Podcast of Tom Conrad’s talk on Apple, Pets.com, and Pandora

Tom Conrad, Pandora CTO

You’ll hear tons of laughing on this podcast, this was certainly one of the most entertaining soapboxes we’ve had here at ZURB. In just over 30 minutes we managed to discuss Tom’s dating life, driving Pets.com into a cliff, and of course how Pandora was started. Enjoy the podcast as you skim through the summary below.

Listen to Tom's Podcast

Subscribe:
iTunes RSSRead Transcript

Follow your passion

At 15 years old Tom saw the 1984 Apple ad during the superbowl. He became fixated on going to California and working at Apple on the Macintosh. He applied for an internship and followed his dream. After his internship he worked on the Finder team at Apple for three years.

Be careful of what you say...

Tom was part of the computer engineering nerd clique in college which was in rivalry with the computer science nerd clique. All Tom wanted is to work at Apple. He learned that someone from the engineering group got a job at Apple. “Who the hell is this doofus?” Tom kept asking everyone. Turns out it the doofus was Tony Fadell, inventor of the iPod.

Don’t be a generalist. If you can - go deep!

Tom loved his experience of working at Apple for three years. However he saw that at Apple people who are rewarded and admired are the multi-talented renaissance men. These people are good at marketing, and can also do product management, and have input on the design, and can code the functionality up as well. These people are generalists. “That’s great” Tom says. However, when Tom came out of Apple he was not great at anything. He certainly wasn't a great programmer.

At Berkeley Systems where he led the development of the You Don't Know Jack game, people were much more focused. There was a team dedicated to marketing, a team dedicated to design, and a team for development of the game. This really helped Tom excel at his craft. Tom mentioned that it's really important to create an environment where people know what their responsibility is. At Pandora the engineers are the best in the world at coding, the business development guys are the best and sales and making deals, and the marketers are the best at their craft. While all the functions interact constantly everyone knows where their line ends.

It's about People and Passion

Tom got a puppy which took over his life. Coincidentally his friends from "You Don't Know Jack" asked him to join them starting this new site where people could buy dog supplies online. Tom loved the idea. It was 18 months from the founding of Pets.com to IPO to going out of business. Yet, when Tom looks back he is grateful for the experience and has no regrets. It was probably the hardest he has ever worked in his entire life on something. He is of course sad to see the company go out of business but he had an amazing experience because:

  1. He loved the people
  2. He was following his passion

Pandora - humble beginnings

Tom had a large music catalog. However he had trouble finding out when his favorite bands were in the area, meet other people who loved similar music, or actually buy the tracks. He had an idea of a software that would digest your music catalog and tell you when your favorite bands are coming to town and connect you with people who like similar music. He was introduced to Tim Westergren who at the time was trying to figure out how to make Music Genome project into a direct consumer business. We are going to leave the rest of juicy details of how Pandora was started for you to listen to on the podcast.

In conclusion - this was one of the funniest and most sincere soapboxes we're had, and as usual, we walked away with some great lessons from Tom's career. We'd like to thank Tom Conrad once again for such an entertaining afternoon!

Don't Miss Out on Our Next Soapbox

Soapbox transcript

Tom: I became like, completely fixated on one thing which was ending up in California, working for Apple on the Macintosh. Be careful of what you think of people because they may end up changing the world. The joke that I think was very, very accurate about Apple was it was the only place in the world where a vote of thousand to one is a tie.

Moderator: We're super excited to have Tom here, Tom Conrad from Pandora. I heard this guy speak at Stanford, and a lot of my friends heard him at the Freemium Summit, and it was an awesome, awesome talk, like juicy stories. He's got just such a rich experience of so much stuff that he's done at Apple where he was just starting his career. He's got [inaudible @00:56] there, Documentum going through that and building that product up and leaving millions of dollars of potential stock and options and going off to start Pets.com where …

Tom: Oops. Exactly. The biggest oops ever.

Moderator: Took that through IPO, threw that off the cliff, and then Pandora.

