The Spotify Play

by Sven Carlsson

  “We should never tell anyone what to listen to,” Daniel told one of his colleagues as late as 2010.

  Soon he would be singing a different tune. Spotify would revamp its radio feature and draw users toward activity-based playlists such as “Italian dinner” and “Dubstep workout.” Slowly, it would build out an à la carte menu of playlists designed for specific moments in the lives of its users. Competitors—such as Jimmy Iovine’s Beats Music—would push them further in that direction.

  FM

  In September 2011, while on vacation in Tennessee, the young programmer Erik Bernhardsson received a call from Spotify’s CPO, Gustav Söderström.

  “We need a really strong radio function. And you know how that stuff works,” Gustav told the twenty-seven-year-old over the phone from Stockholm.

  The bespectacled coder had recently finished a brief stint at a hedge fund in New York City. There, he had used advanced math to help optimize the fund’s high-frequency trading instruments, with the purpose of shaving milliseconds off of each trade in order to beat the market. Since leaving the hedge fund he had accepted a position at Spotify’s North American headquarters in New York. While waiting for his new US work permit, he’d decided to travel south and, like Paul Simon in the famous song, stop in Memphis and spend a few hours at Graceland, the palatial home that Elvis once purchased in the late 1950s.

  Just a few days in, Erik Bernhardsson’s vacation was cut short. Gustav wanted him to head back to New York as soon as possible, promising to compensate him fairly once he had a work permit and could start receiving payouts. Spotify’s head of product had just experienced a moment of clarity and was in a hurry to build a new genre-based radio feature. The idea was to start siphoning users from Pandora, whose listeners could simply pick a song they liked and lean back while being served similar music on an infinite loop.

  Erik hopped in a cab and took the next flight from Memphis to New York. The next morning, he stood outside 111 Eighth Avenue, admiring Spotify’s US offices in Chelsea. The massive brick building took up an entire city block and had, for decades, belonged to the city’s Port Authority. Its new owner was Google, whose New York headquarters claimed several floors of the building.

  Erik took the elevator to the eleventh floor and was greeted by a handful of colleagues in the otherwise empty office. Furniture and boxes lay in piles on the floor. Large windows ran along one side of the space, with the view stretching northeast over the landmarks of Midtown: the Empire State Building; the MetLife building, which appeared to shoot directly out of Grand Central Station; the piercing, metal spire of the Chrysler Building; and Rockefeller Center on 50th Street.

  At this time, Pandora had thirty six million users who listened every month. That made their service three to four times bigger than their Swedish foe. On the New York Stock Exchange, the digital radio company was valued at over $2 billion, around twice as much as Spotify.

  Erik found an available desk and got to work. The millions of user-created playlists, he figured, were a treasure trove in Spotify’s growing data set. The title of each list might say something about the type of music it contained. A playlist called “Chillwave” would be likely to feature contemporary, electronic dream-pop while “Uptempo House” probably contained fast-paced dance music. Within a few hours, Erik had written code that analyzed playlist titles and the songs chosen to fit together in each list. He deployed the code and, as a first test, found eighty four songs that seemed to belong together. That morning, he laid the groundwork for what would become Spotify’s new radio feature, which the company would eventually patent in the United States.

  Three months later, at the Le Web conference in Paris, Daniel Ek walked on stage in a black polo shirt, carrying a water bottle. He launched the revamped Spotify Radio, explaining how users could now choose a song and ask the player to queue a selection of similar music. He made sure to point out that unlike Pandora, his new radio feature allowed for “unlimited skips and unlimited stations.”

  “It was a big use case that a lot of people were asking us for. Today, we’ve covered that use case,” said the Spotify founder.

  A few hours later, when the New York Stock Exchange opened, Pandora suffered a five percent drop in its stock price, wiping out around $70 million from its market cap.

