In 2008 a European music listener could buy songs and albums, or download them without paying. Pirated catalogs held more music than any licensed service, and a downloaded song played at once from the listener's own computer SPX-1. In Sweden piracy had cost the industry more than 80% of its revenue SPX-1. A legal alternative had to compete with free on convenience while paying for every play, and only labels and rights agencies could license the music people wanted, agreement by agreement SP16 SP15.
Spotify
Spotify made a licensed catalog play as fast as a pirated one, gave it away free with ads, spread through Facebook, and turned listeners' playlists into recommendations that made them costly to leave.
How Spotify won
- 1 · 2006–08Speed plus a licensed catalog: better than piracyA 200-millisecond play target (2006) made streaming feel local; a year of guaranteed Swedish revenue won the licenses (2008).Speed
- 2 · 2008–17A free tier as the funnel to PremiumFree kept as good as possible so heavy listeners convert; more than 60% of new Premium subscribers came from free (2014–17).Freemium Model
- 3 · 2011Spreading through Facebook, landing on freeFacebook's listening feed (September 2011) took Spotify's Facebook app from 3.4 million to 5.3 million monthly users in a week.Distribution Partnerships
- 4 · 2014–16Listener scale feeds personal discoveryTunigo's editors got listening data and The Echo Nest was bought (2014); Discover Weekly (2015), drawn from 2 billion user playlists, reached 40 million users in a year.Personalization
- 5 · 2015–Switching costs and scale terms with labelsChurn fell as Spotify programmed 31% of listening (2017); the major labels cut their share of revenue (2017).Switching costs
Versus Rdio: Rdio built the nicer app and charged from the first play, so friends arriving from Facebook met a paywall RD1 SPX-14. Spotify made free listening feel like owning all music, used it to reach far more listeners, and turned their playlists into personal recommendations.
Arena: Market Conditions Before Spotify
How each step happened
Step 1 of 5 · 2006–08
Speed plus a licensed catalog: better than piracy
Spotify's founding problem was that the free alternative was also the larger one. Pirated catalogs held more music than any licensed service, so Spotify had to beat piracy at what people used it for: playing a song the moment they chose it SPX-1. In 2006 Daniel Ek, the founder, set his engineers a 200-millisecond target, the time he had read the brain needs to perceive anything. They reached about half a second, and a loading animation made that feel instant SPX-1. Streaming felt like owning the music.
Speed was worth nothing without the music. Labels would not give away their catalogs on the promise that listeners would pay later, so Ek narrowed the request to his home market and guaranteed the labels a year of Swedish revenue if they allowed the model SPX-1. At the October 2008 launch in seven European countries, Merlin added the largest independent catalog through one agreement, and Spotify offered an advertising-supported free version beside a paid subscription SP15. Listeners got a legal service that competed with piracy on convenience, and Spotify depended on the labels from its first day.
Rivals Rdio launched in the United States in 2010 as a paid service with a clean design, and its founders credited label relationships from their Kazaa years with getting the licenses signed RD1. Rdio measured itself against buying music; Spotify measured itself against piracy, which was free and instant.
Step 2 of 5 · 2008–17
A free tier as the funnel to Premium
The free version was the center of the business, not a trial. Spotify kept it as good as it could afford instead of rationing it. Gustav Soderstrom, who joined in 2009 to lead mobile, reasoned that the more a free listener used Spotify, the sooner they wanted what only Premium, the paid subscription, offered: offline listening and no ads SPX-1. Advertising paid for free listening, and Spotify paid rights holders for every free play SP16 SP3. On a computer free listeners chose any song; on mobile they chose the kind of music but not the track SP16. Premium sold control, not access.
Ek stated the logic in 2014: paid-only services had failed because people already got music free, and free services paid artists almost nothing, so Spotify linked the two: "No free, no paid" SP16. More than 80% of its 12.5 million subscribers had started as free users SP16. The 2018 listing filing measured the same funnel: the free service supplied more than 60% of gross Premium additions from February 2014, and Premium reached 71 million subscribers by the end of 2017 SP3. It also made the next step possible: a product people could pass to friends at no cost.