Tom: I think we drove it into the cliff. I think.

Audience Member: You're going to have to get the Russian translation.

Moderator: And then, Pandora, of course, taking that, having a vision for it and taking it to what we all know today. So, I invited him over here to share some of his lessons learned from all of these experiences. With that, let's welcome Tom to ZURBsoapbox.

Tom: Hi, I'm Tom. I've been doing this, I guess, it's been almost 20 years. I grew up in Columbus, Ohio. I must have been 15 years old when it was that original 1984 ad for the Macintosh during the Super Bowl. I really do remember that ad like it was yesterday. It's remarkable that I was watching the Super Bowl at all because I'm like, then and now, the tiniest, most insignificant sports fan in the world.

Somehow, I was watching the Super Bowl at 15, and that ad came on. I just remember thinking that the Mac was not even — I didn't know it was going to be this huge momentous thing to the point I went the next day, and they had a full page ad in the Columbus newspaper for the Mac. I tore it out of the newspaper. I found it when I was at home at my parents a couple of weekends ago.

I can't believe that when I was 14-years old I pulled that ad out of the paper, but that was definitely pretty consistent with what happened to me the next six or seven years, because I became just completely fixated on one thing, which was ending up in California working for Apple on the Macintosh.

People would be like, "What do you want to be when you grow up?" And I'd be like, "I want to work at Apple." It was that precise. I went to college, not to get a degree, but because I wanted to work for Apple. I went to the University of Michigan, studied computer engineering, and after my junior year I applied for an internship at Apple.

One thing I remember about that period was that after my sophomore year, there was this guy who got an internship at Apple. The way Michigan works is there is a computer science school and a computer engineering school, and they're completely separate.

The way I viewed it at the time, if you were a computer scientist, you had to take a foreign language, and if you were a computer engineer, you had to take electrical engineering. I thought the electrical engineering sounded easier than taking a foreign language. It turns out I'm almost certainly wrong about that because I was the world's worst electrical engineer, the worst. I survived my EE classes.

Anyway, there were computer science kids and there were computer engineering kids, and there were two separate campuses and there were two separate computing labs. I worked in the computing lab for the computer engineering school, my job to help pay for college. There was a whole bunch of us, and we were kind of a little nerd clique. There was another little nerd clique that ran the computer science, and there was a little bit of a rivalry between these two groups.

So, there was this guy that was in the computer science program and worked in the computer science little nerd clique. He was a year older than me, and he got this internship at Apple when I was a sophomore and I remember thinking, "This doofus got this internship. How could they hire him," because we had this whole nerd rivalry thing between the two groups? It seemed impossible to me that they would want someone from the computer science program. Here is what you need to know about that person. That person is Tony Fadell, who invented the iPod, so I think definitively, not a doofus. I guess the lesson there is something about be careful … be careful what you think of people because they may end up changing the world. I don't remember Tony very well, unfortunately, but I do remember thinking that.

So how did this happen? I wrote this cover letter to Apple that just went on and on and on about how I wanted to work there my whole life and I wanted to change the world through computers and all this really dorky kind of stuff. I mailed it off to someone that I knew there and managed to get this internship, I think just on the strength of the cover letter. I don't remember interviewing for it.

So I go off, drive my dad's car across the country and end up at Apple as an intern. The best thing about that story is that by the time I got to California the driver's side door of my dad's car wouldn't open any more, so I spent the entire summer climbing over the console and getting out the passenger side door which is really awesome on a date. "Hon, no, I'm not opening the door for you. I'm just getting in." I've never made that joke before. I need to write that one down. That's a good joke.

Let's see. So, anyway I had a really great experience at Apple that summer and then when I left, it seemed that I most really wanted to work on, I had worked on the graphics engine and the Macintosh OS Quickdraw and what I really wanted to do was work on stuff that had a user interface to it, and, at the time, the big thing that Apple did that had user interface was the Finder, the kind of the core user interface with Macintosh when you booted up. That was a really, really small team. It was like seven people. By my last day at Apple of my internship, I went and found the manager of that group and went, "I was an intern here this summer and I really, really want to work on the the Finder when I graduate. It's the only thing I want to do in the whole world. If you have any positions nine months from now I really want to talk to you about it."

