It’s hard to remember the days of browsing CD stores and paying $1.99 per individual song on iTunes now that Spotify exists.
Like the ungrateful Millennial I am, I often pop in my headphones, start playing the streaming service and don’t give a second thought to the unfettered access I have to millions of songs and playlists that are tailor-made for the trial and tribulations of life.
This kind of democratic access to such a wide variety of music would’ve been unheard of in my parent’s time, yet I can find curated playlists like ‘Walking like a badass’, ‘Knitting party’ or ‘Forever alone’ and deliver it directly into my earlobes with the swift click of a button.
With over 140 million active users as of June 2017, the algorithms and humans at work behind the scenes at Spotify are moulding the music tastes of a generation.
Gone are the days where people watched music videos on MTV to get inspiration for new songs – now, they’re more likely to discover it through technology: Be it a song the streaming service’s algorithm has queued to play automatically, or a playlist that caught their eye.
Spotify’s Australia and New Zealand senior music editor, Alicia Sbrugnera, is one of the people making this happen. As far as jobs go that didn’t exist 15 years ago and are enviously cool, hers deserves a spot on the list (especially considering she says she spends more time at gigs than she does at home.)
She is one of the 90 or so Spotify curators worldwide who is tasked with crafting the thousands of playlists on the streaming service, as well as introducing new music to Australian and Kiwi listeners.
“We have over 30 million songs on Spotify, and it’s my job to help our audience discover new and exciting artists and curate for every conceivable genre, activity, or mood,” Sbrugnera says.
“Our days are pretty dynamic… it all depends on what’s coming down the pipeline – internally or externally, from a new product feature or programming tentpoles to a major new release or artist meeting.”
Being in charge of the playlists on Spotify is no small feat – when combined, they generate around a whopping one billion streams per week. Each is themed to suit a certain genre or mood, such as Coffee + Chill (over 140,000 followers), or Hot Hits New Zealand (over 105,000 followers).
Spotify key statistics:
Subscribers: Over 60 million (as of July 2017)
Active users: Over 140 million (as of June 2017)
Revenue paid to rights holders: $5 billion (as of September 2016)
Number of songs: Over 30 million
Number of playlists: Over 2 billion
While people will always rate their own music taste and devise their own playlists to extremely specific needs (like the ‘I’m getting drunk on a plane’ playlist), according to Spotify, people still appreciate someone doing the hard work and crafting one for them.
About half of its listeners globally (About 70 million people as of June 2017) tune in to the company’s editor-created playlists.
Sbrugnera previously worked as a music programmer for the likes of MTV, but her role with Spotify offers forward a brand-new challenge: the marriage of machine learning and music.
The algorithm software Spotify uses has been key to its success, as it helps serve up user suggestions, as well as the supremely popular Discover Weekly playlist. Every Monday morning, users are delivered a personalised playlist with 30 new songs to try out that are based on their music and listening preferences.
The secret sauce for these suggestions, it turns out, is other users. Spotify looks at the two billion or so playlists created by other users with similar music tastes and bases decisions off that.
The company is also using deep learning, which finds patterns in vast amounts of data, with powerful computers that are “trained” by humans to improve its decisions.
The top five playlists streamed in New Zealand:
1. Hot Hits NZ
2. Just Chill
3. Chilled Hits
4. Morning Motivation
5. Get Popped!
Sbrugnera says the data component has been one the most fascinating aspects of her job to date, as there’s a real-time feedback loop of user data to examine and refine the playlists.
“The way we curate is we come up with ‘hypothesises’, we call them, around particular moods, moments and activities,” she says. “They’re quite targeted to specific audiences, but on the other hand, we have the algorithm, which we see as a really powerful tool in personalisation. We aim to marry these up, as we see it as the way to serve the audience in the best possible way.
“Luckily enough with Spotify, we have access to phenomenal data, so we can see what is going on naturally and organically with our audience base. We can see people’s habits forming and look at a variety of metrics. These playlists are living and breathing entities within the platform, then we sequence them based on how the audience is interacting with them.”
The technology can also pick up on whether there’s a gap needing to be filled in terms of playlists genres. For example, Sbrugnera says fans here are pretty well covered when it comes to pop, hip-hop and electronic playlists, but it’s now focusing on genres that are less mainstream in New Zealand and Australia, such as country music.
“There are underserved genres around the world, so we’re expanding the team and looking at those niche areas,” Sbrugnera says. “That’s where our algorithm is quite a powerful tool.”
Showcasing local talent
Another feature of Spotify is the way its playlists are crafted is a democratic process, as artists aren’t prioritised based on their level of fame.
“Whether an artist is a completely unsigned bedroom artist making music or a major artist making music, it has no bearing on editorial choices,” Sbrugnera says. “It’s really a democratic system that’s based on the strength of the song.”
Case in point: My own Discover Weekly playlist has a song by an obscure Los Angeles-based DJ named frumhere that I fall in love with, but with just 245,265 listens monthly, he probably would’ve never made it onto my radar otherwise.
Editors also divvy up their time listening to 30-second snippets of newly released songs to source new playlist additions. If local editors don’t have the ability to add a song to a playlist at a certain point in time, they can send the song off the global team and see if it would fit in a playlist originating from somewhere else in the world.
This helps with global exposure, opening up Kiwi artists to an industry far beyond the New Zealand music industry’s constraints.
“Over the last 12 months, there’s been some major signings – obviously Lorde is huge, but I think the most exciting music coming out of New Zealand right now is independent music,” Sbrugnera says.
These include local acts like Christchurch-born Theia, who had a runaway success with her single Roam going to clock up 11 million listens. Another is Kiwi Mitch James, who spent two years busking and sleeping on the streets of Europe before being signed to a major label.
He’s now gaining a large following through Spotify, with eight million plays on his song No Fixed Abode. Aldous Harding is another game changer, Sbrugnera says.
“Success doesn’t have to mean a top 50 hit on Spotify, there’s so many other avenues of success. It all comes down to building that audience. It’s exciting times for indies.”
Kiwi artists’ Spotify Success
Mitch James (now signed to Sony NZ)
– 850k monthly listeners
– 80 percent of streams come from Spotify editorial playlists
Aldous Harding (signed to 4AD)
– 900k monthly listeners
– 76 percent of streams come from Spotify editorial playlists
– 450k monthly listeners
– 52 percent of streams come from Spotify editorial playlists
3. New Zealand
- This story originally appeared on Idealog.