Barton, Inghelbrecht, Mukherjee, and Wang
Shazam

Chris Barton, co-founder and former CEO of Shazam, is currently at Dropbox but spent almost eight years at Google prior to Dropbox. He and Philip Inghelbrecht graduated together from the MBA program at UC Berkeley. He also holds an M.Phil. from Cambridge University.

Philip Inghelbrecht, co-founder of Shazam, is currently head of business development at RockMelt, Inc. He is also the founder of Road Hero, Inc., which he started in 2010. When leaving Shazam, Inghelbrecht worked at YouTube/Google and was later also president, and instrumental in the development of, TrueCar, a publisher of industry data in the auto retail market.

Dhiraj Mukherjee, co-founder of Shazam, is currently a Social Media Strategy consultant at Infosys in London. Previously he was Digital Director in charge of Innovation at Bauer Media; a digital consultant with Viant where he started the London office, and a strategy consultant with Bain & Company. He has a BA from Dartmouth College and an MBA from Stanford Business School.

Avery Wang, co-founder and chief scientist of Shazam, has over 20 years of industry experience designing high-performance multimedia signal processing systems for the consumer market and is the principal inventor of Shazam's recognition algorithms and other key technologies. Avery holds graduate degrees in Electrical Engineering and Mathematics from Stanford University, with a PhD at CCRMA on Auditory Source Separation. He also studied Computational Neurosciences on a Fulbright scholarship in Germany. At Shazam his responsibilities include innovation and intellectual property.

Pedro Santos: How did you start from an idea to the actual company? How did you evolve?

Philip Inghelbrecht: This is kind of the sequence of events. There are four founders in Shazam and this is how it all came together. Chris Barton and I entered the MBA class of 2000 at UC Berkeley. In August 1998 we started our MBA in Berkeley and that was when the internet was hotter than ever. It was the dot-com boom at its best and it was very fashionable, if not almost required, that upon graduation you would turn down big job offers because you were going to start your own company.

So, in a strange turn of events, and this shows how coincidental the world can be, in an MBA you have mandatory classes, but if you have taken that class and worked in that area and show you are proficient in it, you can take a quick exam before the class starts and possibly waive it. I had an investment banking background and Chris had a Masters in Finance from Cambridge so we both did that exam and so we both did not have to take that class in our first year and we could choose any other class. The only other class available, and this is ironic, was Advanced Finance. So, on the first day of school, Chris and I found ourselves sitting next to each other, being the only first-year students amongst second-year students. So, in the class everybody knew each other because they had studied together for a year and of course Chris and I were brand-new, so we found each other on the class benches and became teammates on the same project in that particular class and also we became very good friends quickly. It was not long before we decided to start a company together.

We were not sure what it was going to be… It's kind of in our DNA to start companies and do something new and exciting. We had lots of ideas—some of them were actually good and I regret we didn't pursue them, some of them were out of control and I would be too embarrassed to share them… So, in the course of 1998 and early 1999 we just brainstormed but didn't actually do anything, it was all talk. Fast forward, as we approached late 1999 and we had only six months to go until the end of our MBA, we said to ourselves we had to make a decision and to go for something. You can't pursue five companies at the same time—we had to pick and focus. So, the idea of Shazam was the one we went for. It was actually Chris's idea and after some conversation we decided that it was the one we were going to try it out.

Santos: I would like to ask Chris. Can you tell us about some of these ideas and why you actually decided to go through with Shazam?

Chris Barton: Yeah, I think I was brainstorming. There were a bunch of ideas. Among the top two terrible ideas that preceded Shazam. One was, sadly enough—I don't even know if Philip and Dhiraj know about this—to do with selling contact lenses on the web. I thought it was a big, high margin opportunity. But it just seemed so boring that I quickly dropped that one.

Inghelbrecht: I remember that one, yeah.

Barton: Then the second one was one that some of my friends jokingly called “e stalker,” which was an idea where I wanted to introduce the concept of star power to web sites. As you know, a $100 million movie can spend $30 million having Tom Cruise in the movie. The idea is that movie stars would allow themselves to be tracked when they go to certain internet sites. Then if you were on Amazon and they were on Amazon, it would say, “Tom Cruise is on Amazon now, Madonna is on Amazon now,” that type of thing. Then it was selling that essentially to web sites. But anyway, that was also a crazy idea.

Dhiraj Mukherjee: Yes. Chris is still talking about it. [Laughter]

Barton: Yes. Then came the original incarnation of the Shazam idea. My thought was to identify music by monitoring radio stations. I was thinking, okay, what I would do is build software to help radio stations track all their programming and play lists. Then that would be useful to them. Then we would have exclusive ability to monitor what songs they were playing. Then with that, you could provide a Shazam like service but of course it would only work for radio play.

So the final tipping point for the Shazam idea was that I was taking a course at London Business School called Strategic Innovation. They encourage you to really think outside the box, so I had to come up with this idea from every angle. And I thought, “Well, if I built this exclusive network to radio stations and I was the only one who knew what they were playing, what could someone do that would get around me, even if I had that exclusivity?” It was at that point I thought, “Oh, my God. What if you could just know what they're playing using the actual sound heard by the phone?”

Then you wouldn't even need this exclusive deal with the radio station, with the exclusive software. You would literally just be able to do it with the sound. I thought, obviously no one had ever done that.

The first thing I did was I called my dad because he's a nuclear physicist. I mean, he's not an expert in this area, but he's a scientist. I bounced the idea off him and just said, “Do you think that sounds like it could be feasible?”

Santos: Why did it take three years to put this together?

Inghelbrecht: Shazam is not your traditional start-up. I am not sure how I would explain it. If you think about it, most technology start-ups begin with the invention of a great technology. So, it is about how you can commercialize it and create a market, and make money from it of course.