I go back to school, finish my degree and in the Spring I get a call from this hiring manager and he says, "It's your lucky day. The college recruiting department just called and said that if I could hire someone today, I get an extra hiring rack. But if I can't find somebody today then I lose the rack. So do you want to come and work on The Finder when you graduate from college?" And I'm like, "Did I just get a job from Apple without interviewing?" And he's like, "You know, I asked around. It seemed like you did a good job last summer. Blah, blah, blah."

I got this job. I'm finishing up school the last couple of months. I promise not to go into this much detail every year of my life, but this early stuff is funny because I was so ridiculous. So I'm finishing up my degree and I had worked at CompuServe when I was a kid because this was in Columbus and I did little programming things and stuff there during the summers. I was a CompuServe user and all these connect forms and discussion boards and stuff and they had one that was all about the Macintosh and because I had no judgment at all, I decided to post on this discussion board, "I just got a job at Apple and I'm going to go work on The Finder next year."

That probably would have been not very wise in and of itself but I also said, "If you have any ideas how to make The Finder better, I'd love to hear them." So this thread became the longest thread in the history of the Macintosh forum. It was thousands and thousands of messages of suggestions and it became this big deal on CompuServe and then, at the time there was this magazine called MacWeek which was kind of like the Tech Crunch of the Macintosh world back then. It was a print magazine and it was weekly, hence the name, MacWeek.

So they had a reporter call me to talk about this thread on CompuServe and all the feedback that I had gotten and then they sent a photographer to Ann Arbor to take a picture of me. I will give you one piece of advice from this, which has come in handy a number of times in my career. Never, ever, ever let them photograph you with a prop. Not ever. Every ridiculous picture you ever seen of some type of exec is like, it's Kevin Rhoades in headphones and it's Dennis from Foursquare wearing the crown. And in this case, it was me throwing an apple into the air, waiting to catch it. Ridiculous.

So I realize after the photographer has come and all this has happened, I'm like oh, this is bad. I haven't even started yet, I didn't even interview for the job, and here I am, like this dorky 21-year-old kid. I'm sure the guys on the Finder team, the rest of the team are like what the fu — ? Like this is, you know, like who is this guy? So somehow I survived that. So that's what I was like when I was 21.

So I worked at Apple on the Finder for another 3 years and had a really great experience. But the thing is that Apple's a wonderful place. It was full of just really the smartest people that I think, I should say out of respect for the other people I've worked with since, some of the smartest people I've ever worked with and I would certainly, I always felt like I was the dumbest guy in the room. I remember when I was an intern, I spent the entire summer overusing the word orthogonal but I pronounced it ortho- gonal every time. Pretty awesome.

At any rate, dumbest guy in the room and Apple was great. But the challenge, at that time, this was kind of before Steve came back, and the joke that I think was very, very accurate about Apple was, it was the only place in the world where a vote of 1,000 to 1 is a tie. And that is really how it'd work. Like we'd get in a room with this many people and we'd have some discussion about something we wanted to do to improve the Mac or change the user experience or something, and every single one of us would agree on some potential direction and then there would be one guy, Dmitry here would have some point of view that was counter to everyone in the room. And Dmitry would be a really smart guy and we would all think like oh, boy, I don't know. Maybe we shouldn't, maybe we're all wrong. Maybe Dmitry's right.

Actually, I remember one particular example of this. There's a bunch of people that had come together to talk about drag and drop semantics, like you pick an icon up and you move it from here to here, like one window to another window, what should the behavior be? Should it be a move or should it be a copy? And on the Mac and actually on computers to this day, it's actually often not consistent. You drag a file from one folder to another, it's a move. You drag it from one disk to another, it's a copy. And there were people at Apple who were troubled by this lack of consistency and that it would be far better if it was always one thing or another.

