Chris Barton may not have revolutionized the digital music industry, but he ended up forever with a chronic problem that afflicted around 99% of the inhabitants of this planet: listening to a song and not remembering (or not knowing) the name of the song and who is playing it. singing.
Ever since the iPhone came out 15 years ago, just download the Shazam app on your smartphone and the magic will happen. Point your cell phone at the place where a song is coming from and it will tell you who sings it and its name too.
A few weeks ago, the international speaker agency Ideas Collection brought Barton to Brazil to speak at a corporate event. Before going on stage, he spoke to the Sheet.
How was Shazam created? Well, I’ve always had a problem in my life, which was this: I’d listen to songs and I’d be like, “Wow, what song is that?” (laughs). I asked others and kept a list of those favorites. Eventually, I would make a CD of them all later. So, in 1999, I had an idea to create a service that would recognize the music you’re listening to.
Was this new? Not exactly. At the time, I was between San Francisco and London. I was doing my MBA in Berkeley, California, when I had the opportunity to study abroad and I chose the London Business School. That’s where I got the idea for Shazam, in 1999. As I researched, companies came up that were following a similar path. There were a few startups and at least one big company, Sony. Sony had a little device where you clicked to find the name of the song that was playing on the radio. But figuring out what was playing on the radio was an easy problem to solve. Popular radios, together, maybe play 200 thousand songs, very little compared to the 50 million songs you have today on a streaming service. The companies monitored the stations and, when someone asked what was playing, they provided the information that was already tracked, according to the station and the time.
And the cell phone? None of these companies were using cell phones at that time. When I found out about this method, I forced my mind to a different way of thinking and came to the following: “what if I could identify the song by the sound in the air, reaching the cell phone microphone?”. And that was the innovative idea. Nobody was trying it and everybody said it would be impossible.
It’s 1999 and there were still no smartphones, right? Not. They were just phones…mobiles. I had one and all I could do was call and text. I mean, the most advanced thing to do was download a ringtone. Yet two years before that, less than a third of people were supposed to have cell phones. But in 1999, probably 90% of people [nos Estados Unidos] it already had.
This is long before the app we download on our phone, right? Many years before, like eight years before the AppStore. And seven years before the iPhone.
Why did they say it was impossible? There were two major challenges. One was the size of the base to be checked. Let’s think visually rather than aurally. If you have a brother next door and I said “where’s your brother?”, you would say “over there”, because there aren’t many people here. But if we were at Wembley Stadium and I said “where’s your brother?” you would have a hard time, as there would be thousands of people there. So that’s the scale.
And the other? The second problem was the noise. Again, if I said “where’s your brother” and there were people running around in front of you, it would be really hard to see him, right? So we needed to track millions of songs with noises, conversations and various background sounds. That’s why it was so difficult. When we approached professors at MIT and Stanford, they said “there is no technology capable of doing this”. And they had no idea how to invent it.
What was the biggest challenge? The problem was the pattern to create a technology, a recognition pattern that could successfully identify a song against a huge database. There was also software that reduced the sound quality on the cell phone, believe it or not. There are systems that purposely emphasize human voices and drown out all other sounds to make the connection better, which is a bad point for music. So there were a lot of challenges.
The recognition of the songs, how was it done? We took a kind of fingerprint of each song. It’s a mathematical description. Sounds become numbers, but numbers that describe sounds. That was our invention because what we made was a mathematical map, a three-dimensional X, Y, Z-style graph that recognized each track.
Is none the same as the other? Is that why you use the expression “fingerprint”? That. It’s a fingerprint. It would work perfectly with a movie, a TV show or a television advertisement. I could find those things too.
What was Shazam like when it was released? That was in 2002, and our first working model was a service where the user would call us, point the cell phone microphone at the music and hang up after a few seconds. Moments later, he would receive a text with the name of the song and artist. We charged something like fifty cents per call.
What happened when the iPhone came out, five years after that? At the time, Apple dominated digital music, 90% of all digital music was on iTunes. So we thought, “We need to partner with Apple because we identify the songs and Apple has this great phone and they sell digital music.” We were already in conversation with Apple even before the AppStore was launched. And Shazam was selected to be one of the original apps at the AppStore launch. They even did a TV ad that was all about Shazam, showing our app as a way to sell iPhones.
Did you go to work at Google when Shazam was being used on millions of cell phones around the world? I wasn’t making any money from Shazam and startups like this can always make mistakes and disappear overnight. So, yes, I had a full-time job at Google, and then Dropbox. I only spent 10-15 hours of my week on Shazam.
But five years ago you sold Shazam to Apple for $400 million. Yes, it was in 2018. I don’t work there anymore. The first thing Apple does when it buys a company is dissolve the board. All board members are withdrawn. I’d rather not reveal how much of the company I owned, but I’ll tell you it wasn’t much. Basically, we had raised $140 million dollars from investors, which was spent. And when all this money is collected, what happens in the company is the dilution, dilution and further dilution of the percentage of partners.
So you’re not a billionaire? No [risos]. I didn’t even make tens of millions. I don’t complain, life is more than just money. And I’m still happy with the result.
Today it seems like you already have a new idea in mind, right? Yup. I’m developing an artificial intelligence device called Guard, which will work with cameras inside swimming pools and will basically be a lifesaver, sounding an alarm if someone is drowning. In the United States, this is the leading cause of accidental death in the 0-5 age group, more common than car accidents. Of course the Guard doesn’t save the person directly, that would be an even bigger problem, having a device that would jump out of the pool and save the person. If the person is unlucky enough to be drowning and they are the only person at home, it is no use. But most drownings happen when someone is at home, especially children.
Are you starting from scratch again? Yup. I’m not looking for investors yet, but I will. I learned a few lessons from Shazam, and one of them is “don’t raise too much money”. Many entrepreneurs think that getting money is the solution, and sometimes getting the money is the problem because if you have the money, you spend it to achieve what you set out to do. So I always think about how to spend as little money as possible. I didn’t know this when I started Shazam and now I have great respect for it.