recognized this as payola: “Controlling more of the listening experience via programming pretty clearly benefits us more than anybody else.” “The boosted plays come at the cost of other artists.” “Discovery Mode moves money around from some artists to other artists keeping more for ourselves in the process. Artists as a whole get less, we get more.”
The House Judiciary Committee agreed, sending June 2021 letter warning of “race to the bottom where artists feel required to accept lower royalties to be heard.” United Musicians and Allied Workers called it “exploitative, unfair, money grab.” Artist Rights Alliance warned “biggest losers would be working artists, independent labels, and music fans looking to expand/diversify listening.”
But independent labels are caught in bind. One manager: “I don’t like the idea of Discovery Mode, but unfortunately with streaming we’re starting to get pushed up against this wall. It was a useful tool for us. All our artists running through the program did get tons of new listeners, new followers.” The vicious cycle: “We’re losing revenue because of streaming. Then streaming is pushing these weird opportunities on us. We need to take them to make up for lost revenue.”
The deception extends beyond royalty cuts to algorithmic manipulation itself. Discovery Mode works by “re-ranking” tracks in candidate pools before delivery to users—meaning a track enrolled has higher probability of surfacing. But there’s no disclosure. When a listener hears a track on their personalized “Chill Morning Mix,” they have no way to know whether it appeared because algorithms determined they’d like it or because rights holder paid for placement. As one independent label manager realized: “We did have someone reach out like, wow, I’m really hearing [insert song] a lot on Radio. Did you pay for this? And I’m like, if it came from Radio, I guess maybe we did.”
Future of Music Coalition director Kevin Erickson frames this as antitrust issue: “Payola has become integrated into the business model. It’s explicitly a means of driving artists’ compensation down. The FTC has enforcement and investigative tools. Under FTC Act Section 5, a rule-making banning digital payola on any service would be a massive win for musicians.” But as of Pelly’s writing, no regulatory action has been taken.
By late 2023, Spotify introduced “Pre-Campaign Insights”—an algorithm telling artists which tracks are most likely to succeed in Discovery Mode campaigns. This completes the circle: platform determines what music succeeds → offers tool for artists to pay for success → algorithm tells artists which music is worth paying to promote → artists optimize for algorithm → platform determines what music succeeds.
The Playlist Voice and the Flattening of Independent Music
Chapter 15 documents how “independent” transformed from production model (labels, values, artist control) to marketing category (vibe, aesthetic, Spotify-designated consumer segment). The evidence is Spotify’s own “indie” hub: Front Page Indie playlist, spring 2024, contained ~25% major label music, ~25% independent labels with major distribution. The flagship “no genre just vibes” playlists (Pollen, Laura) occupy visual real estate, each a “lifestyle brand” targeting Gen Z consumers.
Laura’s origin: previously “Left of Center,” rebranded as “lorem ipsum” dummy text joke—music so new it doesn’t have name yet. One strategist: “Spotify wanted to create a playlist lifestyle brand for the youth.” The curators form “Indie Global Curation Group,” but “indie” here means aesthetic (bedroom pop, whispery vocals, playlist-adaptable tracks) not production model. Metadata for “POV Indie” reveals it’s often categorized alongside Indie Pop, Pixel, Modern Rock, Bedroom Pop, Alt-Z, Weirdcore—terms that describe target listener, not musical characteristics.
Independent label managers describe the material consequences. One, 2022: “Their indie playlists have turned into pop playlists. Most indie labels are aware and have been really upset. Our distributor is constantly trying to talk to Spotify about it. They’ve pushed what we think of as indie to rock category, which has much smaller listenership. Even when we do get playlisted now, it doesn’t translate into many streams or much revenue.”
Another, operating influential indie rock/folk imprint: “Anything that could be put on coffee shop playlist streams better. Our streaming revenue heavily depends on editorial playlist placements. We used to be in someone’s good graces there, but we went through a two-year dry spell of barely getting any placements, and our revenue has been cut in half.” The culprit, per their distributor: indie editor was replaced by someone “who didn’t like guitar music and instead filled so-called indie playlists with chill, inoffensive Spotify music.”