Well, just thinking it through a little bit, it can't always be a copy, right? Because if you just nudge the icon to the right by an inch in the same folder, you don't want it to duplicate. So you can't make the consistent behavior be copy. So there were literally people who thought the consistent behavior should be delete, which would mean, if you take your expense report and you drag it into an email and send it to someone, it would delete it from your hard drive, because it's always a move.

So we were having this debate, this vigorous debate, about this behavior in a meeting, maybe about 15 people and there was a guy with kind of a white beard, at Apple … we say the guys with the white beard. Now today you'd think they work in a Linux kernel. Then you'd think they came from Xerox PARC. So there's this Xerox PARC white-bearded guy who is arguing pretty passionately for it should always be a move and that the first time you accidentally emailed your expense report and lost the copy of it, you would learn to do the special gesture that meant copy. Like it would just be, you would learn hold down the control key and it's a copy.

I thought that was ridiculous, despite the fact that this guy seemed very Xerox PARC-like and 40 years older than me. And I had this really vigorous debate with him and we left the room and I'm walking down the hall with the other guys with me and I'm like, "who the hell is the guy with the white beard?" And the guy next to me goes oh, that's so-and-so. He invented the icon. This guy literally invented icons at Xerox PARC it turns out. And I'm like, here I am, little tadpole Tom Conrad having this fight with him. I still think I was right. Anyway, that was kind of endless debates we get into at Apple, and it was not the best part about Apple. I think actually in many ways the thing that's great about the new Apple, is that Steve is back, and they don't vote on things anymore. Maybe if it's still a vote of 1,000 to 1, the 1 always wins. I don't know what it is, but it's still very different, and I think that Apple needed a tie breaker. I think that's part of what the genius of Steve coming back is.

Because of that problem and a handful of others, like despite all these brilliant people, we didn't ship very much software at Apple. As Macintosh users, you know about spring-loaded folders. In fact, I wrote the spring folders, my third week at Apple. I'm the inventor of spring-loaded folders. I use that on dates too. That actually works on a certain kind of date. So, I invented spring loaded folders my third week at Apple. Wrote all the software.

Funny little design story about that. When you pick up an icon and mouse over a folder, and it's about to spring open, the icon blinks three times and then it opens. Well, it blinks three times, because at 2:00 in the morning, when I was doing the first implementation of it, I'm not a graphic designer, I'm a bit of an interactive designer, but I'm not a graphic designer. What I really wanted to happen was for the folder to animate open, but I couldn't do that. All I could do was blink it three times, and I knew how to write the software that would blink it three times, so I wrote the software to blink it, and wrote down in my little engineering notebook, get designers to design opening animation. That never happened so they still blink three times to this day. Despite the fact that they completely rewrote that feature when they rewrote the Finder, not once but twice since I left. Be careful what you put in as the placeholder, because it might survive for twenty years.

I wrote that feature my second week at Apple. I stayed for three years. It hadn't shipped. Two years after I left, five years after I wrote that code, it shipped. It was really hard to work at a place where I wanted to work somewhere where the pace was faster. I was at a conference sitting at a table with a bunch of people. We got to talking, and when I got home, I got a call from one of them.

They said, "Hey, Tom. We all thought you were great. We're working on this video game. It's kind of like a quiz show, party game thing. We'd like you to come in and work on the game if you'd like something different." That sounded pretty appealing to me because in video games, they come out every six or seven months, and it's much, much faster paced. That game turned out to be this video game called, "You Don't Know Jack," which I worked on for several years, and was technical director for all the various iterations of that franchise.

I think the big lesson about that change was that Apple's culture is such that the people who are held up and rewarded and admired are kind of renaissance men, who can do a little bit of user interface design, a little bit interactive design, a little bit of programming, and have great business sense, and have input on how the product should be marketed. The more of a generalist you are, the better you do at Apple. I like all that stuff, so I was perfectly happy for three years to flit from one activity to another, meddling in one aspect or another of the product.

