When iTunes launched Radio a few months ago, it was a heart-stopping moment for Tim Westergren, the founder and chief strategy officer of digital radio company, Pandora.com.
Westergren has spent the past 15 years of his life in a nail-biting, white-knuckle ride to create a viable company based on his ground-breaking music product – Pandora – internet radio that plays songs you like before you know that you like them.
It works like this. Pandora uses an algorithm to understand your musical tastes. You start by choosing a song you know and like, and then Pandora searches its database and finds another very similar song for you to try. If you like it, you click thumbs up and you get more along those lines. This way, you discover artists who play music you like even though you have never heard of them before.
Itunes Radio is Pandora’s first head-to-head competitor, and Westergren was very nervous. He has spent millions of his own and other people’s money developing Pandora, which is now listed on the New York Stock Exchange.
At one stage, he had 50 staff working without pay for two years, 11 of his own credit cards maxed to their limits, half a million in debt to friends, family and supporters, and he had pitched his company 348 times to investors without getting a cent.
The 349th time delivered a life-saving US$9 million investment that has seen Pandora grow from 5,000 listeners at its launch in early 2005 to 250 million listeners across three countries today.
Would iTunes Radio spell the end of the dream? Not so far, Westergren told Pandora fans and media at a ‘town hall’ meeting in Melbourne last night.
About 2% of listeners deserted Pandora for iTunes Radio, which uses the same concept as Pandora (although the music genome is Pandora’s own), but Westergren says many have returned. And meanwhile, subscriptions to Pandora’s app are growing.
The Pandora secret
Westergren is touring Australia to help spawn interest in Pandora, which recently became available for Australian listeners and already has an audience of 1.5 million.
It’s an inspring tale. Westergren’s own career as a musician and later a movie score composer spawned the idea for Pandora.
When the romance of being a struggling musician living out of a van wore off in his late 20s, Westergren started working for movie directors. “The job of a film composer is to understand the musical tases of the director,” he explains.
He would take a stack of CDs to his initial brief with a director and play various songs to guage their tastes. “I built an informal kind of taxonomy in my head so that after my interview I could go back and describe what they wanted.”
The dotcom boom was at its peak (late 1999) when Westergren wondered whether this “taxonomy” could be computerised, and with barely an effort, he raised US$1.5 million to test the idea.
The company launched in March 2000, just before the dotcom crash, which decimated thousand of companies and billions of investor dollars.
After a year with a musicology Phd and a few computer engineers, Westergren had analysed 10,000 songs and built a mathematical equation (algorithm) to compute the distance between the attributes of one piece of music compared to another. All 10,000 were loaded into Microsoft Excel and the algorythm encoded in a macro.
The big day came to test their work.
“We decided to start with a Beatles song,” Westergren says. “We hit ‘rank’ and waited with baited breath. After about four minutes it came back with a Bee Gees song. We said, ‘oh fuck, we are a year in and it doesn’t work! It’s flawed!’ But we listened to the song and it sounded just like the Beatles. The Bee Gees, before they were a disco band, were a Beatles sound-alike.”
This, says Westergren, is the strength of Pandora’s algorithm – it doesn’t make recommendations based on popularity, but on musical proximity. “Pandora is blind to the cultural part of music,” he says.
Back from the wilderness
After the years in the financial wilderness, Pandora was finally able to attack the market with its US$9 million in series B venture capital. Within a year, the company launched its computer-based internet radio site which grew exponentially with first thousands and then tens of thousands of fans signing up every day, just through word of mouth.
Then came the company’s big break. It was one of just a few music apps on the very first iteration of the iPhone. “This guy bought us the [first] iPhone in a briefcase handcuffed to his wrist. He looked like he was from the Matrix. We had a specially outfitted room that only our CTO has the key for and they installed a special desk with a bullet-proof plexiglass case. The phone was inside, and a cord connected it to the developers. We were not allowed to touch it, and strict instructions from Steve Jobs himself not to breathe a word about it to anyone.”
The iPhone doubled Pandora’s growth rate, and today, 70% of its audience listen on their phones.
The goal: to steal another 61% of the AM/FM radio market
Westergren’s vision is a bold one. He says 80% of the hours people spend listening to music is still spent listening on radio and that is the market he wants. “We think we can completely change radio. We are 9% of radio in the US today, and we think we can be 70%. That will have two great impacts. It will bring people back to music. But it is also great for musicians.”
Westergren says Pandora will soon allow musicians to communicate with their fans directly, to plot out where they are geographically and plan their tours accordingly, to build their audiences sufficiently to make a living from their music.
“This is the cornerstone of Pandora’s mission,” he says.
In Westergren’s world, the talented musicians currently without audiences will make a living from their music.
Pandora pays royalties – last year writing cheques for US$340 million. In the US, AM/FM stations do not pay to play, but Pandora does. “That is a tidal shift of money into the hands of artists,” he says.
The future
There is still a question mark over that future – the company, which relies on advertising for revenue – is yet to turn a profit. At January 31, 2013, it posted a loss of US $38.2 million – double the loss of the previous year.
But this is content marketing at its best – content of such quality that it spreads by word of mouth and, with the right business model, can deliver a living to musicians, a market to advertisers and a world of undiscovered brilliance to us all.