David Baggett is the co-founder of ITA Software, a company that Google bought for reported $700 million. ITA makes software that’s the search engine behind sites like Kayak. It actually figures out prices, time and other data like flights.
Today he’s launching Inky, a mobile app that makes clearing your inbox easier. It will categorize your email for you. Help you unsubscribe from junk and quickly respond to what’s important.
I invited him here to talk about his past companies. I want to ask him about Inky, including a couple of challenging questions that some of you may have. I think this is the right app for me, but I have one big question that we’ll talk about later on.
David Baggett, ITA Software
David Baggett is the co-founder of ITA Software, whose software created an up-to-date database of flight information from airlines.
Andrew: Hey there, freedom fighters. My name is Andrew Warner. I’m the founder of Mixergy.com, home of the ambitious upstart. The place where over a thousand of entrepreneurs have come to tell the stories of how they built their businesses, had challenges along the way, overcame them, or not, and mostly told you what they learned from it. And thousands and thousands and thousands of other entrepreneurs have listened and learned, and used what they’ve learned from these interviews.
Today’s guest is David Baggett. He co-founded ITA software, a company that Google bought for a reported $700 million. ITA makes software that’s the search engine behind sites like Kayak. It actually figures out prices, time and other data like flights.
Today, after selling, he launched, and is running, Inky, a mobile app that makes clearing your inbox easier. It will categorize your email for you., it will help you unsubscribe from junk. And it will let you quickly respond to what’s important.
I invited him here to talk about his past companies. And personally he’s a guy who gets a lot of email and doesn’t want to waste time on it. I want to ask him about Inky, including a couple of challenging questions that some of you may have. I think this is the right app for me, but I have one big question that we’ll talk about later on.
First I should say that this interview is sponsored by Scott Edward Walker. Go to WalkerCorporateLaw.com. Later on I’ll tell you more about him, but I’ve got to welcome David. David, Thanks for doing this.
David: Thanks a lot. Great to be here.
Andrew: You sold, it’s 700 million dollars. Which bank account is it in?
David: It’s in checking, I have it all in checking.
Andrew: I thought so. Do you ever go to the ATM, print out a receipt, look at it and go, “700 million dollars, I did it.”
David: Usually it doesn’t work that way. When you have this kind of exit, lots of people are involved. We raised a hundred million dollars so we had investors who owned more than half the company. There’s a lot of dilution any time you get an exit that big. There are a few counter examples. I sure Facebook, that was a big one for Mark Zuckerberg. But usually when you have a big exit or a big IPO it’s been carved up into lots and lots of little pieces.
Andrew: Are there so many pieces that right now you have to get a job at the local Wendy’s while you’re building this app.
David: No it’s not like that. I can self-fund Inky. And that’s what I’ve been doing so far. Obviously, any kind of exit like that is life changing at some level, and it was for me too.
Andrew: When you say, “Life changing,” what, specifically, about your life now has changed, and is better because of the sale.
David: Well, you don’t have to worry about money things on a day to day basis. That’s a huge difference. Just personally that removes a whole stress from your life. I had a little bit of that from Crash Bandicoot. I worked on game called Crash Bandicoot in the back in the 90s. I actually didn’t own equity in that company, Naughty Dog, But I did have a royalty on the game. So one thing it let me do is invest in ITA. So part of the reason I had shares in ITA when it sold is because I was kind of the angel investor, or one of the two, in the beginning of that company.
Andrew: And we’ll talk about one of the strings that came with investing in that company in a moment, but the reason that you were able to do it, to code and to angel invest and so on, seems to me that part of it goes back to something you and your dad did together. Here’s what you told Jeremy Weis [SP] in the pre-interview.
“I taught myself to code. I’d read source code listings out of a magazine.” What is that.
David: Back in the day, in the 70s you didn’t really have floppy drives, you had floppy drives, but you didn’t get software on floppy drives. You’d get these magazines like Creative Computing, or Byte Magazine. And you’d just type in a program listing. And the thing about that was that it would never work because the Basic was a different dialect. So it was a great learning experience. You’d have to figure out how to make it work. I’d type in something that was meant for the Apple II computer into my Heath Kit computer that my dad and I built. I kind of helped. As a seven year old you can imagine what that help was like. He really built it. I’d type them in. He often says now that he never got to use the computer after it was built, because I just completely monopolized it.