We were thinking, “How cool that would it be if you could just recognize the song by holding up your phone?” But the music recognition technology out there but by no means was it fit for that particular scenario. We had to recognize ten seconds of a very noisy audio and in the matter of milliseconds out of the database with millions of tracks. So, we said: hey, what a great concept, let's go build the company around it. That's not how you start companies unless you are in biotech or something like that where you have ten years of clinical trials. So going a little bit back, it's late '99, there is Chris and myself, and by that time we had roped in Dhiraj Mukherjee. Dhiraj had set up the Viant office in London. So, the three of us sat on this great idea, but we really didn't have anything.

We didn't even have a business plan and most importantly we didn't have the implementation of the technology. The first few months of Shazam were spent trying to find the fourth person—the technologist, the engineer, who could implement this and “crack the code” on music recognition in a noisy environment. We literally went around the world trying to find that person. We would go to Xerox PARC, MIT Media Lab, Stanford University Center for Computer Research in Music and Acoustics. In all those places we would describe the problem and ask if they could build this. Most of the time we would be laughed away. People would say, “Ha, ha, ha, it's not possible. You boys go take a job at McKinsey or something when you graduate.”

And we just kept going at it. People would say, “Very intriguing, very challenging but… yes, I guess you could build it, but it would require a computer park that every single search would cost you a dollar or two dollars and you can't build a business around it.” That was the first three months of the company, trying to find that person.

Santos: What was the main thing that made you pursue Avery out of all the others?

Barton: Well, what we did is a bunch of research on the web. We found people who had published articles in digital signal processing. Almost all of them came from either MIT or Stanford. We then had a list. Philip and I had a list of probably about forty of these people or something like that. We then found Julius Smith. He was a professor at Stanford. Many of the other people on the list were PhD students or recent graduates.

We tried to get to him a couple of different times. He actually ignored us the first few times. Then eventually, because we had another advisor from Berkeley named Dan Ellis… By having Dan Ellis as an advisor, that kind of opened the door with Julius Smith.

He finally said, “Well, if Dan Ellis is involved in your project, then, sure, I'll talk to you.” So we went and met with Julius Smith in his living room in Stanford in Palo Alto. He loved the idea. He said that he didn't know of any technology that could do this, at the time, but he felt like it could be feasible to invent it.

Then in our second meeting with him, I brought the list of forty people that we had researched, who were all PhDs in electrical engineering, focused on digital signal processing. He knew almost all of them because it's a small community.

I basically said, “Julius, I want you to rank the five smartest people on this list. I want you to rank it not only for their ability in acoustics signal processing. Because we actually have to invent something, we don't have to just engineer something, but we have to actually invent something. They need to be deeply theoretical at a mathematical and statistical level. So they need to be very strong in mathematics and statistics at the theory level, essentially.”

Then finally, I wanted them to also have the ability to code and execute, because we're a start up and we can't afford to have coders as well.

Based on all those criteria, he ranked the five smartest and most recommended people. Avery was the number-one ranked on that list. There was another guy who was at Liquid Audio at the time, I think, and I can't remember who the other three or four were. But anyway, yeah, we met with Avery and liked Avery, and thought that he could invent the solution, even though he thought it was impossible.

Inghelbrecht: Sometimes you need luck along the way so, Avery was actually living in Palo Alto at that time. His own company wasn't really moving anywhere. So, we just went after Avery. Pitched him on the idea, on the concept and promised to make him the fourth fully-fledged co-founder. If he was going to do this, his contribution was going to be as significant as ours. That is when we finally had the team together. So, the first five to six months were spent trying to find and build the team.

Santos: Avery, what were your thoughts when they came up with the idea for you?

Avery Wang: From my perspective, I got an e-mail from Chris. Initially, I said, “Oh, whatever, some random person” and I ignored it. One thing about Chris is he's incredibly persistent and basically doesn't take no for an answer. Then he tracked me down and forced me to go to lunch with him at a cafe in Palo Alto, Cafe Brioche, which is now one of our favorite places, and kind of gave me a pitch about what he had in mind.

Philip was there too. So I thought, “Well, okay. It's sort of a crazy idea and I have no idea how to do this but I'll give it a try.” Then I started out and I started working with Julius Smith.

He's actually one of the top researchers in digital music and audio signal processing. So we started thinking about how to do it and tossed around a bunch of ideas and for actually several months didn't really have a solution.

While this was going on, Chris was saying, “Okay, we need to have an algorithm by this date and raise money by that date, otherwise we're going to start running out of money and we'll have to find other jobs.” [Laughs] Get a real job.

Santos: When did you actually come to the conclusion that you have a working solution or you have an idea?

Wang: That's kind of interesting. I was working on this and not really getting anywhere. The guys had already moved to London where we decided the business would be based. I was saying, “We need audio recordings.” So it took a while but the people in London got me some sample recordings. There's a whole story behind the London crew as well, which I probably will defer to one of the other guys, possibly Dhiraj.

Anyway, so I'd been working on this for a couple of months and then I heard the first audio recordings. Understanding how the service is supposed to work as it was envisioned by the other three co founders, the audio was being recorded in a noisy environment with the music being in the background and going through a tiny little microphone on a mobile phone which meant it also went through voice compression.

So by the time we got the recording on the other end of the phone, you couldn't really hear much music in most of the cases. I was in despair. I was starting to think how to break the news gently to my co founders that this was actually indeed impossible just like all the other professors had been telling them.

Then Chris was following his timeline and saying “Okay, you need to get this done.” I think we were up against the deadline in our project plan. Then he went to Croatia, I believe, on vacation.

Barton: That's right.

Wang: During that time, there was one afternoon where I was sitting in a café. Too much coffee and probably spending half my thoughts on how the other guys would feel after I let them down. [Laughs] I was staring at some graphs when I realized that I had the solution. There is a certain trace and a scatter plot of matching audio fingerprints. I could go into that in more detail, but just leave it at that. Then I realized that it was a very strong statistical indicator of being able to detect music. I ran some examples and found that it was actually able to recognize music even in the very noisy cases where you couldn't actually hear music in the recording with your own ears.

