Culture

This startup finds the music megastars of the future

Instrumental spots hidden gems before they go big. But is it ruining music?

At Instrumental, Conrad Withey has built a new approach to finding musical talent. Instead of relying on human preference, which by its nature is clouded by subjectivity, Withey looks only at data – social media and streaming numbers, transforming a practice that has remained static for nearly a century. The musical landscape is sparkling with “overlooked gems”, Withey argues, because people are looking in the wrong way.

The origin of Withey’s approach lies in video tapes. After graduating from university in 1993, he secured a job at entertainment company PolyGram, developing projects for home video. In this position, he learned an “audience-driven way of thinking,” which is to say he identified audiences by analysing viewing figures across television, stand-up comedy and theatre, and targeted them with a VHS spin-off product.

In music, it usually works the other way around – as Withey learned when he began working at Warner Music in 2006. Instead of capitalising on existing audiences, the industry has traditionally relied on artists and repertoire (A&R) executives to spot talent by listening to it. They’ll then spend big money to develop a fanbase. Although Withey scored some major hits, his colleagues never bought into his “counter-programming” approach.

So, in 2013, he launched Instrumental with his wife, Abi Hanna, a former marketing director at Penguin Books. With the help of data scientists, he designed a bespoke system that he says can pinpoint the moment an artist becomes meaningful to the wider music industry. The upsurge in digital music, social media and YouTube, where many aspiring artists were uploading their work, gave him new datapoints to identify his ready-made audiences. Where labels saw a teenager singing covers in their bedrooms, he saw 500,000 subscribers.

By looking at track uploads and playlists, Instrumental creates artist profiles that it enriches with data from social media. It then tracks the artist and ranks them against others, giving them an “Instrumental score”, which is continually tuned through machine learning. At first, Withey envisaged making his money by licensing the tool to anyone in music who wanted an edge over their competitors.

The initial results were “astounding”, Withey recalls. Among the first successes was singer-songwriter Calum Scott, who signed to Capitol Records in 2016. The former Britain’s Got Talent contestant was topping the Instrumental charts because of the speed with which he was attracting subscribers to his YouTube channel after uploading his cover of Robyn's “Dancing on My Own”. When Withey encouraged him to release the song officially, it sold over 600,000 copies. The signs were all there in the data, Withey says.

Instrumental also flagged Arizona Zervas, nearly two years before Columbia signed the American rapper. Lil Nas X was on Withey’s radar months before Columbia came calling. Withey knew of Tones and I long before the Australian-born artist signed to Elektra Records. “Each week it [Instrumental] finds thousands more artists with huge potential,” Withey says, “opening up a world of talent that you wouldn’t otherwise be aware of.”

In 2018, Withey changed tack, launching his own label rather than licensing Instrumental to others. Whereas traditional labels tend to be style-specific, frtyfve works cross-genre, and its entire catalogue has been found through Instrumental’s technology. “We’re attracted to music with an audience, whatever it is they’re doing,” Withey says. He’s even signed musicians without having ever heard them.

While Instrumental can find music that’s popular, it’s not always “good”, Withey admits. And because it looks at online engagement, the platform supports internet stars rather than musicians, and so it’s often perceived as a threat to musical artistry.

Withey, however, sees his work as an upgrade to notoriously inefficient methods of talent discovery. According to the International Federation of the Phonograph Industry, record labels spent $4.1 billion on A&R in 2017, but their commercial success rate is as low as one in ten. This wasn’t as important when labels could afford to spend vast sums on talent, but now data-driven insights can improve results and minimise costs. Withey sees Instrumental as a tool to support A&R executives, not replace them: with more than 40,000 tracks added to Spotify each day, it’s impossible to keep abreast of these new creators, let alone identify those with the most potential. “The lovely thing about closing your ears and eyes, and being led by data, is that you might discover the next big thing,” he says.

More great stories from WIRED

This article was originally published by WIRED UK