Andrew: You just knew that’s the thing for you. What was it about the computer that made you feel, I want to sit and work on this thing?
David: It was like an infinite toy box. The thing that still really appeals to me about software is that you can create something from nothing. And it’s just such an amazing feeling to be able to literally imagine something and just make it. My son now plays Minecraft. I’m sure everybody knows about Minecraft. I ask him, why don’t you just go across the street and play with your friend Noah? Why do you play on Minecraft?
And he says, “But dad, I can’t build a giant building in the real world. With Minecraft I can make a building.” And software’s just like that. It’s all virtual. And the fact that it is means that someone like me, a seven year old kid, could make a game and actually see it come to life. And of course, at that time the games were just primitive. But it was still just magical, this ability to create an infinite number of toys from this toy box.
Andrew: But you did actually make a game at that age.
David: Yeah, I made a little space invader type games, mostly character graphics. They’re really primitive. I think the first program I actually sold was a game when I was like 13. It was a little space invaders game for a basic programming book, so very, very primitive. One of the liberating things about that was that commercial games were very primitive too because you just couldn’t do that much with the hardware. So the gap between what I was able to as a young teenager and what was commercially available was not that big. And now I think it’s a totally different world. It’s daunting for kids now because there’s so much to learn to make anything that’s even a basic app. And I just can’t imagine what it’s like to be someone like me when I was seven, now. I think a lot about how should we teach kids how to program. And how do we get kids to get kids to get excited about computers like I did.
Andrew: And it seems like the answer is to get them to make something of their own as quickly as possible, something that was in their head, and then is now in the real world.
That’s the other part of you. You were selling. You actually sold your mom one of your early programs:
David: That’s right.
Andrew: What did you sell her?
David: Actually, the reason that my dad built the computer wasn’t for me to have a toy. It was that my mom was a freelance writer. He told her, “Hey, there are these things, computers now.”
And she said, “Well why do I want a computer? Why would I want that?” And he said, “Well you can just edit stuff on the screen, you don’t have to retype it from scratch.” And she was like, “Sold.” So basically this very early word processing machine, it was meant for my mom to write a book. So indeed she did.
Back then, it’s hard to imagine, but even the idea that you would double space when you printed, or that you’d put a header and a footer and a page number on every page, the word processors didn’t do that by default. They didn’t come with that. That was like a fancy feature. So I’d write her a little program that would take her documents and put headers and footers on.
And I actually found, recently, a manual that I wrote for one of my programs. I’m this 12-year-old kid. I wrote a manual. I put my phone number, which was also her phone number, of course, in there and said, “If you have trouble, call this number for Tech Support.” For me it was like pretending to be a real software guy, a real software producer. I obviously must have liked that at the time. That was part of it for me.
Andrew: Did your dad get to see – Is your dad around still?
David: Yeah, he is.
Andrew: He is. So got to see what you got to do with this. He was an electrical engineer who got you started. How did he feel when he saw you coding and producing and creating professionally?
David: I think it’s probably just like any other kid with a talent. Your, at some level, probably amazed by the talent. And your also sort of dismayed by how much time it takes. Because I just became absorbed in it to the point where I would spend all of my waking hours on it. And that I was all the way through high school, and then in college. Then at some level they probably wondered, my parents, whether this was actually a good thing over all, that it’s becoming this obsession.
But I think it was pretty gratifying for him to see that, especially since I tell the story about building that computer with him. I mean this was a Heath Kit, which nobody’s heard of any more, I’m sure. But they made Ham radios and stuff. And in this brief period in the 70s they would sell you a kit computer. Not like a kit that you put the video card in. No, a kit you solder every resistor onto the motherboard. Since my dad was an electrical engineer he was able to do this, but it took six months. Every few days we’d work on it. He’d say, “Get this color resistor out of the box.” And I would go do that. And he’d solder it in.
And I thought the whole time, “This is never going to work. My dad’s gone crazy, this is just . . .” I’m humoring him while we’re making this thing. And then, after 6 months, he turns this thing on for the first time and this little prompt comes up. And it was an H: because it was a Heath Kit, so H. And I was just amazed because, Oh My God, my father is a God. He can make this thing. I couldn’t even imagine how this think worked. All of a sudden it just booted up the first time. It was just a magical moment for me.
Andrew: I can’t imagine that. My dad made kabobs.