Then Chris insisted that I fly immediately to London and help out with the fundraising. The good news is we were still essentially on Chris's project plan. We weren't really slipping schedule. We had essentially implemented the technology basically just in time so we could go and raise money.

Santos: Impressive.

Mukherjee: I polished off the business plan during that week while Chris was away and he came back and he said, “Wow! I should go away more often,” those were his exact words. [Laughter]

Avery's work was much harder, but now we could get into the fundraising.

Santos: Just one question to Avery. Did you actually realize when you came up with that solution how good a solution it was?

Wang: It wasn't quite in my mind at that moment, I think. I realized that there were certain properties, advantages and disadvantages, of the solution and that it was largely focused on exact recordings, but I realized at that point that it could scale dramatically. I don't know how much detail you want to go into.

We knew it had to be able to scale to millions of recordings in a database and be able to handle perhaps hundreds, if not thousands, of recognitions per second without using some huge amount of servers.

And also be able to handle a large amount of noise from the environment. So that was a pretty tall order. But when I saw that graph, I realized while thinking almost all those thoughts within a few minutes, that this would actually solve all those problems.

Santos: I suppose this phase was all bootstrapped.

Inghelbrecht: We didn't really spend money. Our only expenses were phone calls and a little bit of travel. We couldn't build anything, we were not engineers. I have to be honest with you, Dhiraj was really busy in London, he had a real job, but Chris and I were in our last semester in business school and we had plenty of time to kill. That last semester is either for partying or starting your own company. It is the time when school doesn't really matter anymore. So, as Avery mentioned, he comes on board and we literally wrote the project plan.

Santos: Slightly optimistic…

Inghelbrecht: You can't do this, this is just ridiculous. But Avery came through. Surely the first incarnation of the technology is nowhere close to what it is today. It wasn't as scalable, it wasn't as accurate, you name it, but the cornerstone was laid. And it was enough for us to start raising angel money, build a team and start building a real company.

If you think about the Shazam service in the early days, the technology, the algorithm, was a very critical part, but it was only one out of three. Because you don't only have a great algorithm, but you also must have a database of music to match against. You have to look at things in perspective. In late 2000, there was Napster, but people were buying CDs, there were no iTunes stores. Digital music was very new. There wasn't much out there. So, we had to go and build overnight so to speak the largest database of music fingerprints ever built. We had one chance.

The other one, for this to work if we were going to launch in the UK on day number one, was to make the service available on all the UK operators. There were four of them at that time. And we had to have with all of them a single short dial code like 2580, because we couldn't have multiple numbers, we had to have a system—what is called a reverse SMS charge. Once it has recognized the song and the text message was sent to your phone with the name of the track and the artist as a consumer, you would pay to receive that text message.

So, we had to force their technology in that way and convince all those operators, all at the same time that they would share that revenue back with us. I mean this is early 2000. None of these three things that I've have just mentioned—the algorithm, the music database, and the implementation with the operators—had been done by anybody before. I am still mesmerized and happy how some of the investors, angels and VCs believed that we were going to put all these pieces together and orchestrate it in one giant launch in 2002.

And we did it, we came through. Ten years ago you could start a company with a PowerPoint presentation. Eight years ago, with a PowerPoint presentation and maybe a prototype and maybe we were at that area and we had that. Three years ago you could start it with beta product or a working product. Nowadays to raise money, you'd better not only have a working product, but you'd better demonstrate traction that you have—like one million users. Otherwise you don't raise money. Shazam, early on, was very capital intensive. We had to crack a few nuts that were never cracked before. And it took time.

Santos: That leads me to a double question. One, you were all in the US, well three of the four founders were in the US, and you launched in the UK. What made you make this decision? Where did you raise the business angel funds—in the US or in the UK?

Inghelbrecht: The company was always meant to operate in the UK. There were a few reasons for that. In 2000 you couldn't even send an SMS cross-network in the US yet. There was no “premium charge SMS” in the US. And our whole business was predicated on sending SMSs back to the end user, not to mention charging for it. That infrastructure was simply not available in the US mobile market. That is one. Also, the music market per capita in the UK is higher than anywhere in the world, actually. At least back then higher than Japan and higher than the US. So, that's a good mix as well. So, those were two important details. Finally, investors in Europe understood consumer mobile businesses at the time whereas the US investors were not there yet. Besides we thought it would be fun to live in London for a few years.

Now, Avery had a family. Dhiraj, Chris, and myself, our opportunity cost was low. Dhiraj had a nicely paid job. Chris and I never took a job after graduation, so we never tasted what it is to make a $100,000 or $150,000. So, for us, we didn't really care what we did and where we moved, as long as we had fun along the way. For Avery, of course, it was a little bit more difficult. We had to appreciate and acknowledge that. From day number one it was well understood that Avery would stay in Palo Alto and it made sense back in the day, because he was really focused on the algorithm and as a deep theorist and truly an inventor, some of his best ideas are generated at three o'clock in the morning. We didn't need him to stay in the office. Besides he bounced off a lot of his thoughts and ideas with Julius Smith, his professor at Stanford. So it was better to have them both together than across different time zones. So, it is weird, but it was actually the best for the company.

Santos: It is a very unusual mix.

Inghelbrecht: Very unusual and I'm not sure if I can recommend it to anybody today.

Santos: Well, now on to you, Dhiraj. You were the only one in the beginning in London.

Mukherjee: Yes.

Santos: So how did you connect to the rest of the team? What's the beginning from your side?

Mukherjee: Sure. Chris and I were good friends. When I lived in San Francisco in the ‘90s, he used to live in San Francisco as well. Then I was in London in the summer of '99 and Chris was there as well. He was an intern that summer and I had a job at Viant. We hung out and talked about starting a company together, but we didn't have one specific idea in mind. Chris always said he had this great friend, Philip, who was his classmate at business school. He wanted to involve Philip in whatever we did together.