The downside to that is that I wasn't really great at anything when I left Apple. I certainly wasn't a great programmer, that's for sure. When I got to Berkeley Systems, who made "You Don't Know Jack," the environment was really, really different. They had a team of people whose sole job was game play design. All that group did was design the game plays. Another group of people only wrote the questions and the comedy bits and recorded the audio. Another group just did the graphic design. How do we take this game play and turn it into something beautiful. The development team's job was to take all those assets and make really high quality, really efficient, really well tested code and get it done on time.

My job was to lead that team of developers, and I actually wrote a lot of the software myself in the early part. That was a total and complete change for me. Sure I had the opportunity to express an opinion on these other things, but, at the end of the day, I wasn't going to be judged on how much of a contribution I made in those other areas. I was really going to be solely judged on whether the software high quality, is it bug free, does it ship on time, are the other engineers happy, did you hire the right people? That kind of thing. I feel incredibly lucky that by just happenstance I ended up in an environment like that because it really let me get good at programming in a way that I don't think I ever would have been had I not left Apple.

Despite the fact that I am a bit of a generalist ever since I've tried as I've had the opportunity to be the one of the people that shapes the culture of the organizations I've been in, I really prize this thing of "the doors are always open, people can have input into things," but I think it's really important to create an environment where people know what their job is, people are judged primarily on being really, really exceptional at that one thing rather than encouraging every single person to be a corporate gadfly that gets involved in everything.

Certainly that can work. In fact, it might be, in some cases, the common Silicon Valley culture. But I think it's really helped Pandora in particular that we have this much more "our business development are the best in the world at business development and our engineers are the best in the world at writing software and our interaction designers are the best in the world at interaction design and everyone knows their job." We're all friendly but at the end of the day we know where the buck stops.

I had this great experience at Berkeley Systems on "You Don't Know Jack" particularly when I became a programmer and it was really fun to have this, at the time, "You Don't Know Jack" was pretty well known and a bit of a phenomenon. Amazingly, I think this holiday season "You Don't Know Jack" for Xbox and PlayStation and Wii is coming out, it's almost 20 years ago that we put the first version of this game out. I can't believe they're still making versions of it. It was actually part of the reason that I left after a couple of years. The writing was on the wall that you could make a career out of doing You Don't Know Jack sequels and after a while it just got to be too much of the same.

The other thing that happened, which is maybe a little bit of a cautionary tale, was the consumer stuff that I had done to that point in my career, which is six or seven years at that point, I felt like, "I'm getting pretty good at this and it's really not that hard and I understand how it works but the stuff I don't understand is all this fancy server side engineering stuff." A friend of mine was starting a company that was kind of search technology and had fancy conceptual indexing and it could take documents and figure out not just what they said but what they were about.

So you could have an article that talked about Puget Sound and Microsoft and the Space Needle and this technology could be, "Oh, this document's about Seattle," even if the word 'Seattle' never appeared in the document. I went off to do that largely because I thought it was going to be hard and challenging in a way that maybe the consumer stuff wasn't so much anymore. It was starting to lose some of its luster.

I learned two things in that experience, which lasted a couple of years. The first was that it wasn't particularly harder, it was just different. I actually think there's a lot of tech-elitism in our industry. It's like this software is easy and that software's hard. Those engineers over there are the most valuable ones and these guys are replaceable. I think it's just a bunch of bullshit. I think it's all kind of hard and it's all kind of easy and looking for ways to keep the culture respecting the contribution that everyone makes is something I've really tried to do at Pandora and places like Pets.com, which I'll talk about in a minute.

I did that for a couple years. I won't dwell on it too much. They got bought by Documentum. I had a bunch of stock options. Honestly, it's the only start up exit that ever had any positive outcome for me, and I walked away before most of that value really managed to make it into my bank account because I wanted to get back into consumer software. I missed being involved in consumer stuff so much. I just wasn't passionate about this information retrieval- document management-enterprise software stuff.