David: You know at some level you still – there’s a craft to that, too. So this idea that you’d use a soldering iron to make something, it was exciting.
Andrew: I used to watch it, it would be like magic because when I was a kid I got the Radio Shack, I forget what they are but you didn’t need to solder anything because there were springs that would attach the wires together, and I couldn’t get that to work.
Andrew: And I would have gone mad if I sat and typed out every line of code from a magazine and it didn’t work. For you it wasn’t something that would make you say, “Forget this, It doesn’t work, they cheated me.” It was an opportunity for you to say, “Something’s a little off, because I have a different Basic than they do. What do I do to fix it?”
David: Yeah and I remember there were tough times, though. I remember one night when my parents went out and I had a babysitter. And they decided to do nothing but play on the computer all night. And I remember I typed this whole thing, and I hit some key and it erased everything. It was just this absolute shock and tragedy for me though. So and I still remember it even though I was seven at the time. I still remember, “Oh my God, my program is all wiped out” and it’s like a terrible thing. So it wasn’t all . . . I was still a kid, so I still remember getting really frustrated and giving up and coming back to it.
I certainly wasn’t robotic in my ability to deal with the challenges, but yeah, I mean, I would get these little basic programs and type them in and then try to figure out why they didn’t work.
Andrew: How did you get into creating games professionally?
David: Well, see, that’s a great story because I had always made games as a semi-professional. I actually had a game development system that I made with a guy actually in Scotland. We collaborated over the Internet and it was like, you know, mid-80s. So this was early. We’d send tech files back and forth over the Internet that were interesting coded binary. It was just very primitive file sharing and we made this game development system. But it didn’t really get traction, and so I was never really a professional game developer.
And then I was fortunate to get into the artificial intelligence lab at MIT for graduate school. And they have an orientation where once they’ve decided they want you then they try to sell you on going there. They invite you up there. So first day, literally, the first day I’m there the first guy I met was this guy, Andy Gavin (sp). I’m talking to him, “Yeah, I really like making games.” He said, “Oh, I’ve sold six games.” And he talked about, you know, “I made this game for the Sega Genesis” and it sold 70,000 copies. So I guess, wow, that’s what happens when you go to MIT. You meet someone who had been selling games and he’d been selling games with his friend, Jason, since they were in seventh grade.
And so these guys were just amazing. They sold six games commercially when they were kids, and I just thought these guys were . . . First of all, they were really great friends. And second of all, you know, they wanted me to be their first employee when they actually went out and did a AAA game. Triple A is kind of the term for big budget, you know, put out all the stops, get a great game. And now AAA games are $400 million, but then it was a million and a half dollars.
So when they asked me to join them as their first employee, I bailed on my PhD and just left, and that was the beginning of my involvement with Naughty Dog and Crash.
Andrew: Oh wow. And then you weren’t as . . . Well, I guess you were an employee who got royalties on the creation of the game.
David: Yeah, that’s right.
Andrew: I heard from Jeremy that Sony would Fax you bug reports. Talk about bug reports.
David: Yeah, so we had, you know, two of us programming on the first game. It was just brutal because that’s not nearly enough people. We had seven on the team and two programmers. So we were just working crazy hours and towards the end, as you said, we would get bug reports from Sony. They had 60 testers in a room testing all day, and they would test all day and then they would Fax over these pages and pages of bug reports.
And so Andy would get them off the Fax machine at two in the morning, and we would start slogging through bugs. It was brutal, but it was obviously one of those things that you remember forever because it’s unforgettable that you’re there. And it’s the middle of the night, and you’re working on bugs and you’ve got to ship the game in two weeks. But, yeah, we used to get bug reports literally on the Fax machine.
Andrew: The shocking thing to me about that isn’t just the Fax machine, it’s just that there were people who were just paid to play games and look for errors in those games.
David: Oh man, you’d think that’s fun, but actually it’s grueling.
David: After you play the same game 12 hours a day for three months, you start to go crazy. [laughs]
Andrew: And after you’re doing this, there are other MIT . . . Were they graduates at the time who were working on ITA?
David: Yeah, well that started actually around the same time we were working on Crash 1.
Andrew: What was the problem that they saw that made them say, “It’s time to create software.”