So we basically agreed to start a company together, but we weren't quite sure what. I continued my job, and Chris was back at business school when he came up with the idea for Shazam and we talked it over on the phone and agreed we would go for that one.

So as it happens, it worked well to be based in London, as Philip mentioned, because the mobile market was much more advanced in Europe than in the US at the time—the usage and the capabilities. It was just coincidence that we were together in London at that time, but the location actually fit the business. That's why after Chris and Philip finished business school, they moved out to London.

It was lucky that—you heard the story from Avery—Chris and Philip were actually in the Bay Area and that's in fact where Avery was, so it was just a lucky combination of circumstances.

Santos: While Avery was developing the actual solution, you guys moved to the UK and you started trying to get all the other pieces together. What was the first challenge that you addressed beyond the actual solution?

Barton: Without a doubt the challenge was attracting capital. Essentially, it's that whole thing that to attract capital, you have to have addressed the key concerns of the people that are looking to invest. One of the key concerns was obviously getting interest from mobile operators. We had a lot of meetings with mobile operators and the mobile operators were unwilling to jump on things at such a nascent stage given we could disappear overnight.

We actually focused most of our efforts so that our initial set of investors were angel investors. We decided that we wanted to raise angel money, but we also decided that we wanted the money that we attracted to also validate the business. So we wanted high profile angel investors. Specifically, we just decided to go after these famous angel investors.

We ended up getting the former CEO and chairman of EMI, one of the major record labels. The former chairman of BMG, one of the major record labels. The founder of Amazon.com in Europe, amazon.co.uk. The inventor of the 56K modem. The former chief technology officer of British Telecom. All these different guys. The founder of Liquid Audio. All these people became angel investors.

That was a big hurdle for us. We wanted them. We wanted their names behind it because we knew it would help the VCs get more excited about us.

We also needed the money because we had no money at all. That became our primary focus immediately after inventing the technology and filing the patent.

Santos: You're a small start up with the technology completely brand-new and an unproven business model. How do you convince these top business angels to get on board and support you?

Barton: I can jump in with a really simple answer on that. It's all thanks to Avery. Because Avery built this demo. When you saw this demo at the time—I mean, now, music recognition has become mainstream. But we had something at the time that you could show people that they had never thought to be possible. Show this little demo and it literally was the thing—combined with a PowerPoint, of course—that just won people over, including the venture capitalists that eventually invested.

Literally, in this demo, you would play a little audio clip that was recorded over a mobile phone. So we'd have that sound snippet that we had captured over a mobile phone. Then we'd run it and it would identify that song essentially against a database of songs.

We got to the point where we could do it actually with a limited number of songs right there, with a demo mobile phone. When people saw that, they just thought, “Wow.” It was like magic. That alone basically sold them.

Wang: I don't want to take all of the credit, because a lot of the momentum was starting already before I had actually implemented the technology. My fellow co-founders had already been leveraging their contacts. They're fairly well connected in knowing a lot of people who are in various strategic places. Then it was basically an exercise in leveraging credibility.

You've got one guy who's credible. Then another guy saying, “Oh, if he's involved, maybe you're interesting.” Then building that up and basically bootstrapping credibility, I guess that probably worked to some extent. But at the end of the day, you have to have something to show.

So it's certainly not just the technological innovation that built the company.

Mukherjee: This was a voyage of discovery. So, we didn't even know what kind of technologist we were looking for. We had to first discover that digital signal processing was the right field. And then we could track down Avery. Similarly, as he has described with the fundraising, you'd better have all of the different parts of the story come together. A big piece that we didn't know about was the music database.

Santos: So, you raised the angel money, you managed to have the algorithm done. How did you manage to get the database?

Inghelbrecht: Here, I can't give you the full details. Some of this is still very active and confidential. But back in the day, a company called Entertainment UK was the largest wholesaler of music in all the United Kingdom. If anybody had music in their warehouse, it was them. So, we started with Entertainment UK. It is important to realize that if you look at Shazam, we don't really need the music itself. To recognize music you don't need full digital copies.

We help the music industry… there is nothing more promotional than Shazam.

Santos: It is free promotion for the artist.

Inghelbrecht: It is proven. This is how we built the music database in the early days. And these were critical deals.

Barton: There really were no digital music databases at the time that we could access, so we had to start from scratch. The other part of that is, in order to get that database, we needed to get the CDs because that was the way it existed. We wanted to avoid buying one hundred thousand CDs because you're looking at a couple of million bucks.

So, Philip was able to go do the world's best start up deal in history. He essentially went and did the deal with Entertainment UK. Entertainment UK was the biggest wholesale distributor of CDs in the UK. The largest record store in the UK was Virgin Megastore, with about fifteen thousand or twenty thousand unique CDs. And Entertainment UK had one hundred thousand unique CDs.

Philip's deal basically said that you let us co locate on site. We're going to pick your CDs off your shelves. We're going to create this database using your CDs. Entertainment UK was able to keep a copy of the database of metadata for their own purposes. We did that deal and co created this database with Entertainment UK, and it saved huge amounts of money.

We also had twenty people on a seven days/twenty-four-hour basis, keying in all the metadata—the song titles, album titles, track lengths, you name it. Because you have to put yourself back in time—this was the early 2000s. The name of the game was to sell CDs online. The only way you could sell CDs online is if actually you could describe those CDs, right? And so Entertainment UK didn't have that metadata. And so we keyed it in and gave it to them.

Santos: And the operators… did they receive you well or was it like, “Mmm, this small company is wanting to strike a deal with us?”

Inghelbrecht: We definitely had to push them because, if you think about it, we had to get a reverse SMS charge and a revenue share. Chris has a long background in the mobile industry and he really positioned and pitched it better than anybody else where he demonstrated that this was some kind of new data revenue against voice that was going to become a commodity.

