In 1993, the senior vice-president of a successful hedge fund quit his job to set up an online bookshop. To his friends and colleagues, this must have been madness. In 2020, the easiest, fastest and cheapest way to set up almost any business is on the internet. But in the early 90s, the idea of quitting a well-paying job to have a go making money on this new fangled ‘internet’ was unheard of. Brick and mortar stores still reigned supreme.
But while it’s true that 90% of startups fail, this one didn’t. The company was Amazon, and the senior vice-president was Jeff Bezos, who today is the richest man in the world.
Risking it all on relatively new technology was what took Bezos to the heights he now sits at, and the speed of technology’s evolution has only increased since he founded Amazon. A speed that the music industry, at times, has struggled to keep up with. It was undeniably slow to figure out legal music downloads, with multiple ventures failing before iTunes managed to make it stick. Next came streaming services, which have been accepted more quickly – yet are mired in controversy over payments to artists.
But both of these deal with sales, and while sales of records, tickets and merchandise are how the majority of the industry make money, they aren’t the key component. The artists are. The music industry sells art, and comprehending art and its worth have traditionally thought to have been beyond the ability of technology. Until now…
Robots and Recruitment
From friendly robots like Star Trek’s ‘Data’ and Lost in Space’s imaginatively named ‘Robot’, to the villainous Skynet and The Matrix, artificial intelligence, or A.I., once only existed in the realms of science fiction. And while these are all fanciful examples of a technology that is really only in its early days, A.I. is in use today.
Take Eightfold, an online recruitment service. Its A.I. searches through thousands of CVs and offer innovative suggestions for hires that may have fallen through the cracks of human error. In one example highlighted in an New York Times article, Eightfold themselves were looking for a Data Analyst. Thanks to their A.I. system, they eventually landed on a prospect whose experience lined up with what they needed, and yet had never held a Data Analyst job before.
Automated Talent Searches
Of course, the tasks of hiring a Data Analyst and finding the ‘next big thing’ in music are worlds apart. Exactly how would a computer go about discovering talent? It could certainly determine singers with the biggest range, or seek out guitarists who can shred faster than anyone else. But these kinds of factors aren’t necessarily what makes a star.
Enter Snafu Records. Marketing itself as a 21st Century Music Company and the first ‘full-service A.I.-enabled record label, the Swedish company uses its own A.I. algorithms to find exciting new talent. But it’s not looking for musical talent, at least not directly.
Each week Snafu’s A.I. trawls through 150,000 tracks on online platforms like YouTube, Spotify, Tik Tok and Soundcloud. Its data analytics take in everything from the amount of plays a song receives, to the reaction it garners on social media, to which playlists it is added to. Through this, Snafu aims to pick up on new tracks within days rather than weeks, as is the current norm.
With backing from ABBA’s Agnetha Fältskog, and a cash injection of over £2 million from venture capitalists, the idea definitely has some legs. Already the label has found some success, with 16 signings worldwide. One of their first artists, Joan, were signed with days of releasing their first single and now boast an impressive 20 million streams across all platforms.
What does this mean for music?
This isn’t the first time that analytics have been used in the music industry, far from it. All forms of social media offer data to their users, and industry professionals have clocked on to its usefulness.
Any agent worth their salt will check out where their artist is most popular before routing a tour. In some ways the uses are limited for larger artists, there are only so many arena-sized venues available, but for smaller artists, the information is invaluable. It clearly makes sense to book shows where people are definitely listening to their music.
The real question is: what do services like Snafu mean for the future of the music industry? A&R is one of the bedrocks of the music industry, and a keen ear for ‘the next big thing’ has helped make the career of many an industry executive. If a computer program can find these artists on its own, what need is there for talent scouts to spend their nights in dingy clubs or industry showcases?
The good news is that this doesn’t completely spell the end of the human component to the music business. Just as booking agents are still required to negotiate deals, Snafu realises that computer programs shouldn’t be alone in deciding who deserves a record deal. After analysing an artist, Snafu implements a ranking system that is looked over by a team of real people who then check out the artists and decide who to sign.
It also remains to be seen when, if ever, AI will become the standard method for discovering new talent. Music is subjective, and while Snafu’s algorithms take into account human opinion, a real person might find an artist worth developing that hasn’t yet been found by the general public. There’s also the fact that Spotify alone is adding 280,000 new tracks a week, so the label’s algorithm only covers around half of what is being released. But in a time when labels are tightening their belts, this seems to be a surer way of securing returns on investment.
Revolutions in technology are always worth following, and they often create new opportunities for employment. Keep an eye on Snafu, but unlike Skynet, don’t expect them to take over the world very soon.
If you dream of achieving success in the creative industries, come along to one of our Open Days. We can’t wait to see what our alumni do next – and we’d love to welcome you into the ACM community.