Voice of the people

The music industry is increasingly relying on the songs we play to help it discover who the next superstars are. Matthew Hussey explores this people-based revolution

If you want a good example of the term ‘power to the people’, turn on the television and tune in to the channel MBC 1 during the Arab Idol final. The annual talent contest, designed to unearth the Arab-speaking world’s next big singer, is determined almost entirely by the audience who watches it.

In last season’s finale, it was a tightly fought contest between Hazem Shareef of Syria, Haitham Khalailah from Palestine and Majed Al Madani from Saudi Arabia. The panel of judges, made up of singers, producers and composers, found technical errors in all of their performances. But the audience thought that Shareef’s performance was flawless, and promptly voted him the winner.

This model of talent discovery has been sneered at and looked down upon by music’s aristocracy. Paul Weller and Elton John among others have both publicly claimed that the shows are “killing talent”. However, record labels are quietly canvassing public opinion behind the scenes to discover the artists they should sign, where they should tour and who they should work with.

The rise of music-streaming services such as Spotify and the app Shazam, and the social media platform Twitter, have made it even easier for record labels to monitor what songs people like without the need to pay talent scouts to trawl bars and rely on the opinions of a select few. A revolution in music is quietly taking place, and it’s all thanks to you.

We do believe the information we obtain for this new label will provide very useful signals that will bolster our ability to find the stars of tomorrow

Rob Wiesenthal, Warner Music Group Chief Operating Officer

Since its launch in 2000, music-recognition app Shazam has tagged some 12 billion songs, helping people across the globe to discover new music. It has also quietly collected every song people have requested. This information it then uses to tell record labels exactly what you’re listening to.

Shazam can identify which songs are catching on and where before just about anybody else. In 2013, New Zealand-born singer-songwriter Lorde rose to prominence as if from nowhere. However, Shazam’s engineers were able to pinpoint the exact day her success went viral in cities across the United States. At the end of 2014, Shazam even predicted the artists and songs that will be big this year based on what people are ‘Shazaming’ right now.

Warner Music Group has capitalised on all this data by setting up its own record label driven entirely by what you listen to. “While data and crowdsourced analyses will never be a substitute for the expertise and instincts of our A&R professionals, we do believe the information we obtain for this new label will provide very useful signals that will bolster our ability to find the stars of tomorrow,” Warner Music Group’s Chief Operating Officer, Rob Wiesenthal, said in a statement.

“Millions of fans use the Shazam app to discover music, so when a song debuts strongly on the Shazam chart, it is an  exciting, early indication that it has hit potential,” says Mike Caren, President of A&R at Warner Music Group. The deal, announced last year, means that Shazam and Warner will share profits made from artists discovered directly through the app.

This type of context allows artists, managers and labels to better determine how much social activity to expect on a daily basis

Liz Buley, Journalist at Next Big Sound

300 Entertainment, a US record label, has signed a similar deal with Twitter to mine its data for signs of emerging talent. Spotify, too, is using your listening habits to tell record labels things about their artists they could never possibly know. It now gives artists free access to their analytics so they can see which of their songs is played most and where. It lets them push merchandise to avid listeners, and concert promoters have even started to tailor tours according to which towns have the most fans for a prospective artist.

But perhaps the most close examination of what we like online comes from Next Big Sound. The US company, which has penned a deal with Spotify to share its data, monitors Spotify plays, Instagram mentions, Facebook likes, YouTube views and any other platforms where the mass public congregate to learn about what you like. It can now with confidence label almost every recording artist with a digital presence.

“This type of context allows artists, managers and labels to better determine how much social activity to expect on a daily basis, and sets benchmarks for how long an artist will spend in each stage of their career,” says Liv Buli, resident data journalist for online music analyst Next Big Sound.

That means there are huge swathes of artists who lie undiscovered, all in the early phases of what could be careers as big as the Rihannas and Eminems of this world. The company offers this discovery service to record labels for a sizeable six-figure fee. However, while the biggest growth area of data mining has come as the digital realm plays an ever-growing part in our lives, older analogue formats have relied on measurable metrics to determine which songs to play, too.

Shazam was able to exactly pinpoint when New Zealand singer-songwriter Lorde hit a tipping point in the US
Shazam was able to exactly pinpoint when New Zealand singer-songwriter Lorde hit a tipping point in the US

HitPredictor, a US company that works closely with radio stations, has millions of users in its database, which it farms music out to before they’ve had a chance to be heard on the radio. If a song ranks above 65 out of 100, according to its selective audience, they’re considered likely to be a hit. SoundOut, RateTheMusic and Mscore all offer a similar service to radio stations looking to only play the songs that their audiences are going to love, before they’ve heard of them.  

But what does this mean for songs and music more generally, if everything is being influenced by what people click on now? In 2012, a report by the Spanish National Research Council said popular music was becoming increasingly similar.

So are we all ruining music by letting record labels discover what we like? Not exactly. According to David Huron, a musicologist at Ohio State University, at least 90 per cent of the music we listen to we’ve heard before. It’s a phenomenon psychologists are calling the “mere exposure effect”, first studied in the 19th century by the German experimental psychologist Gustav Fechner, and more comprehensively by the American social psychologist Robert Zajonc in the 1960s. Fechner and Zajonc demonstrated that simply exposing subjects to a familiar stimulus led them to rate it more positively than other, similar stimuli, which had not been presented.

The theory doesn’t just apply to music either. Words, paintings, faces, even Chinese characters have all been ranked more positively the more they have been seen by people. In essence, we like songs we’ve heard before. Elizabeth Margulis, a music psychologist, thinks that the mere act of repetition underpins something more profound when we hear music that’s familiar to us. “By just having someone repeat a word to you a number of times over the course of a minute, you lose the semantic association that those sounds normally have,” Margulis told US public radio network NPR. That means our brains are free to explore other aspects of the music we wouldn’t normally have. We can leave ourselves temporarily.

In reality, when record companies mine our personal music choices, they are trying to find songs that are more effective at helping us escape for a few minutes. So the next time someone chastises you for tuning in to Arab Idol, or any other talent show for that matter, you can be safe in the knowledge that you’re simply making music better for everyone. Now that really is revolutionary.