And he pitched it so well that all the operators worked with us. They wrote us letters of intent so that we would have something to show for when we went to Series A fund-raising. I make it sound easy, but putting it all together and signing documents was hard work. But they were definitely cooperative. Even the recording labels were cooperative. Here is the company that hasn't launched its product yet and has nothing really to show for it…

If you are in the recording industry the best thing you can do is sit back and relax and let the CDs sell themselves and cash in the money. Don't try to fix something that isn't broken. Don't move too fast with new distribution opportunities or radical business models. So, they really didn't care much about it. So, when talking to the record labels about this, I think we had three types of response. We had “okayish.” And then people who went out of the way quickly, most people. The majority didn't care—there was like one meeting and nothing would happen. And occasionally we would run into those I call the eternal optimists.

I will never forget the meeting with [an executive at] BMG. So, we wanted to demonstrate technology to him, so the next thing he does … we had a prototype on our phone, he puts in one of his CDs, he plays the music, let's say it was Britney Spears, and we tagged the song and the answer was that it was Christina Aguilera. That's the worst that can happen. So, it's his music and not only did we miss it, it's the worst false positive. Despite all that, he stood up and said “Wow, this is amazing. I don't care if it gets it or not, but it is still amazing.”

Santos: You need people like this in the beginning because the company is still quite fragile.

Inghelbrecht: You do, you look at some of our advisors—they were all great people. These were people who truly believed in it.

Santos: And that makes them great advisors. They have the optimism that you sometimes need.

Inghelbrecht: It reinforces your vision. When you put all these deals together… until then you didn't raise any capital … once you had the deals with all the three components, you had the prototype, the database and you had the operators interested and then you went to raise cash.

We raised a total of $1 million in angel money. In July 2001, we raised our Series A led by IDG Ventures Europe and two other VCs in Europe. So, one million in angel and $7.5 million from VCs.

Barton: Just a quick story about a VC meeting. We were a teeny little start-up with very little money, and so we rented a tiny little box like office that was in a very cheap area. But we also wanted to be in a hip area. And so if you want hip and you want cheap in London, that puts you in Soho. The shady part of Soho. And so that's where our office was and our neighborhood had a little bit of a red-light district type of atmosphere.

Well, it turns out, sadly, there was, I guess, a murder that occurred, actually, in a store that was directly below us in our building. So anyway, it just happened that the next day, we had scheduled some venture capitalists to come by and meet us at our office so we could pitch to try to raise money from them.

Normally, we would visit venture capitalists in their offices, but I think these ones were visiting from Belgium. They came up to ring our door buzzer and as they pushed the buzzer, some policemen approached them and said, “Do you know anything about the murder?” They were bewildered. [Laughter]

And the police were serious. So … the first impression of these venture capitalists is like, “My God. What kind of company is this?”

Santos: How did that meeting end, actually?

Barton: I'm trying to remember. Do you remember? Was that FLV?

Mukherjee: It was FLV and they invested in us.

Barton: Yeah, they invested. That's hilarious.

Santos: Any more VC stories before we move on?

Barton: Well actually, I have another one. Shazam was working very well before in beta testing before release, except it seemed when tested by our venture capitalist, who was very nervous indeed. He kept trying it. And for him and him alone, when he tagged songs, it kept coming back always saying that it was “The Pogues” no matter which band he tagged.

Avery dug and dug and dug, trying to figure out why this was. And then eventually we figured out it was because he was so eager to make it work that when he was tagging songs, he was putting his mobile phone literally right next to the speaker, so close to the speaker that it created a feedback sound. And that feedback sound was identical to the sound that was in this particular song by The Pogues.

Santos: So you solved the database problem. You had funding. You had the algorithm. You still have to convince the music industry. Did you have any problems or issues with the music industry?

Barton: Obviously, we were always hoping that the music industry would embrace us, they would promote Shazam, and you name it.

In the early days, the record label executives would laugh us away. Obviously, that has now changed. If you're a music industry executive and you don't use Shazam, you're toast. You know what I mean? So in the early days, it was very hard to get traction with anybody in the music industry.

Mukherjee: We've always wanted to work with the music industry, and Shazam, to this day, does do promotions with the music industry. We were not reliant on the music industry in the way that other music businesses are that want to sell music or play music.

Santos: And the telecom operators. How did you convince them to incorporate this new service into their networks?

Barton: A lot of persistent phone calls. [Laughs] So the good news is that there were really only two things that we were reliant on the telecom companies for. One of them was a standard product. So the two things that we needed from them for an initial launch were premium SMS charging— that was a standard product. Almost any company can sign up and get premium SMS charging. That's how we do it. We send the SMS to the user with the name of the song, and then the user would also receive a charge on their phone bill of fifty pence. Then we would get a revenue share on that. The really, really challenging thing, and that was the second thing we wanted, was not a standard product. We had to convince them to do something that they did not do with any other companies. And that was to give us the four-digit phone number or “short code.”

We didn't want to have a full, long phone number. We wanted to have a number like 911, 411, or 192. One of these short digit numbers. The one we went for was 2580, which happens to be the only four digits that went right down the middle of the phone, so it had a nice geometric, memorable component to it. It was recommended to us by a talented user experience guy who Dhiraj knew. We had to convince them to give us this four-digit number and that was actually very, very challenging. We had many, many meetings and lots of resistance over many months.

Ultimately, our strategy was basically to try to get one, and once you get one, then you get two. And once you get two, then the third one—and they feel like they don't want to be the only one missing out. We tried to hype it up and say this is going to be a really cool service. You don't want to be the only mobile out here that doesn't have it.

Santos: Why was this four-digit number so important?