The artists caught in this system described bending to platform pressures. “They’re getting on playlists like ‘driving at night,’ ‘chilling with your friends,’ ‘coffee and tea,’ ‘surf rock sunshine’—the most inoffensive music you’ve ever heard. It’s really easily digestible, doesn’t ask much of listener.” The optimization temptation: “I don’t know a single artist who listens to ‘hanging out and relaxing’ playlist, but they’re all looking at Spotify for Artists dashboard thinking, oh wow, this is the song that got playlisted, this song performed well. Maybe I should lean into that. It could happen to any artist. It’s almost just human to think, well, this thing is kind of successful. Maybe I’ll try it.”
The transformation of AWAL (Artists Without A Label) from boutique distributor → major label acquisition target reveals the endpoint. Founded on letting artists retain copyright ownership, AWAL offers three tiers: Core (15% of royalties, self-serve distribution), Plus (30% of royalties, dedicated rep + playlist pitching), Recordings (full label deal). Sony bought AWAL for $430 million in 2022. Penny Fractions newsletter: “AWAL is the minor league training ground for Sony.”
Universal struck similar deals with DistroKid (data access for “upstreaming” program identifying signings). Warner partnered with Boomi (AI music generator). The majors were “hellbent on growing market share, acquiring portions of DIY and independent distribution sector.” Simon Wheeler (Beggars Group): “In today’s world, when people say indie, it can mean anything that’s not three companies. You’ve got all the long-tail creators, AI companies, you name it. And apparently we’re all the same. We’re all just indie.”
The Living Wage for Musicians Act and the Limits of Reform
Chapter 18’s documentation of UMAW organizing provides the book’s most hopeful material—but also its most sobering reality check about scale of change required. The pathway from pandemic-era Zoom meetings (April 2020) to introduced federal legislation (March 2024) required: forming working groups, studying historical precedents (AFM’s 1942-1944 recording strike, Audio Home Recording Act 1992, Digital Performance in Sound Recordings Act 1995), assembling pro-bono legal team (Rohan Grey, Henderson Cole, Harvard Cyber Law Clinic), building relationship with Rep. Rashida Tlaib’s office through volunteer work, year of research meetings (one hour/week), draft → redraft → redraft.
The Act’s core provisions:
Artist Compensation Royalty Fund: new royalty stream paid directly from platforms to artists through nonprofit administrator
Funding: new fee on streaming subscribers + 10% of platforms’ non-subscription revenue
Distribution: paid to artists according to stream counts
Cap: maximum payout per track/month; after 1M streams, money returns to pool for other artists
Public funding: leaves door open for state/federal contributions
The legal basis is sound—two 1990s precedents prove creating new royalties that bypass existing contracts is possible. But Pelly includes crucial caveat via UMAW member Michael Abbey: “A penny per stream or this new royalty isn’t the horizon for our imagination. There was talk of maybe having more radical stuff in the bill. At least from my perspective, I thought it was important to find some emergency fix now because conditions are so extreme.”
This is the concession at organizing’s heart: the Living Wage Act isn’t the solution—it’s a band-aid. It doesn’t challenge pro-rata, doesn’t eliminate Discovery Mode, doesn’t break up major label oligopoly, doesn’t address surveillance capitalism, doesn’t reimagine digital infrastructure as public good. It adds a new royalty stream while leaving extractive architecture intact.
Why accept such limited ambition? Political realism. Rep. Tlaib senior policy counsel Andy Gaudiris: “To us, this is kind of the bare minimum. This is basic common decency and basic economic justice.” Passing even this bare minimum requires: sustained organizing, politician who actually cares about musicians (Tlaib represents Detroit/Motown), pro-bono legal work, media attention, and luck. The alternative—comprehensive restructuring of digital music economy—is politically unviable in current moment.
Alternatives: Cooperatives, Libraries, Public Funding
Pelly’s conclusion pivots from critique to construction, surveying alternative models:
Catalytic Sound (creative music cooperative, 30 members): Combines streaming + direct music sales. 50-50 split after expenses: half to infrastructure/editorial, half to artists divided evenly 30 ways regardless of stream counts. Monthly rotating selection (few hundred albums, mostly exclusive releases). Saxophonist Ken Vandermark: “We made decision there’s a limit at 30 musicians. But what we want to do is say Catalytic is one model for musicians’ collective. It’s not the one. The idea is to motivate other communities.”
Bassist Luke Stewart (Irreversible Entanglements): “We’re all running it together, which is testament to collectivity we think is necessary. The music industry pushes machine of celebrity, where it’s all about one person. Catalytic is all of our thing, and it truly operates in that fashion.”