David: Well, they looked at things like Savor. Savor was incredibly impressive. It was developed in the 60s by IBM and American Airlines, but it was kind of built on mainframe. It hadn’t changed much. It was still kind of based on this older technology stack of, you know, mainframes and the language called TPF. In particular the search function where if somebody, a travel agent, wanted to figure out for a customer who called because, of course, they would call on the phone. There was no Internet yet. There were no Internet sales of tickets, at least.
So someone would call on the phone. They would say, “I want to go to Boston and I’m in LA.” And the travel agent would type in to this thing and it would do a search. And it would figure out, what are all the ways to go from LA to Boston, and what do they cost.
At that time it was all text based systems. They weren’t hooked up to anything like a TCP/IP connection that you’d need for a web site. And they did a really poor job. You’d do something like Boston to LA and you’d get 9 answers, all on United. And it was kind of like, Well, I, know there’s other airlines, what happened to them?
Andrew: Why didn’t they show other airlines?
David: It was a hard problem. Partly it’s a really hard problem to come up with a representative set of solutions because, for any given query like that, if you look at something like a transatlantic flight, Boston to LA, there’s a thousand reasonable ways to go Boston to LA. And of course there’s a thousand reasonable ways to come back. Right there, before you even talk about prices, you’ve got a million combinations you’re looking at.
And the challenge is to figure out what all those million cost by applying the fare rules, and making sure this fare can be used with that fare and so on. And then to provide a summary of that search space is really challenging. So the way all the systems worked then is that they used heuristics. And they’d say, “Well, it’s really good to go through Chicago, so let’s do some of those and see what they cost. And maybe out of Dallas or something.” But is was all rules of thumb. It was heuristics based.
So their insight, Jeremy Worthheimer [SP] started the company and Carl de Marken [SP] came in and wrote the majority of the early system. Their insight was that we could actually do a comprehensive job of this and find all the valid solutions. And then we could take that huge space, the number of solutions in this would be like 10 to the 32, giant, giant numbers, and then find a way to produce a summary that covered that space and summarized it. So it was just a very hard problem.
Andrew: So is it that the raw data would say, if you want to go from LA to San Francisco, there’s also a way to go from LA to New York to San Francisco, but you needed to figure out that that should be excluded and others need to be included. Am I picking up on it?
David: Yeah, if you think about it there are trade-offs, right? You might have an inconvenient itinerary where it’s 2 stops, but just the way the fares work out, it’s cheaper. So you still want to show that inconvenient 5 dollar cheaper thing, or 50 dollar cheaper thing. What you need to do is generate a thousand outbounds and a thousand returns. And you price out all the combinations. And then of those millions, there’s typically a millions ways to price a single itinerary because the fares are so complicated.
The airlines didn’t intend to do this but their rules are so flexible that the combinatorics are absolutely staggering. You want to look at all of these possible combinations because you don’t want to miss something good. But there are just a lot of them, a very large number.
Andrew: I wouldn’t have thought, it seems like it’s just so easy to do that the only epiphany was that Carl said, “You know what, I can do it online,” or that Jeremy saw that the internet could be added somehow, like magic. But now I’m seeing what you’re talking about. So if it’s the 2 of them working, why do they need an investor? Why do they need money? Why can’t they just continue to figure out this problem on their own.
David: I knew Carl really well because he was my office mate at the AI lab and we studied computational linguistics together. We were in the same group, had the same adviser. So I knew Carl, knew what he was doing and he showed me the early prototypes of the product. And I thought, “This is really cool.” Basically, what the state of the art is, is right now is 9 answers, and Carl’s got this 500 answer awesome prototype. And I really wanted to be involved as an investor.
And Jeremy was very canny because he said, “You can invest but only if you also write code.”
Andrew: I never heard of that before. You have to write code as an investor.
David: He basically said, “I’m not going to let you invest just passively. You can invest, but then you’ve got to do work. You’ve got to help us,” very clever in that he needed and programming. So he basically got both of them at the same time. I wasn’t the only investor. His father put a bunch of money in. Eventually we actually got a deal done with a company called Amadeus which was the European Sabre. So at that point the company was self- sufficient. But at the early stage it was on Jeremy’s credit card, that kind of phase.
Andrew: What did you envision for it? What did you say that made you want to invest in it? And by the way, earlier when I said, “Inky,” I accidentally slid my phone into the garbage pail which is always empty. So I’m going to pull it out. There it is. No it’s not empty at all. It has the dry cleaning bag that this shirt came in. Phone is back.
David: There you go.