Barton: First of all, we had to have one consistent number across all mobile operators. So that was one thing that was critical. Otherwise, marketing would be a disaster. Then secondly, if you don't have a four-digit number, then that means you have to have a full phone number and a full phone number probably would have various charges associated with it, depending on where you call from in the UK, and it would also be very, very hard for people to remember it, and so it would be very hard for marketing. So that was the reason.

Santos: And this was all in preparation for the launch in 2002. How did it feel? How was it to each one of you to actually experience it?

Mukherjee: Well, when we launched, it started off with lots of drinks. I do remember the launch party very well.

Barton: Obviously it was fantastic. I think everyone was delighted. We were big believers in celebrations. Dhiraj introduced one of the most important fundamental philosophies at Shazam, which was that after a successful meeting with a partner, we would have to go directly to the pub to have a drink and celebrate.

Mukherjee: That was okay at six o'clock in the evening. At eleven o'clock in the morning, it was a little harder.

Barton: I can remember going with Dhiraj to the outskirts of London and meeting with—I think it was T Mobile, one of the mobile operators, and then after the meeting we were out there in the middle of nowhere and the first thing we did, it was probably around midday, we went straight to the pub and had a pint and celebrated just because it was a good meeting. Yeah, we're big believers in celebration. Certainly the launch was very exciting.

Mukherjee: One of my roles was to run the launch or to coordinate the launch with all the different pieces we've talked about in place: the music database and the mobile operator deals and the actual technology, etc.

We had just hired our new CEO. He started a couple of weeks before our launch. I said to him, “Look, who knows when we're going to launch because it's so hard to say when Orange is going to agree to give us this number. Can we just make it a moving target?” He said, “Nope. Can't do that. You have to pick a date. You have to tell me when it's going to be, and we have to launch then.” I said, okay. So we had this giant spreadsheet with every single thing that needed to happen.

I came back and said, “Well, it's going to be the sixteenth of August.” He said, “What day of the week is that?” So, I said Friday. He said that's not a good day for a launch because it's a Friday. How about we make it a Monday? I said, “Fine, we'll have it on Monday, the nineteenth of August.”

On Monday, the nineteenth of August, everything was done. So all the pieces came together, we launched, the CEO was on television, and then Philip has already told you how we celebrated that evening.

Inghelbrecht: I think there was one other detail there. Chris said, “Yeah, besides that's my dad's birthday, so that's a good day.” Chris, is that right?

Barton: Yeah, that's right. It's still my dad's birthday.

Mukherjee: Still your dad's birthday? That's crazy.

Barton: Exactly. It really helps me remember his birthday, so it's great.

Mukherjee: I still celebrate even ten years later. I never fail to think of our launch when the nineteenth of August comes around.

Barton: Philip's wife Mekhala, who had joined us as one of the earliest employees at Shazam, still sends an SMS out to people saying it's Shazam's anniversary. It's great.

Santos: As Dhiraj said, a bit before the actual launch you hired a CEO, which now is not seen as a normal step.

Inghelbrecht: It's a funny world. Today there is a concept of CEO/founder. VCs only fund companies where the founders will become the long-term CEO. That's what they want. They want the founder to remain the CEO unless it goes awfully wrong. Back in the day, it was more fashionable and it was also required from us that upon us raising our Series A, we would find a gray-haired, experienced CEO. And that led us to Jerry Roest.

Some CEO candidates were pushed upon us and we really did not want them and we would fight that very hard. But the concept that the CEO would have to come in was the norm. That was how start-up companies came about back in the day. After raising money, you would hire a CEO. It was a fundamental belief it was a good thing because as a founder you should be very happy to step down if a guy or a girl would come on and turn us into a much bigger company. It was a belief that it was a better thing for the company. Nowadays, we'd be hard-pressed to do that.

Santos: How hard was it, from your point of view, to actually find a CEO that you were happy with and to get everyone behind your new CEO, Jerry Roest?

Barton: I would say it was very challenging. I remember interviewing multiple CEOs and I remember there were ones I didn't, actually we all didn't, like and the VCs did like, and as Philip said, that was very problematic. But we always knew that we were very willing to hire a CEO. In fact, we even agreed to it right up front when we signed a term sheet with a venture capital firm, and we had no problem with it. But we wanted to make sure we had someone who is a great fit.

Actually, I've read the Google story about how Larry and Sergey, they took a long time to find a candidate that they liked, as well. And just trying to find that right fit is so hard. When we met Jerry, we really, really loved him. He had a great, great energy. He's a very smart guy, he picked up things very quickly, and he had been a proven CEO and leader.

So we're very fortunate, because after meeting a lot of other candidates that didn't have that kind of connection, we were very pleased to have him join.

Santos: What were the things behind the energy that somehow made him a good fit for you and the investors?

Barton: I think the investors liked him because he was a blue chip CEO. He had run CompuServe Europe, which was a large technology organization. He had been one of the senior directors at NTL, which is a humongous cable company in the UK. So he had all the big leadership experience. What we liked about him is that he came from a technology background, which was great. Shazam has a lot of passion in the business. It's a consumer-facing business. So you need someone who is not just a technology person, but someone who has a kind of energy. Because it's music and it's consumer-facing.

And Jerry, his early career was actually in sales. He was a superstar sales guy and rose up to be one of the top salesmen at Xerox. So basically he had a sales like personality. He was very, very skilled at building excitement with partners and also building a great sort of excitement within the culture of the company and amongst the employees.

He's very much of a people person and he's one of the most talented people I know in the world in terms of understanding and dealing with people, all types of people. And he can just walk into a room and, just like the best salesman on earth, he can just get to know who you are and what you want, how to connect with you. And he can do that across the employees of Shazam, as well as partners of Shazam. In that perspective, he was a huge, huge asset.

Santos: Who did the search? You or the VCs?

Barton: The VCs found one of the top recruiting agencies and through them, brought in all the candidates.