Library Streaming Projects: Dozens of public libraries (Iowa City, Seattle, Austin, Pittsburgh, Minneapolis, Ann Arbor, Edmonton) have launched local music collections. Model: musicians submit recordings during open calls; community curators (library staff + scene-embedded locals) select 40-50 albums; artists paid $200-300 one-time license fee for two-year term. No per-stream tracking, no algorithmic promotion, no data harvesting.
Edmonton Public Library’s Rekalman: “Local music is part of local history. Traditional focus on famous authors, famous music doesn’t help us learn about ourselves. You learn about your neighbors, who lives where you live.” Ann Arbor’s Eli Neiburger on controlling digital infrastructure: “Early web had lot more variety and opportunity. As you’ve seen so much consolidation, public libraries are one of few forces resisting that. Very few in corporate world have incentive to think how collections will be useful in 500 years.”
Rapper/producer Cadence Weapon (Edmonton): “Engaging with streaming companies is like Wizard of Oz to me—you can’t see them, you don’t know who’s behind these playlists, you don’t want to ruffle feathers and never end up on playlist. If it’s somebody from your community you actually know, there’s certain level of trust.”
Public Funding Models:
Ireland (2022-2025): 2,000 artists (584 musicians) receiving €325/week basic income pilot; early research shows decreased anxiety/depression, more hours on arts practice
France (since 1936): Intermittence du Spectacle—musicians clocking 507 hours/year as performing artist get ~€1,300/month unemployment benefits accounting for irregular work
Norway: Project grants + salary-style basic income periods; musician Jenny Hval: “Allows for practice involving your whole being instead of just the product where you’re content creator. I am valued somehow. I am a citizen.”
These models prove public music funding works—but they require political will, robust social safety nets, and cultural consensus that art is public good worth supporting. In US context, where healthcare is privatized and social services gutted, public arts funding faces uphill battle.
What Pelly Proves, What She Doesn’t, and Why It Matters
Proven Beyond Reasonable Doubt:
Spotify’s founding mythology is false: Ad-tech entrepreneurs seeking traffic source, not mission to save music (documentary evidence: patents, corporate registrations, founder statements)
Major labels designed streaming system for their benefit: Equity stakes, advances, guaranteed minimums, most-favored-nation clauses, influence over platform evolution (leaked contracts, financial reporting)
Perfect Fit Content program exists and is substantial: 100+ playlists over 90% stock music, €61.4M gross profit annually, systematic replacement of artists with cheaper alternatives (internal Slack, Swedish journalism, musician testimony)
Algorithmic systems optimize for engagement, not discovery: Success metrics are session extension, retention, not musical diversity (former engineer testimony, patent applications, product feature analysis)
Surveillance and data selling are core business: 67 tracking companies, Axiom partnership, mood data sold to WPP, emotion detection patents (GDPR case evidence, SEC filings)
Discovery Mode functions as payola: 30% royalty cuts for undisclosed algorithmic promotion, €61.4M profit, no user labeling (internal documents, House Judiciary letter)
Independent musicians face precarious livelihoods: Median income insufficient for living expenses, 50%+ require non-music work (UK Musicians Census, Princeton/MusiCares survey)
Proven with Strong but Not Definitive Evidence:
Spotify caused lean-back listening culture: Correlation clear (mood playlists, sleep streaming), causation murkier—could be responding to existing preference rather than creating it
Algorithmic personalization creates cultural silos: Evidence of self-referential recommendations, but proving this narrows rather than expands listening would require controlled experiments
Platform pressures shaped streambait pop sound: Strong evidence of optimization (chorus-first, 30-second hooks), but correlation with artistic choices doesn’t prove causation—young artists might genuinely prefer this aesthetic
Microgenres harm music culture: Hyperpop case study shows real harm (erasure, commodification), but doesn’t prove all classification systems equally harmful
What Remains Unproven or Underdeveloped:
The Counterfactual Problem: If streaming didn’t exist, would musicians be better off? Chapter 1 documents how CD sales collapsed 60% (2000-2010) before streaming dominance. iTunes also paid poorly. Piracy was rampant. What was the viable alternative path?
User Agency Question: Book treats users as victims of manipulation, but doesn’t adequately address users who want lean-back listening, who enjoy chill playlists, who find algorithmic recommendations genuinely useful. Are they wrong, or are their preferences legitimate even if platforms exploit them?