  Spotify Radio was not perfect. Over the coming years, picky listeners would mock the song selections, calling them predictable and simplistic. But the new feature did mark the beginning of the end of Pandora’s dominance within digital radio in the US.

  Viva Las Vegas

  While Spotify was testing the waters in the US, Jimmy Iovine was hard at work building a new type of streaming company, with its roots in the music business.

  In January 2012, a few weeks after Spotify Radio launched, the Brooklyn-born record executive found himself in Las Vegas. The sun was shining on the City of Second Chances, and it was time for the annual Consumer Electronics Show, CES.

  The board of Beats Electronics had convened in a conference room at The Wynn, a five-star hotel on the northern end of The Strip. One item on the agenda was “Daisy,” the streaming project that was still under wraps and, for now, shared its name with Beats president Luke Wood’s dog. Trent Reznor, frontman of the alternative rock band Nine Inch Nails, was seated at the table. He had been tapped as chief creative officer of the project and had worked on a conceptual pitch with the Swede Ola Sars, who was the chief operating officer of “Daisy.”

  The duo’s presentation expanded on the concept of “human curation,” with playlists by Beats-affiliated stars such as Dr. Dre, Will.i.am, and Gwen Stefani. One aspect that really piqued the board’s interest was a feature called “the sentence,” which Sars and his team had cooked up in Stockholm. The idea was that the listener would describe their setting and mood, and let the streaming client set the soundtrack. A listener who was “in a house,” “with their friends,” and felt like “starting a riot” might be served a rowdy song by The Beastie Boys or DMX.

  The executives in the conference room were impressed. At this time, Spotify’s partnership with Facebook was still fresh. The Beats concept felt like a natural next step. The room was in agreement: their blueprint was already better than anything those “robots in Sweden” had cooked up. The presentation ended in cheers, with Dr. Dre clapping loudly.

  Runnin’ Down a Dream

  In a matter of days, the elated mood in Las Vegas had dissipated. Jimmy Iovine seemed impatient and displeased since he, much like Daniel Ek, was only really at ease when things moved fast. The energetic music executive had thought that a Beats streaming service could be ready to go within a few months. That might be possible, Ola Sars would tell him, but only with non-interactive licenses. That would make Beats’ new product a radio service in the style of Pandora, not an on-demand player that would compete squarely with Spotify.

  Sars had, by this time, recruited his former colleague, the previously skeptical Fredric Vinnå, as the project’s Chief Technical Officer. Vinnå moved to Los Angeles and quickly became close friends with Trent Reznor. The two had a lot in common—the long-haired Swede also liked to dress in all black and talk about product design. They would hang out in Reznor’s modernist home in Beverly Hills and frequently return to the idea of a streaming service with a heart and soul. The duo convinced Iovine to take the time to build a fully functioning streaming service. They wanted to challenge Spotify, which the Beats team thought resembled a glorified FTP server: a modest improvement on file-sharing software like Napster, but too technical, and hardly culturally relevant.

  Time was of the essence for the Beats team. To buy some, they figured they needed to acquire an existing streaming company. One option was Rdio, which had its head offices in San Francisco. Jimmy Iovine met with co-founder Niklas Zennström, the Swede who had recently cashed in spectacularly when Skype was sold, for a second time, now to Microsoft. Iovine is said to have left the meeting thinking Zennström wanted too much money fo
r his streaming company. Instead, he and the Beats team decided on MOG, another San Francisco-based streaming service with forty employees. The company was founded a year prior to Spotify, yet didn’t stand a chance against them. Beats acquired MOG in March of 2012 for around $14 million, which would be pocket change in the ensuing streaming wars. Ola Sars would take credit for landing the deal. But soon, the “Daisy” project would move into a new phase that would result in his departure from the company.

  Fredric Vinnå, however, rose through the ranks. He was allotted shares in the company and became Head of Product, filling his team with top talent out of Silicon Valley. It was a perfect setup, but taxing work. In 2013, Vinnå lived in a swank apartment with a view of the San Francisco Bay, but he never had the time to furnish it properly. Most nights he would work until midnight, only to get up early the next morning and do it all over again.