Andrew: So what was your vision? Why did you say, “I want to be a part of this because I see the future.” What’s the future?
David: I thought it was a really interesting hard problem and I thought that would be fun to work on. We had no idea how hard it would be. Even though we knew it would be hard, we had no idea. It was this hubris of naivete. Right? But I also thought, Jeremy just went through the numbers around the travel industry and at that time it was two billion flights a year. Travel’s huge. If you look at the amount of commerce done, the dollar amount of travel, it’s vast. Indeed, now travel sites are probably the biggest spenders on google search. It’s just enormous amounts of commerce involved in travel.
So I thought the idea that we would be able to get some tiny fraction of that was a really compelling business opportunity. Now I would call it, “The total addressable market.” The total addressable market was really large for that space.
I wasn’t thinking of it that way at the time, but I could tell this is a big space. Also I kind of liked the idea of working on something that people would understand at some level. Like games, people understand, travel, they may not know how it works, but they know, oh I get it, when I go to Orbitz something has to go from point A to point B. Your software does that. And that’s kind of gratifying.
Andrew: This was 1997.
David: ’97, ’98. They had a prototype by the end of ’97.
Andrew: Were you thinking, “This is going to power other websites.” Or were you thinking web sites at all?
David: We were. I would like to say that the whole thing was planned from the beginning to take advantage of the shift of ticket distribution onto the internet. It wasn’t nearly so sophisticated, but we did, obviously, know the internet. We followed Netscape and things like that. We knew browsers were starting to become important.
It really wasn’t clear in 1995, yet, at least to me, how rapidly companies would adopt the internet. I remember saying to somebody at the AI lab, “Yeah, this thing’s going to take five years at least, maybe 10. Because companies have to be on this thing. Otherwise nobody’s going to care. And how long is it going to take companies to get around to making web sites. It’s going to be a long time.” And indeed it was, like, 8 months. And then every major company had a web site. It was fascinating how fast it happened.
We knew about this, and we knew it was going to happen, but we didn’t know how rapidly.
And the other thing that was a structural opportunity that we weren’t thinking about but that was a huge advantage for us was that because nobody was really selling many tickets on the internet yet, the airlines didn’t care that much about that channel. Which meant they were willing to trust a crazy company of MIT guys with their business on that channel, because it was kind of de minims. If they were putting us in the middle of 75% of their ticket sales they’d be really nervous. Right? Unproven technology. squirrelly guys from MIT.
But that fact that it was a channel that was just starting to become relevant was really valuable. Because we could get in there. And then all we had to do was not blow it. Right? As the channel exploded we just had to grow and maintain our operational integrity as the volumes went up. And we kept the business.
It’s a really interesting lesson for entrepreneurs, I think. Usually when something like this happens there’s some structural opportunity that you end up taking advantage of. I think a lot of successful entrepreneurs like to say, “Of course we thought about that from the beginning and we were geniuses.” But I’m not so sure that’s true. In our case we kind of had this insight but it wasn’t planned out from the beginning.
Andrew: Was there a minimum viable product?
David: Yeah and one of the things that’s fascinating about both travel search and email is that the minimum viable product is really, really not minimal at all. And it took us a long time, in both cases, to learn what that minimum product was, and to get to it.
Andrew: What is the minimum viable product? It can’t be as easy as, “I will manually do it for them.” What does Eric Ries call it, the concierge MVP, which is the simplest, it seems, of all the MVPs. What did it have to include.
David: For travel search, if I talk to a programmer who doesn’t know anything about travel, but they’re a good programmer, they’ll typically say something like, “Well isn’t looking up the price a database transaction?” They assume there’s just a price somewhere you look up. And in fact it’s much more like assembling a million piece jigsaw puzzle for every query. You know it’s lots of combinations of fares. And fares have rules that say, “Well I’m allowed to be used if there’s queue availability, but I can’t be used if this other fares used. And I can’t be used on Tulsa’s trips. And I also can’t be used between 5 PM and 7 PM.”
So there’s just these rules that go along. And there’s hundreds and hundreds of fares in each market. So the minimum viable product was one where you read in all these rules, you read in all the fare data, which was not trivial, it wasn’t easy to read the fares. It wasn’t rocket science hard, but it was slogging to figure out how you’d just interpret this data. Originally they’d give it to us on big mag tapes, like, literally, those tapes you see on old Star Trek Episodes, the big tape drives.
