How Spotify is insidiously changing your taste in music
Over the past few months, my music-obsessed corner of the internet has found itself baffled by a streaming-borne phenomenon. Dozens of Spotify users report that after they’ve pressed play on a popular song – the latest chart-topping release from Taylor Swift; another instalment in the war of words between US rappers Drake and Kendrick Lamar – and have listened to the end, the track that has automatically followed has been Sabrina Carpenter’s ‘Espresso’.
This breakout single by the American Disney star, which has now been Number 1 in the UK for five weeks, is Gen Z’s answer to ‘Barbie Girl’: sickeningly sweet and seemingly unavoidable, hinging on a catchy chorus that’s stuck in your head after just one listen. Yet Carpenter’s popularity points not just to her ability to craft an earworm, but also to the power of the algorithm: the secretive mechanics that guide certain listeners to certain artists or genres. Spotify, the world’s largest streaming service, has effectively abandoned its position as our foremost “digital musical library”, home to 100 million songs, in favour of pointing listeners towards what it thinks they should hear next.
Founded by Swedish entrepreneurs Daniel Ek and Martin Lorentzon in 2006, Spotify prides itself on an intimate knowledge of its consumer base, a knowledge gained, in large part, through monitoring their listening habits. With more than 615 million users worldwide and an estimated market cap of $60 billion, it has become the world’s pre-eminent source of music discovery. Just how it uses personal data, and what it means for the future of curiosity, is the knotty question that Glenn McDonald, a former “data alchemist” at Spotify, attempts to answer in You Have Not Yet Heard Your Favourite Song, his sprawling history of the music industry’s increasing reliance on streaming services.
McDonald starts in an era when consuming music meant three options: records, radio, or if you wanted your pop with a pint in the pub, jukeboxes. “Before the internet,” he writes, “when music was already amazing and felt like it was everywhere, listeners had to be hunters and collectors first.” One could stroll into a record store at noon and emerge, bleary-eyed and clutching an empty wallet, just in time for dinner, after spending hours rifling through gems from Pink Floyd, Louis Armstrong or NWA.
Radio handed the listener more power over what they wanted to hear. If an analogue station plays a song you don’t like, you can simply switch over to a regional or digital alternative. Jukeboxes, too, offer an opportunity to rifle through songs while you search for old favourites. But even those two options had limited breadth, with a finite number of channels – and songs – available. Spotify, by contrast, has spent the past two decades promising to be a turbo-charged amalgamation of all three providers: a one-stop shop for all the music in the world.
McDonald worked at Spotify until December 2023, and analysed the streaming habits of 500 million users during his time there. He also categorised, as part of a wider data team, the platform’s millions of songs, from around one million artists, into more than 6,000 niche sub-genres, from Aarhus Indie to Zydeco. Here, he reveals just how Spotify – which he describes as an advanced form of “surveillance capitalism” – decides what to recommend to each listener, when, and why.
The main barometers are obvious: name, age, gender. Others, such as location or existing tastes, can be determined by IP address, the version of the app being used, or how much of a certain song you played before switching over to another. It explains why the aforementioned pop music-obsessed corner of the internet is being plagued by Espresso; Spotify understands that a person who has regularly listened to Swift will probably enjoy the music of similar acts (Carpenter was an opening act on the Eras tour). But there are limits to this knowledge, just as there are always limits to how much computers can learn about humans. “They know what we play,” McDonald writes, “but they can’t see whether we’re dancing enraptured while the music spins, or two rooms away folding distractingly crinkly laundry.”
Throughout McDonald’s book, personal anecdotes and his own love of music spill out in witty, conversational prose. Even chapters that delve into streaming’s complex finances – unsurprisingly, your £15 monthly fee does not go directly to your favourite artist, but is split between that month’s “most streamed”, meaning that megastars such as Swift and Ed Sheeran stay at the top of the pyramid – are told in layman’s terms. Critics often deplore streaming for cheapening art, but as McDonald explains, at least consumers are regularly paying for their music again. Half a billion people spending £15 a month is a better musical ecosystem than one in which 1,000 people buy one record a year, or 100,000 people stream songs for free on YouTube.
Recent controversies, admittedly, have made it easy to vilify the platform. Neil Young and Joni Mitchell removed their catalogues from Spotify in 2022 in protest at anti-vax comments made by star podcaster Joe Rogan; yet, two years later, they’re back. Even for artists of their stature, with a loyal, multi-generational listenership, streaming is too big a beast to fight. Better to accept its limitations, then, and celebrate the positives: paying customers, aspiring artists being able to upload their own songs, and the wondrous availability of the world’s music at the touch of a button – as long as you’re willing to dig.
You Have Not Yet Heard Your Favourite Song is published by Canbury Press at £16.99. To order your copy for £14.99, call 0808 196 6794 or visit Telegraph Books
Solve the daily Crossword