Rivals Rdio stayed paid-only until it added ad-supported web listening in January 2014, two days after Spotify made its own free listening unlimited SPX-15. As late as June 2015 CEO Anthony Bay called free on-demand listening unsustainable and kept Rdio's free tier to radio stations SPX-9 SPX-11.
Step 3 of 5 · 2011
Spreading through Facebook, landing on free
The free tier also decided how Spotify spread. At Facebook's f8 conference in September 2011, Spotify joined the new listening feed, which showed friends what each person played SPX-4. In one week Spotify's Facebook app went from 3.4 million to 5.3 million monthly users, and more than a million new users joined SPX-4 SPX-5. A friend who clicked a post could play the song free.
The channel alone gave no advantage: MOG, Rdio and Slacker joined the same feed SPX-4. What differed was where a curious friend landed. At Spotify it was free listening; at Rdio it was a subscription page SPX-14. The partnership paid off because step 2 had made the product free to try.
Rivals Rdio put its listening into Facebook feeds too, but had no free tier until 2013–14 SPX-4 SPX-14. Former staff said it never kept a marketing chief for more than a few months SPX-14 SPX-13.
Step 4 of 5 · 2014–16
Listener scale feeds personal discovery
With the whole catalog available, the listener's problem became choosing what to play SP4. Spotify bought Tunigo, a team of playlist editors, and gave them data on skips and on how each playlist performed; Soderstrom says they became the most data-driven editorial team in music SPX-1. In March 2014 it bought The Echo Nest, whose recommendation technology also served Rdio and other services, and Spotify's editors built playlists from its analysis of every song GFA-2 GFA-3.
Discover Weekly, launched in July 2015, made recommendations a habit. The team found Spotify's personalization technology buried several layers deep in the product, so it delivered suggestions as a two-hour playlist refreshed every Monday SP4. The input was scale. Matt Ogle, who led the product, says it draws on what all users add to their own playlists, more than 2 billion of them SPX-2. Each listener's playlists improved everyone else's recommendations; a service with fewer listeners had less to draw on. Discover Weekly reached 40 million users in its first year SPX-3. Where rivals licensed much the same catalog, personal discovery gave listeners a reason to choose Spotify SP14.
Rivals Rdio built discovery around following other listeners RD1 and dropped The Echo Nest within a week of Spotify buying it, moving to suppliers such as Gracenote GFA-1. In 2015 it took about $1.5 million a month in subscriptions SPX-12. Apple Music and Tidal offered editorial, blog-style discovery SPX-3.
PersonalizationData network effectsM&A StrategyAggregator Strategy
Step 5 of 5 · 2015–
Switching costs and scale terms with labels
Personal discovery also made leaving costly. By 2017 Spotify programmed about 31% of all listening through its playlists, up from less than 20% two years earlier SP3. A listener's library, playlists and taste history lived inside Spotify. Premium churn, the share of subscribers cancelling each month, fell from 7.5% in the fourth quarter of 2015 to 5.1% two years later as engagement rose, while Apple Music, which the filing called its closest competitor, went after the same listeners SP3. Spotify also told SongShift, an app that moved libraries between services, to stop transfers away from Spotify SPX-8.
Scale changed the terms with the labels, within limits. In 2017 the three major labels cut their share of Spotify's revenue from 55% to 52%, and gross margin rose from 14% to 21% SPX-6. Universal accepted lower royalties in exchange for holding some new albums on Premium for up to two weeks SPX-7. CFO Barry McCarthy described a dependence that ran both ways: the labels needed Spotify's listeners, and Spotify still needed their music SPX-6. Universal, Sony, Warner and Merlin still owned about 85% of what was streamed SP14.
Podcasts offered audio whose economics did not run through the labels SP14. In 2019 Spotify bought Gimlet, which produced shows, and Anchor, which let anyone make them SP11, and in 2020 it signed Joe Rogan's show as an exclusive SP12. It later revised the bet: a 17% staff cut in December 2023 SP7, and a 2024 Rogan renewal that put the show on other platforms SP6. The advantage that lasted stayed where step 4 built it: in each listener's library and recommendations.