  Long before the product was finished, Iovine was in talks to sell his company. He met with Ted Sarandos, the Chief Content Officer at Netflix, who was quickly becoming one of the entertainment industry’s top executives. One source would recall how Iovine also courted representatives from Google and the team at Microsoft’s Xbox division in Seattle, which had inherited digital music distribution from the discontinued Zune project.

  Many big technology companies had contemplated entering the music-streaming market at this point. Iovine’s main goal, however, was to sell his company to Apple.

  We’re Not Living in America

  Despite its code name, Beats’ intention to challenge Spotify was no longer a secret. By 2013, details of the project had begun to leak to the press. Among those who took notice were Spotify’s founding duo, and one of them was especially annoyed.

  “Trent? Damn him,” Martin Lorentzon said during a meeting in Stockholm, having just been told about the rock star’s influential role at what would eventually be called Beats Music.

  Just a few years prior, the Nine Inch Nails singer was one of many influencers in the music world that helped promote Spotify in the United States. The way Martin saw it, if there was a dark side in the digital music space, Trent Reznor had just enlisted. He now worked for Jimmy Iovine, whose ties to Apple had been clear since he helped Steve Jobs launch the U2 special-edition iPod back in 2004.

  The meeting in Stockholm had been requested by Ola Sars, who was about to start a new streaming company with Spotify as a major shareholder. Daniel Ek was also present to discuss the new venture. Soundtrack Your Brand would offer music streaming to shops and restaurants. But the conversation lingered around Sars’s turbulent time with Iovine, Vinnå, Reznor, and the others. It had now been a year since he left Beats with just over a million dollars as thanks.

  Martin floated the idea of getting Vinnå to come work for Spotify. Daniel seemed especially interested in how Beats Music was meant to work. Sars, bound by a nondisclosure agreement, painted a general picture of how their music curation—with playlists tailored for each user—would work once the service launched.

  Spotify was already heading in the same direction. Besides Spotify Radio, they had launched a “Discover” tab with music recommendations served up by algorithms. But Daniel knew that he needed to program much more of the listening on his service if Spotify was to become a true hit in the United States. He had recently seen market research that showed Spotify lacking a unique advantage over their American competitors. Online radio listeners in the US had responded that their instinct was to turn on Pandora. Those looking for a specific song would tend to search for it on YouTube. Spotify was not the leading contender, not even for the lean-forward user case that it had always centered around.

  The results fascinated several of Spotify’s top executives, who had begun to realize the extent of the problem. They were too far-removed from the average music listener in the US. How might they reach thirteen-year-olds who just wanted to hear the latest cover track aired on American Idol? Spotify’s new tagline—“Music for every moment”—was a roadmap of sorts. But exactly how they would deliver on it remained unclear.

  Daniel had already tried to partner with third-party app developers, once again borrowing a page from Mark Zuckerberg’s playbook. For around eighteen months, leading music magazines like Rolling Stone and Pitchfork had been building apps within Spotify that contained music charts and user reviews of new releases. Def Jam, an influential hip-hop imprint under Universal Music, served Spotify users their music in playlisted form. Other developers offered anything from touring schedules to dating services based on a person’s music taste.

  But the app that truly caught Daniel’s attention was Tunigo. It had been built by a company with offices in a basement on a tucked-away stretch of Kammakargatan, a back street in central Stockholm. In workstations starved of daylight, coders and music editors had built playlists that fit into categories like “romance,” “working out,” “travel,” and “partying.” They had been heavily inspired by Songza, a startup based in Queens, New York, that would later be purchased by Google.