The way I ended up at Pets.com. I had just gotten this puppy. It's a true story. For those of you who have never had a dog or kids or anything, not that dogs and kids or the same, but I'm just saying. There are lots of differences. You can put a dog in the kennel when you don't want to be with the dog. I had this little puppy and it was the absolute center of my life. It had just taken over. Suddenly, I went from just living this completely selfish life where all I did was program, do whatever nerdy things I did, and now I had this dog. I had to take care of it. If the dog wanted to go out, the dog had to go out. I thought it was the best thing ever.

So in the middle of all that, I get this phone call from the people who had done "You Don't Know Jack" with me and they said, "We're putting the team back together. We're going to do this internet thing and it's called Pets.com." The combination of these people that I love in consumer land and I had this dog, suddenly, "Oh, that's a really good idea. I don't like going to the pet store to get dog treats and dog food. And so it would be great if it could get delivered to my door."

It was 1998 or whatever when everything was all frothy and crazy and E-toys had just gone public for $7 billion of market cap or something, and there were all these stats like "people spend more money," this is so disturbing, "people spend more money on their pets than they do on clothing and toys for their children combined." So we had a big, giant $28 billion U.S. market category, blah, blah, blah. So we started this thing and it was 18 months from the founding, to going public, to going out of business, 18 months, which is just insane.

The funny thing is that you would think that I had walked away from this enterprise software thing where I had a bunch of options that were worth something to this really kind of catastrophic glorious disaster, that I would look back on that and think that was a really, really bad decision. But the truth is, that is actually not the case at all. I worked with absolutely remarkable people there who taught me tons. We worked harder there than I think I've ever worked on anything. It was unhealthy how hard we all worked. Somewhere there might have been Crazy.com comings and goings during that era, but they certainly weren't having it at Pets.com. We were working on like a pair of filing cabinets with a door on top of it and just chaos.

I look back on that, incredibly glad that I left. And certainly, having had that experience building a large-scale website and consumer space helped me when we got around to building Pandora years later. But I think mostly I loved that experience because of two things. One, I did love the people. It was a great group of people and I was really passionate about it. I had this dog and I loved the idea of the service and I loved that for a while there people loved it, until we took their grandparent's pension.

At least for me, I think when I've made decisions that where I'm following my own sort of personal passions and things I'm really interested in, particularly when I was going to work with people I really cared a lot about, the outcome was always good even when it was maybe, you might look at it and say, "Wow, that was sort of [inaudible @28:46] a disaster."

Somewhat to that point I went on after that to another enterprise software thing which I won't bore you with the details of, although it was sort of like an AJAX platform before there was the acronym AJAX and spent three years there. Here again, it was another one of those enterprise software things. I really wasn't that passionate about it. I had a little bit of passion about making software to make programmers more productive and make better user interfaces and things, but mostly, my consumer products are kind of in my DNA. I spent that whole period really anxious to get to some other consumer offering, and finally, after …

This was kind of the nuclear winter period post the bubble bursting and there were just no jobs and there was no investment money for people to start their own things and so I was just grateful to have a job at all. I had friends in that period who went years and years and years, super- talented people, who couldn't find any work at all in technology. It's hard to believe that was just seven or eight years ago.

Finally, in 2004, I couldn't do it anymore. I figured, "Gosh, even it I can't get a job, I'll set up a laptop in my basement and write some software on my own." It was kind of in that period where Flickr had just started and Delicious had just started and there were a handful of consumer-oriented things being done. They were being done on a really, really small scale, in people's basements and sometimes finding acquirers from basement phase to acquisition phase. I thought I'll design something in my basement, I'll write all the software myself and see where it goes.

As I started to think about what I might want to do, thinking about the problems in my life, I had been somewhat late to the game ripping all of my CDs. I had been a bit of CD collector my whole life, one of side passions is collecting music, so for the first time I had all this digital music in iTunes and I remember in that period being really disappointed about how little I could do with it. Sure I could play it but I couldn't learn more about the band, I couldn't see when they were coming to town, you couldn't buy the tracks, you couldn't meet other people who shared similar interests.