Scale of Alternatives: Catalytic Sound (30 artists), library streaming (dozens of libraries, few hundred albums each), public funding pilots (2,000 Irish artists) are inspiring but microscopic compared to Spotify’s 615M users, 100M+ tracks. How do these models scale without replicating streaming’s problems?
Labor vs. Capitalism Distinction: Is the problem streaming technology (could be fixed through regulation, better ownership) or capitalism itself (requires dismantling)? Book oscillates between reformist (Living Wage Act, FTC enforcement) and abolitionist (cooperative alternatives, public funding) without resolving which theory of change is primary.
Major Label Complicity: While book documents majors’ privileged deals, it doesn’t fully explore whether independent musicians would be better off if majors weren’t involved. Pro-rata system benefits majors, yes—but majors also subsidize platform viability. Would Spotify exist without major catalogs? If not, would alternatives emerge?
The Methodological Triumph and Its Limits
Pelly’s greatest achievement is making the invisible visible. She obtained internal Slack messages showing PFC gross profit tracking, Discovery Mode expenditures, Strategic Programming team’s metadata tagging work, playlist editors’ ethical objections. This level of access is extraordinary—comparable to Frances Haugen’s Facebook Files or Susan Fowler’s Uber testimony. Combined with 100+ interviews (former employees, musicians, label workers, engineers), Swedish journalism, patent filings, financial documents, and her own embedded research in DIY scene, the evidence base is formidable.
The book’s structure—moving from founding mythology through technical systems to cultural/economic consequences to political protection to organized resistance—creates cumulative persuasive force. Each chapter provides evidence for subsequent claims. By the time you reach Discovery Mode (Chapter 16), you understand why independent artists have no choice but to participate (Chapters 13-15 documented their economic desperation) and how the system was designed to exploit them (Chapters 1-2 showed extraction was always the goal).
The weakness is normative framework sometimes obscuring rather than clarifying. Pelly’s commitment to DIY values, grassroots organizing, and anti-corporate politics produces powerful moral clarity—but it also forecloses certain questions. When she writes that streaming “relegates music to something passable, just filling the air to drown out the office workers’ inner thoughts as spreadsheets get finalized,” the judgment is clear: this isn’t real listening. But who decides what counts as real listening? Pauline Oliveros’s “deep listening” philosophy is one tradition; there are others.
Similarly, the Brian Eno discussion around ambient music gets tangled in definitional debates. Eno wanted Music for Airports to “accommodate many levels of listening attention without enforcing one in particular”—it should be “as ignorable as it is interesting.” But Pelly argues streaming’s functional music betrays this by becoming only ignorable. The distinction is that Eno’s ambient was compositionally sophisticated, meant to enhance spaces and induce reflection, while Spotify’s mood playlists are generic filler meant to disappear. This is aesthetically defensible but doesn’t resolve whether some users genuinely benefit from functional music—even if corporations exploit the category.
The book’s populism—championing DIY ethics, challenging corporate power, centering working musicians—is its moral strength. But it occasionally produces analytical blind spots. When independent label managers complain about losing playlist placements, Pelly frames this as Spotify’s betrayal of independent music. But it’s also possible the definition of indie has genuinely changed—that younger listeners hear “indie” as aesthetic (bedroom pop, chill vocals) not production model, and Spotify’s playlists reflect this shift rather than cause it.
The Question Pelly Forces Us to Confront
The book’s ultimate provocation isn’t whether Spotify is exploitative—Pelly proves this comprehensively. It’s whether exploitation can be reformed or must be abolished.
The reformist path (Chapters 13, 16, 18): Living Wage for Musicians Act creates new royalty stream. FTC bans digital payola. GDPR-style privacy laws limit surveillance. User-centric payments replace pro-rata. These are achievable through legislation, regulation, organizing.
The abolitionist path (Conclusion): Cooperatives like Catalytic Sound, library streaming projects, public funding replacing market mechanisms, “delinking from harmful systems” (Brandon King, Resonate). These require reimagining digital infrastructure, music as public good, rejecting venture capital/corporate ownership entirely.