Rivals Rdio lost about $2 million a month, and its former design lead said the labels took the largest share of revenue whatever Rdio did SPX-14 SPX-13. It filed for bankruptcy in November 2015, and Pandora bought selected assets for $75 million RD2.
Key dates
- 2006Ek sets a 200-millisecond play target SPX-1
- 2008-10Launch in seven European countries, free and paid
- 2010-06Rdio opens a paid US preview RD1
- 2011-09Facebook listening feed; Spotify's Facebook app goes from 3.4 million to 5.3 million monthly users SPX-4
- 2014-01Free listening made unlimited; Rdio adds free web listening two days later SPX-15
- 2014-03Buys The Echo Nest; Rdio drops it GFA-2 GFA-1
- 2014-1112.5 million subscribers More than 50 million active users; more than 80% of subscribers started free SP16
- 2015€1,940M revenue Full year SP3
- 2015-06Rdio's CEO calls free on-demand listening unsustainable SPX-9
- 2015-07Discover Weekly launches SPX-2
- 2015-11Rdio bankrupt; Pandora buys its assets for $75 million RD2
- 2016€2,952M revenue Full year SP3
- 2016-05Discover Weekly reaches 40 million users SPX-3
- 2017€4,090M revenue Full year; 71 million Premium subscribers, 159 million monthly active users SP3
- 2017-04Universal lowers royalties in exchange for Premium-first releases SPX-7
- 2018-03Filing to list shares on the NYSE SP3
- 2019-02Gimlet and Anchor bought
- 2020-05Joe Rogan exclusive
- 2024-02Rogan renewal puts the show on other platforms
Sources
Oldest first.
- SP15 Merlin to join Spotify at launch. Rights-holder launch announcement
- RD1 Rdio kicks off US launch: a chat with founders and management. Founder interview
- SPX-5 Spotify gains 1 million new users from Facebook integration. Trade report
- SPX-4 Spotify Users Jump 3.4 Million To 5.3 Million In 1 Week Thanks To Facebook. Trade report
- SPX-15 Rdio Caves To The Competition, Makes Web Streaming Free For All Users. Trade report
- GFA-2 Spotify Acquires The Echo Nest. News article
- GFA-1 Rdio Will Drop The Echo Nest For Music Recommendations Post-Spotify Acquisition. News article
- GFA-3 Inside The Spotify - Echo Nest Skunkworks. Feature article
- SP16 $2 Billion and Counting. Founder explanation
- SPX-9 Rdio CEO talks radio, Spotify, indies and freemium (interview). Participant interview
- SPX-11 Spotify rival attacks free streaming models. Participant interview
- RD2 Pandora asset purchase agreement with Rdio. SEC filing
- SP4 What made Discover Weekly a successful feature launch. Product development account
- SPX-12 Why didn't Pandora buy Rdio? It was $220m in debt. Trade report on bankruptcy filing
- SPX-14 Too good to survive: The rise and fall of Rdio. Trade report
- SPX-13 "The classic startup mistake": What went wrong for Rdio. Participant interview
- SPX-2 Matt Ogle talks Discover Weekly and Spotify's personalised evolution. Participant interview
- SPX-3 Spotify seeks more personalized playlists after Discover Weekly finds 40M users. Trade report
- SPX-7 Spotify trades free-tier album delays for smaller royalties. Trade report
- SP3 F-1/A registration statement. Listing filing
- R15 Acquired — Season 2, Episode 6: Spotify’s Direct Listing. Podcast transcript
- SPX-1 How to build trust fast. Podcast transcript
- SPX-6 Spotify: We don't expect major labels to 'hand us an additional pile of gross margin'. Trade report
- SP11 Audio-First. Founder strategy announcement
- SPX-8 Spotify is blocking apps from transferring music libraries to other services. Trade report
- SP12 Joe Rogan exclusive partnership. Contract announcement
- SP14 Spotify, Netflix, and Aggregation. Strategy analysis
- R19 Acquired — Spotify CEO Daniel Ek. Podcast transcript
- SP7 December 2023 organizational changes. Founder allocation decision
- SP6 New multiyear partnership with Joe Rogan. Contract announcement