So I started developing some ideas around a community-oriented recommendation service that would digest your iTunes catalog and tell you about bands that were coming to town and tell you about albums you were missing from your collection and introduce you to other people with similar tastes and introduce you to other bands and music you might like based on your taste. Actually that idea was probably much more like Last FM than it's like Pandora. When I was in the process of working on these ideas I got introduced to this guy, Tim Westergren, who had a little tiny company in Oakland called Savage Beast.

In this company, there were about five or six engineers, a business development person, a couple of founders and that was about it. So there were nine or ten people. As it happened, they were looking for an engineering manager, somebody to come in and hire some engineers and manage software development projects which is something that I had done at that point in my career, but I had spent the last many years in a somewhat more strategic role where I had much more influence over product direction, was a member of the executive staff and did a bunch of outbound stuff. This job felt very constraining to me. But I really liked Tim and I really liked the team of engineers and I thought this thing, the music genome they had, was a really interesting approach to music recommendations.

Very briefly about the music genome project, the idea, which is still the reptilian brain of Pandora to this day, is that we have a team of professional musicians who come to the office everyday, that put headphones on and they listen to music one song at a time and score each piece of music on between somewhere 200 to 500 musicological dimensions. Not just are there guitars but what kind of guitars, how are the guitars played, what role does the guitar play in the overall composition?

What that allows us to do then is to find songs that are musically similar. If you imagine that we collected just two dimensions, maybe how breathy is the vocal and how much falsetto is used, and if you took every song and plotted it in two dimensions so here's breathy and here's falsetto. The really breathy stuff with lots of falsetto there'd be a bunch of songs up here, the stuff that has falsetto but isn't breathy at all would be here, neither and so on.

So you'd end up with this constellation of songs in two-dimensional space and the ones that are close together are similar. Here are all those that are breathy with falsetto and here are all the ones that have neither. So by computing distances between points, you can see how similar songs are in this two-dimensional space between falsetto and breathiness. We do exactly that but we do it in 400 or 500 dimensions rather than two dimensions. The math is actually very, very similar to the math you'd use to do it in two dimensions, it wouldn't be that unfamiliar to you. That's the basic idea.

They had this piece of software that did that and they had this team of musicians who were analyzing songs and, at the time, they were trying to figure out what to do with it. There were figuring it out by going and talking to other people that had a relationship with music consumers. They didn't want to be a consumer software company they wanted to power other consumer companies.

They went to AOL and helped make recommendation technology for AOL music. They went to Best Buy and Best Buy said, "We really need is something to allow people to listen to music in the store before they make purchases. We need some kind of kiosk. Could you make a kiosk for us?" Tim, the founder would say, "Well, we really don't know anything about kiosks, but if we made you a kiosk could we have music recommendation?"

And they go, "Yeah, yeah. That'd be great." So he'd come back and say to the engineering team, "Do you guys think you could make kiosks?", and the team would say, "Yeah, I think so." Long story, longer, and I'll try to wrap up here, because I know we're running out of time. I joined the company in the Spring of 2004. About two weeks after I joined, our CEO, Joe Kennedy, joined. Very quickly, Tim, Joe and I started talking about what kind of direct to consumer software company we'd like to be when we grew up.

Those conversations led to this idea that Joe had about "one click personal radio," he called it. Within six months or so, we were off building what became Pandora.com. We changed the name of the company somewhere along the way. The rest is just simply the last seven years of my life. It's been incredibly gratifying and by far the most fun thing that I've done. I think we're kind of running out of time, but I know there may be some questions.

Audience Member: What's your guy's overall philosophy [inaudible @36:26] all the devices now being connected to the internet. I know that earlier you embraced that.

Tom: Yeah.

Audience Member: [inaudible@36:33]

Tom: That's a big dimension of what Pandora is about these days. When we launched Pandora for the iPhone, unbelievably only two-and-a-half years ago, I think at the time we had on the order of 13,000,000 listeners, which is pretty good for three and a half years of the life of the service. Fast forward to today, we now have 65,000,000 and a lot of that growth is on the back of this explosion of mobile listening.