Pelly doesn’t resolve this because she can’t—it’s a question about theory of change under capitalism. But her evidence suggests both paths are necessary. The Living Wage Act is “emergency fix” (Abbey’s words), not horizon of imagination. Cooperatives demonstrate alternatives exist but face scaling challenges. Public funding works (Ireland, France, Norway) but requires political systems that treat art as public good, not market commodity.
What Pelly proves definitively: the current system was designed to extract value from musicians and listeners while enriching a small class of executives, venture capitalists, and major label shareholders. The “democratization” narrative was always a lie. Every technical choice—from royalty calculation complexity to algorithmic recommendation optimization to Discovery Mode’s undisclosed payments—serves extraction, not access. Every market expansion (from music to podcasts to audiobooks) extends the logic. Every “innovation” (AI DJ, mood playlists, generative music) turbocharges it.
The book’s power is this: after 400 pages, you cannot unsee the machine. When Spotify Wrapped arrives each December, you’ll recognize it as surveillance rebranded as cute gamification. When you press play on “Chill Vibes,” you’ll wonder whether you’re hearing music or hearing a Swedish production company’s response to Spotify’s prompt for cheap ambient filler. When Discovery Weekly suggests a new artist, you’ll question whether this is recommendation or paid placement. The veil is lifted. The emperor has no clothes. The playlist is a lie.
But Pelly leaves readers with more than critique. The conclusion surveys musicians organizing (UMAW, Music Workers Alliance, Swedish engineers’ union), cooperative models (Catalytic Sound, Resonate), library streaming, public funding. These aren’t utopian fantasies—they’re working models, proving alternatives exist. The question is whether we have political will to build them at scale.
Nati Linaris (Resonate board member) names the framework: “Fighting the bad and building the new. It’s about doing both at the same time.” Demanding better deals from corporate streaming while constructing artist-run alternatives. Regulating surveillance while developing cooperative infrastructure. This is the only viable path: reform to reduce immediate harm, organize to build collective power, construct alternatives to demonstrate different models are possible.
The book ends with this: “On a collective level, we have to be active participants in the cultural economies we want to see flourish. We have to validate the culture we want to see in the world. The corporate culture industry entrenches its power not just through controlling the marketplace, but also by controlling the popular imagination, by convincing us there are no alternatives. The alternatives are growing all around us though.”
Closing: The Cost of the Perfect Playlist
What’s the cost? Pelly’s answer is comprehensive: $0.0035 per stream for musicians. €61.4 million annual profit from pay-to-play schemes. 86% of tracks demonetized by “artist-centric” royalty reform. Billionaire executives while 50%+ of working musicians require day jobs. Major labels consolidating distribution while calling it democratization. Ghost artists replacing real musicians on playlists with millions of followers. Algorithmic homogenization flattening aesthetic diversity. Surveillance apparatus harvesting emotional data for targeted advertising. Political lobbying protecting extraction from regulation.
But the deeper cost is epistemological and cultural. When music becomes background utility, when listening becomes data generation, when discovery becomes algorithmic regurgitation of your own taste profile, when fandom becomes metadata labor, when “perfect playlist” means perfectly optimized for corporate profit—we lose the very reasons music matters. The moments where ineffable becomes real, where loneliness dissipates, where world briefly makes sense. The possibility of being surprised, challenged, changed by sounds we didn’t know we needed to hear.
Anoni, whose “Why Am I Alive Now” was algorithmically shoved into a “Chill Vibes” playlist despite being, in her words, “so despairing,” names what’s lost: “There used to be a system where harder music had a place within the pantheon of the economy of music. Even within capitalism, there were smaller economies, smaller worlds where smaller musicians were thriving. Losing the physical object of the record and instead leaning on monetization based on plays really abends the transaction between songwriter and listener. It is a very politically astute maneuver that favors a narcotic relationship to music over a complex meditative relationship to music.”
This is Mood Machine‘s final provocation: Spotify didn’t save music from piracy. It perfected a more sophisticated form of theft—one that compensates artists just enough to claim legitimacy, surveils listeners just subtly enough to avoid revolt, and extracts value efficiently enough to enrich billionaires while musicians work day jobs. The perfect playlist isn’t perfect for you. It’s perfect for them.
Pelly has written the definitive account of how we got here. Whether we can get somewhere else remains an open question—one she’s given us the tools to begin answering.
Tags: Spotify platform capitalism, streaming surveillance economy, music industry labor exploitation, algorithmic cultural homogenization, Perfect Fit Content investigation