Santos: Regarding the launch of the service, it didn't go viral. Did you feel a lot of pressure from the investors and the people who backed the start up because the growth was not as high? Or not really?

Barton: I don't think anyone expected it to go viral, but we definitely felt pressure because ultimately we had achieved everything we said we were going to achieve. I don't think that there was an expectation that it would just be instantly a financial success. By the way, the investors behind us were almost as green in some ways as we were as entrepreneurs, so it was a learning experience for everyone. I think we were one of the first investments by this guy from IDG Ventures named Ajay Chowdhury.

Once we had accomplished all these things, we recognized, “Oh, my gosh, we have a significant burn rate.” Meanwhile, the big, unknown, unforeseen event was that attracting capital became very, very difficult. As you know, 2001, 2002, 2003 were tough years.

When you have a high burn rate, a need for capital, and don't have significant revenues, even though you may be achieving all of your milestones, which we absolutely did do, the revenues are king. You need revenues. It was definitely challenging. It was very, very challenging. I think everyone felt pressured, both the VCs and the entrepreneurs and the employees. I don't ever recall people thinking it would completely go viral and we'd have it be instantly profitable.

Inghelbrecht: We had prepared for launch. We had hired a VP of Marketing … and tied it all up. And the reality is that the whole concept of virality and social components simply did not exist then. It wasn't as if we flipped the switch on Thursday and then on Friday traffic went through the roof and grew exponentially. It wasn't like that.

I will take a big step back now. Shazam, in its first incarnation, sampled the music through your handset instead of speaking. It was what is called an IVR or Interactive Voice Response system. The same way as you would call your bank and push numbers to navigate. It was a bit the same system and you would get a SMS back, limited it to a few characters, and you could not buy the track immediately. It was just an ugly experience. And when we launched, surely we had traffic but it wasn't the traffic that we hoped for. And, really, it took the coming about of what we call the smartphone and, of course, the iPhone, when we became a much better user experience with a better business model, the freemium model, for Shazam to truly take off. We were way ahead of our time. We built, not to toot our own horn, the killer app for music—but there were no smartphones yet. It was five years before the smartphone came about. Obviously it gave us an early mover advantage. When iPhone did launch, we were there right at the start.

Santos: But you still had to wait five years for that.

Inghelbrecht: The company was growing and it got a lot of growth by launching in other countries outside the UK, very often on a white-label basis. But that is not what I would call a healthy organic growth or traffic. It took the smartphone to get us there.

Santos: Were you ever worried that the company might fail?

Inghelbrecht: Yes. As you may have noticed, I always make it sound as if it was all fun and peachy as we started the company. Part of starting a company is riding through the lows as you ride the high, as well. Surely, we had our very hard moments. We launched a service, it went okay, it did not take off like a rocket.

At the same time the entire VC market imploded. In 2002, the whole market went to the gutter. We had to raise more money, we had up to seventy people, we had our layoffs, and everything like that as well. So! Your early hires become your friends because they go through thick and thin with you, they take almost as much risk as you do, and when those moments come that you actually have to lay off your friends—it is not a lot of fun. I don't wish that to anybody. And surely we had those moments as well.

Santos: Was it in 2002 that you had to fire people?

Inghelbrecht: I think it was late 2003—early 2004.

Santos: That part is always hard.

Inghelbrecht: Yes, but there is nothing you can do. You can either all go bankrupt together or you try to minimize your burn rate and go on in survival mode. Some of these people are still friends today. They were not laid off because they were underperformers. It was just because we couldn't pay them. It was just the brutal reality. They were all smart people and they got jobs very quickly. We tried to help them in everything we could. Looking back at the history of the company, some of my fond memories are about people that we hired and the stories around that.

When we started Shazam, Dhiraj was part of the team running the Viant office in London. That was pretty much when the internet went downhill. Viant was an internet consulting company. It was even traded on the stock exchange and was doing really well. And, of course, as the internet went down, they lost all of their business. Shazam was like a hyena circling around Viant because as they went out, we literally took on their best people. We cherry picked them. So, from the first ten employees of Shazam, five were Viant people. Some of them were investors. When they closed an office, we would take over all their infrastructure, down to the chairs.

Yesterday, I had dinner with one of the guys who helped to build the database of music. We had to rip a lot of CDs to extract those fingerprints, so it was like a 24/7 operation when you had young kids flipping CDs day and night. You have to manage that. You need a strong manager for that.

So, I actually hired a guy who had done similar work in the US before and he was an ex-military man. His name is Bart and I will never forget. We first interviewed him over the phone. So, we thought the guy was good and very strict and he would be a good manager for those kids. I flew him out of the States, because he was there, to interview him here. And at the end of the interview, we all for some reason went to a bar, at four o'clock in the afternoon, and they had tequila tasting. So we got into tequila tasting and I started to realize I still had to drive him back to the airport, so I stopped drinking. So Bart had all my tequila tastes himself.

So, everybody was blissfully drunk as part of the interview, except me. This was getting interesting because I was getting a very sober view on what was going to happen. Bart helped himself straight through all of that. When people get drunk, this is when you really get to know something about them. But Bart ended up well. I drove him to the airport and the next day Bart was hired and a week later we flew him out to the UK. So, he literally had the tequila test.

Some other guys, as well. We had these three engineers from Viant, they were all in Germany, so we just hired them in one go. So, we told them: Viant in Germany is going to close but here is a job even before you lose yours. And we would fly them to London and called them the Dream Team because they were very efficient in what they did.

In terms of employees, the very first employee, Mekhala, actually became my wife. She was a student in London and helped Shazam in the early days for free. So, she worked for the company, we became friends, from friends we became something more, and fast-forward, in five years we got married and now have two beautiful daughters, thankfully taking after their mom. So, Shazam to me is obviously more than a cool service or company. I got my wife and children through it as well.

Santos: It is life-changing in every sense of the word.