On the mobile side, we learned the hard way that part of the magic formula for success for Pandora is focusing on smartphones. We did all of these experiments with feature phones [inaudible @ 37:14] and [inaudible @37:15] and they were never successful in any interesting way. Something about the smartphone customer and the form factor is a regular headphone jack, and the expectation that it's a music device in addition to being a phone, has been a real catalyst for us. We do all of the development in house.

Interestingly, we've never really hired a mobile-centric engineer. Everybody that works on our mobile products started as a web developer, a back-end developer, and migrated over. It's encouraging that you can be successful with that model, because it is terribly difficult to find people who have tons and tons of deep mobile experience. We found that really talented engineers can make that transition pretty easily.

On the consumer electronic side, Pandora is also available on more than 100 consumer electronics devices, like television, and Blu-ray players and set top boxes and those kinds of things. We have a slightly different model in that case because there are all of these different platforms. We basically have an API that we've licensed to our partners. Samsung has a team of engineers that built the Pandora tuner, the software that lives in the Blu-ray player, and their televisions that use our API, and all of our server side infrastructure, but they do the development and certification work. They send us a sample, and we run through a series of tests and certify that it matches our standards. There's a developer programmer and developer support and a certification process and documentation and sample code that goes along with that. We don't have to do implementations for all of these different devices in the typical case.

Audience Member: [inaudible @38:56] Do you have a guideline on it?

Tom: We do. For ten foot UI's, for things that run on televisions, we have a reference design, that includes a bunch of information on how interactions should work. It's actually a really tricky question though, I have to say. TIVO is a really good example. TIVO recently launched Pandora. It's been very, very well received by the TIVO audience. TIVO has a really exceptional team of interaction designers. They have a whole aesthetic and set of standards for their platform. On the one hand, it would be tempting for us to tell TIVO, "You can't use the standard TIVO interface when they activate Pandora, because it's too hard to support. We think it's confusing. You made it for television. You didn't make it for music," and so on. But, we haven't done that. We said, "You guys know your customer. They've been trained to use your product. We're happy with you in implementing the TIVO look and feel against our service. I think the same thing has happened with Netflix and Apple TV. Netflix in the context of Apple TV looks more like Apple TV than it does, like Netflix, and I think that that's largely to everyone's benefit. It's to the consumer's benefit, it's to Netflix benefit, it's to Apple's benefit. I think everybody wins and so if we we're to ever end up on Apple TV, I think that we would likely be very open to doing the same.

I think where it gets a little bit confusing is like when we're on a Blu-ray Player, it's sort of like Blu-ray manufacturer X from Korea. They don't have a strong sense of user interface design. All of the bits and pieces of their stuff looks different from all the other pieces and that's an area for us to just say you know your customers better than we do. Design something probably isn't the best answer and so we're still kind of sorting out like when do we really assert ourselves and when do we let - I think the unfortunate answer is, it depends.

Audience Member: Would you say that the end product is an audible product makes it more difficult or easier [inaudible @41:12].

Tom: One of the really difficult design constraints for us is that we think that part of the reason that Pandora has been successful, and something that we are trying to do from the very beginning, was to emulate the best of traditional radio. We never really thought of ourselves as a website. We thought of ourselves as a radio service and radio is really simple. You press a button and music comes out and so, on the one hand it means that you don't have very much user experience surface to work with really, like presets, pick a channel, listen, turn it up, turn it down, pause it, skip, a few little things.

But as soon as you start to move outside of that, it stops feeling like radio and starts feeling like a website or an application. But the good news is we have a bunch of constraints that help keep us honest, but for those of you who are designers, you know the more constraints you put on problems, the harder it is in many cases, and so we're constantly sort of wrestling with this. How do we keep it really simple but continue to innovate and expand the capabilities of the system and add new content and those kinds of things without violating the core tenants of the product.

Host: If you guys have any more questions ask away. Let's thank him for his awesome soapbox.

Tom: Thank you so much for having me. It was great.