What the avalanche of new music means – from overwhelmed listeners to algorithms that seek out unsigned talent for labels
By the time you finish reading this sentence, three new songs will have been uploaded onto Spotify. This time tomorrow, the number will have risen above 20,000 songs — a daily deluge of music which would take you a month and a half to listen to . . . without sleep. By this time next year, more than 7 million new tracks will have been crammed onto the service.
Spotify boss Daniel Ek revealed these numbers back in March at his company’s Investor Day in New York, and the music business hasn’t stopped talking about them since. This trend isn’t limited to Spotify, either. On other music-streaming services, it’s even worse: the CEO of Rhapsody/Napster, Bill Patrizio, confirmed in June that his platform ingests 24,000 new tracks each day, or 1,000 tracks per hour.
One executive whom many in music regard as having a sharp intellect is Greg Maffei, CEO of Liberty Media — an $8 billion-a-year U.S. corporation that owns SiriusXM, will soon own Spotify rival Pandora and also owns around a third of Live Nation. Speaking to Liberty’s investors in mid-November, Maffei warned that a slightly less intense, but no less exhausting, trend was now hitting the world of scripted entertainment.
He cited figures from FX Networks CEO John Landgraf showing that in 2017, 487 original scripted TV series aired in the United States — with online shows up by 680 percent since 2012. Maffei warned that, in general, “content businesses have become a scary place,” and cautioned that, across a media landscape that also includes Instagram, Fortnite, Facebook and Twitter, there is “massive competition for audience . . . and a limit of only 24 hours in a day.”
As for YouTube? Eeesh. Owner Google/Alphabet says that more than 400 hours of video are uploaded onto the platform every single minute.
The music business has reacted to this onslaught of content in a few different ways. Such conditions have given rise to tech startups like Instrumental, Asaii and Sodatone. These Artificial Intelligence-driven tools use machine learning to mine tens of millions of unsigned artists on Spotify, YouTube and SoundCloud. Then, largely based on streaming performance, they highlight the best and the brightest.
In response, paranoid that tomorrow’s superstars are being lost among the dross — or worse, nabbed by a competitor — the music industry has reached for the checkbook: Apple acquired Asaii (technically acqui-hired) in October, and Warner Music Group snapped up Sodatone for an undisclosed fee in March.
Interestingly, far from the oft-predicted story of “self-releasing artists” destroying the world’s biggest record companies, the major labels have arguably benefited from this profusion of (largely) DIY music. Universal, Sony and Warner love this cacophonous digital noise, because it gives them the opportunity to convince artists that they are the sole masters at cutting through it, via their marketing and promotion expertise. To be fair, they may have a point: According to MBW’s calculations, major-label-distributed music claimed more than 80 percent of all payouts from music-streaming services around the world in 2017.
Meanwhile, forward-thinking artists have twigged that the old-world model — that of limited releases being sold on limited shelf space via limited formats — is now dead and buried. Instead, they are experimenting amid the avalanche.
Witness Kanye West’s decision to release five shorter-than-usual albums, all of which he produced, across consecutive weeks in May and June this year. (Those albums, in order: Pusha T’s Daytona, West’s Ye, Kids See Ghosts’ self-titled LP, Nas’ Nasir and Teyana Taylor’s K.T.S.E.) Yes: In today’s punishing, never-ending release environment, even Kanye West has to invent new ways to make people pay attention to him.
As for Spotify? Its reaction to these escalating numbers seems to be “Bring it on.” The Swedish firm recently announced that it now allows artists to upload their music direct onto its platform, in a style similar to SoundCloud. As a result, you can expect that 20,000-a-day figure to rise considerably in the months ahead.
Where this explosion in music content has created the biggest impact, though, is in front of music fans. With such a crazy volume of music being uploaded, the art of recommending tracks — and having the public trust in those recommendations — has become a precious commodity.
Spotify’s biggest weapon in this recommendation game, naturally, is its playlists, which are powered partly by human programmers, and partly by algorithms. Yet questions exist as to whether Spotify’s leading playlists, in this context, are actually doing their job.
According to data obtained from ChartMetric, the three most popular (non-chart) playlists on Spotify worldwide are Today’s Top Hits, Rap Caviar and Vivo Latino! — with more than 40 million followers between them.
With 7 million songs being uploaded to Spotify each year, remember that these lists promise to do the hard work for you: plucking out all the brilliant bangers so you don’t have to.
Except Today’s Top Hits is composed of a not-inconsequential 50 tracks — as is RapCaviar, as is Vivo Latino! Between these three playlists alone, which all update at least once a week, that’s nearly nine hours of music.
The solution to further refining this mind-boggling volume of music may lie in the steely palms of the robots. Spotify has already garnered user acclaim for personalized playlists like Discover Weekly, Daily Mix and Release Radar, which serve you, and only you, a list of tracks based on your personal tastes.
The company has now also been caught testing out semi-personalized playlists. In other words, playlists that appear to be the same for everyone — but which sneakily alter certain tracks based on your past individual preferences.
To sum up, then: Computer algorithms are recommending artists to record companies that, amid a mushrooming quagmire of user-uploaded tracks, these labels hope their customers might like. At the same time, similar computer algorithms are directly recommending tracks to those very same customers, based on their idiosyncratic tastes.
All the while, the enormous conglomeration of recorded music cradled by Spotify’s servers continues to grow bigger and bigger, and — quite literally — more and more unlistenable.