Inghelbrecht: Pedro, we had this fun thing in the early days. Whenever we had milestones, we all had to dye our hair. So, I had green hair, red hair, blue hair, black hair. At one point or another there was always one of us who walked into the office with his hair completely dyed for whatever reason.

Santos: That must have been interesting to go back to VCs and talk again.

Inghelbrecht: We have pictures of that. I actually think it may have given us some credibility that we were really experts on the music market.

Santos: How did you actually grow into new markets?

Mukherjee: We hired an international director who was one of our original angel investors. He was older than us by about ten years. He had gray hair, gravitas, and a lot of experience. But he also loved to travel all over the world. He flew all over doing deals with mobile operators in some cases, with content companies in others, just literally traveling around pitching the Shazam story and looking for local partners—which is why we ended up with a huge mobile footprint. I think it was nineteen countries within a year after our UK launch, because wherever there was a deal to be done and partners interested, we established those relationships.

We ended up being called by different names, in different territories, on different operators. But that was the nature of the beast at the time. Now, it's good that Shazam is known as Shazam everywhere in the world. But when we started out, we just got on with it the best we could.

Inghelbrecht: This actually really highlights how fortunate we were that the apps stores came about. Because that allowed the company to move forward under the trade name of Shazam, rather than the white label name—music finder, music ID, and whatever names operators or licensees would give to it in various countries. The apps stores allowed us to proceed under the name Shazam, which obviously has become a brand in its own right, now. It definitely has made the company much more valuable than it otherwise would have been.

Santos: What's your advice to the readers of this book that want to follow in your footsteps? What have you learned from this amazing experience at Shazam that you would have liked to know from the start that would have prevented you a few mishaps?

Wang: A few things come to mind. One is persistence and another one is friendship, which I think on both points, we've been very strong on. The persistence part, there were many points along the way where things were looking very discouraging and we might have given up, but somehow there's this “can do” attitude. This attitude of where nothing is impossible. That, I would say, was greatly exemplified by Chris Barton, who basically doesn't take “no” for an answer.

When I was working on the recognition algorithm I would come to these roadblocks and then I would just say, “I don't think this is possible.” I suppose this ties back to the development process of the algorithm that I was talking about before. Chris would, even though he doesn't have a technical background, he would dissect my statement and say, “What are your assumptions?” I would give some factors and he would take every single one of my statements and then drill deeper. Eventually, I would get to the point where I would say, “Because I have a PhD. Because I say so.” [Laughs]

Basically, I would find that I wouldn't really have a definitive reason as to why it wouldn't be possible. Then he would say, “Okay. We're going to do a project plan and you're going to solve this thing because you don't have conclusive proof that it's not possible. You're going to solve that problem.” He'd rip it down into little pieces and didn't take “no” for an answer at any point.

That worked really well. It's also true for a lot of business negotiations and that kind of thing. I think all of us were very persistent. Philip did a bunch of deals with the recording rights industry. All of us were blazing trails where nobody had done this kind of thing before. A big challenge was creating new markets, which is quite difficult. Especially in a situation where—this is after the bubble hard burst in the year 2000, and we basically got started right after that happened. So that is very challenging and Chris, Dhiraj, and Philip were raising money in this environment. Going back to my point, their persistence really paid off.

Another one is friendship. Throughout all of this we've—even though we went through some challenging times—we remained very good friends and stuck together. There are some quite difficult situations where one would think that our goals might not be aligned. I'm not sure I want to get into that, but we stuck together and made sure that there was a fair outcome to a larger extent.

Mukherjee: I would definitely agree with latter point—we enjoyed each other's company, we were always fair to each other, and I think we put our friendship before the business in some sense. I think the first point—I agree with Avery as well. I think the Shazam story is a crazy story and I would strongly advise any entrepreneur not to try to follow this in any shape or form at all.

Come up with an idea which is impossible then try to find somebody who can make it un impossible and then do deals which have never been done before. It's hard to imagine how one could set up a worse plan, but I'm just incredibly glad that we're even in this conversation with you, because so far it's all worked out.

Santos: I think that actually comes in line with what Philip said, as well. That everything fell into place.

Mukherjee: Yeah. So persistence was a good plan of attack, but trying something smarter would be another one.

Wang: I think in general, we kind of do it both ways. Adding to what Dhiraj said, we're blazing new trails in a couple different areas, technologically as well as business wise. We were about five years too early in the mobile market, I think. There wasn't really a marketplace for apps at that time. It's very easy nowadays to launch an app—for example the iTunes or Android stores. The thing is that it can go both ways, certainly trying to blaze new trails is very risky, but then one advantage there is that if you succeed, you're a first mover in the market and you gain certain advantages there. But then it is higher risk because nobody has done it before. And then, if you fail on any of the points, like creating the technology or doing the business deals or if it's not possible to actually make a profitable business, then the company is dead. It's high risk and high reward.

Mukherjee: I don't know if the other guys know. When we agreed on the idea for Shazam, as Chris said and Philip said, there were other ideas on the table. So I basically did a spreadsheet. I looked at each of the risk factors and I looked at the probability of success for each one and I just multiplied out the answer. I think my spreadsheet said there was like a four percent probability that we'd make it. I said, “Okay, well, you know, I'm in!” Chris said, “But I thought you said your spreadsheet says there's like a four percent chance of success?” I said, “Yeah, but it sounds like it'll be fun. So count me in!” But actually four percent was a wild, wild overestimate. I think it was a fraction of a percent. So I'm glad we're here 10 years later to tell the story.

Santos: Why the name Shazam?

Wang: I don't know. Anyway, the working name for the company was going to be Aardvark.

Santos: Aardvark?

Wang: The reason for that was with the double A at the front, it would always be at the top of any alphabetical list.

Santos: [Laughs] Okay. What made you guys decide to move to Shazam? Do you remember?

Inghelbrecht: I don't think anybody really appreciated the name Aardvark.