  In May 2013, Spotify paid more than $6 million for Tunigo, the company’s first acquisition since μTorrent in 2006. The new subsidiary was tasked with helping Spotify sift through the roughly twenty million tracks in its catalogue. Several members of the Tunigo team joined Spotify, among them the company’s CEO, Nick Holmstén. He would become director of the section within the Spotify player called “Browse,” where listeners could find new music.

  “Tunigo helps users find great music for every moment,” Nick said when the deal was announced.

  The concept of “moments” would, over time, become central to Spotify’s product strategy. The following year, in an interview with the New Yorker, Daniel would sound like he had found his edge.

  “We’re not in the music space—we’re in the moment space,” he told the magazine.

  Thinkin Bout You

  As the battle for streamed music heated up, Jimmy Iovine was not alone in talking to potential buyers.

  During 2013, Daniel Ek once again held several meetings with representatives from Google. This time, the stakes were higher—as were the dollar sums.

  The Swede had had a complicated relationship with the search giant ever since he, as a teenager, was rejected for a job there. YouTube was now one of Spotify’s core competitors. Many Spotify staffers were bothered by the fact that Google’s sprawling video platform earned advertising dollars on music videos while paying the music industry considerably less than Spotify did.

  Four sources would recall that Daniel spoke to Google about selling his company two years after the US launch. According to one account, Daniel flew to San Francisco to meet Larry Page, the revered co-founder whose crowning achievement—coming up with the “Page Rank” algorithm, the backbone of Google’s search infrastructure—had since spawned an internet empire. By many accounts a problem-oriented, introverted person, Page was now serving as Google’s chief executive, having taken over from Eric Schmidt two years prior.

  Both Daniel Ek and Martin Lorentzon have remained tight-lipped about the discussions. One source, however, would describe how Daniel had seemed interested in a deal only if he could serve as head of both Spotify and YouTube, as a new online media division. A deal with Larry Page would need to include a grander vision about music and video.

  “Daniel was interested in the deal,” as one source would say, “but he didn’t want Spotify to become yet another business unit within Google.”

  The Swede did not appear interested in cashing in on his company and joining Google only to leave once his entire payment package had matured, the way many acquisitions in the tech space tended to pan out. One forthcoming example was Facebook’s $16 billion acquisition of the messaging service WhatsApp.

  Daniel and Page didn’t click, and Daniel left the encounter feeling slightly rejected. The numbers didn’t meet his expectations either, according to two sources.

  During the talks, Daniel was said to have asked for $10 billion and t
hen lowered his ask to $8 billion, one source recalled. Google’s representatives offered $3 billion, or maybe $4 to $5 billion, depending on the source. Regardless, the parties remained billions of dollars apart. The negotiations never amounted to a formal bid.

  “Google knows all about advertising and weren’t very impressed with that side of Spotify’s business,” one source in the record industry would recall.

  In a regulatory filing from December 2013, Google wrote that it had “recently pursued but discontinued a potential buyout of a foreign company, with a valuation estimated in the range of $4 to $5 billion.”

  According to one source, Daniel would have had a hard time pitching Page on the idea that he would also run YouTube.

  “That was the part of Google that everyone wanted to work at,” the source said.

  Daniel Ek once again had walked away from acquisition talks, meaning venture capital was the way forward. His finance team had been drumming up interest for a new round of funding that would bolster Spotify’s valuation and allow its CEO to keep investing in growth.

  It turned out to be a hard sell. The Spotify team met with a wide range of investors but did not have the metrics to show that their business would eventually yield major profits. To convince the skeptics, the Spotify team started making aggressive projections for its future mobile growth, one person would recall. They wanted to issue new shares amounting to 6 percent of the company in exchange for $250 million. That would indicate a total value of more than $4 billion for the company.

  Ordinarily, a deal of that size would mostly be swallowed by one or two lead investors, with smaller shareholders sharing the scraps. But in this case, there was only one interested party: Technology Crossover Ventures, a fund based in Silicon Valley. To ensure its influence over proceedings at Spotify, TCV secured a board seat at the company.