Andrew: They’d ship them to you. You’d have to put them on an old tape deck looking thing upon a reel to reel looking thing. And then you’d pull out all the data. So it wasn’t even up to date.
David: No and it was just test data too. So in the beginning we had to write a device driver for Linux to talk to this Kennedy tape drive. So there was that kind of scut work around just getting the data and understanding it. And that was a decent amount of work. But to have a real minimal viable product for travel search you’ve got to do all that rule validation properly. There’s thousands of different kinds of rules that the airlines have.
People know about things like Saturday stay restriction and that sort of thing. But there’s just thousands of these things, and lots and lots of detail. Long story short, it basically took 5 years.
Andrew: Five years to get to the MVP.
David: And it’s similar with email. With email people don’t think about it, but there’s just vast amounts of complexity. I like to call it a fractal problem because the more detail you get into the more you realize there’s all this stuff I didn’t know I had to do.
And the thing that I think is interesting, from a lean start-up standpoint is, how do you think about a lean startup in the context of a five year MVP. Because not all software problems are equally difficult. Some are harder than others. How do you fund something where it takes 5 years to get to the MVP. And I think these are interesting questions for the startup philosophy, the kind of Eric Ries lean startup narrative.
Andrew: If you have an MVP that takes 5 years, presumably, not presumably, the reason you have an MVP is to show to a customer and say, “Is this what you guys were looking for, yes or no?” And if not, I’ll go back and tweak it. In fact, even if it is yes, I’ll go back and tweak it. Can you take back and tweak something you spent 5 years creating. Can you get back and adjust it?
David: Yeah, and to be fair, we weren’t waiting 5 years before we showed it to anybody. We were showing it to airline people the whole time. It’s just wasn’t really viable in terms of being usable by any of them.
Andrew: I’m assuming airlines did not use it until 5 years later.
David: There are other reasons for that. There were internal conflicts within Amadeus as to whether they should use our engine or their own, so they didn’t really end up using it, but not for that reason. But let’s say somebody like American Airlines, somebody like that, they would give us feedback, they’d say, “Wow, this is really cool, and this is going to be great but you’re getting all of our prices wrong because you’re not doing these rules properly.”
Andrew: I see.
David: We could show them something that kind of looked and felt like the end product, but it wasn’t really minimally viable. And that’s really true of email too. You can make client in a weekend, but it’s not really minimally viable because no one could actually use it instead of mail.app or Outlook. That’s another multiple year effort to get to that point.
Andrew: There’s one company that I saw that did that, Triage, that seemed like that was an MVP that all you can use is reply or archive.
David: The problem with that that we found, at least with email, is that people won’t use multiple email apps. So you either wholesale replace the thing that they’re using, or they’re just not going to use it.
Andrew: Oh, yes. Yes. Right.
David: There are services where they help you unsubscribe from mailing list. The problem with those is that they can’t become someone’s primary mail client. And most people don’t want to use multiple mail solutions. They want to have the thing that comes in their phone, or an alternative app which replaces that. So every single thing that a mail client does, you’ve got to cover somehow, whether it’s attachments or the handling of drafts. And there’s still things that people say they want out of Inky that we don’t have yet that are on our list to work on.
So minimum viable, there’s sort of the minimal thing to show to somebody to gauge market reception to it. And then there’s minimum viable from the standpoint of, someone could use this instead of the alternative. And they’re really different for a mail client or a travel search engine.
Andrew: Your first client was Orbitz, 2001.
David: Yes, they were the first, they were kind of our marquee client. They were the first ones that really put it into production. We also had an airline called America West, which people may not remember, but they sort of evolved into US Airways and now Doug Parker runs American. He was the guy that started America West, or the CEO at the time. And so we had America West and Orbitz.
And there again it was a fortuitous opportunity because Orbitz needed something differentiated from Expedia and Travelocity, and there we were with this new, fancy search engine that did different stuff and produced a very different-looking kind of output. So again, a lot of what makes things successful, I think, is being in the right place at the right time with the right stuff. And so, that was the right place at the right time because of Orbitz and their need for something at that time for something differentiated.
Andrew: Wasn’t Orbitz founded by the airlines?
David: Yes, Orbitz was founded by the airlines as a way to try to challenge Travelocity and Expedia, so that those two entities wouldn’t have too much leverage over the airlines.
Andrew: So did it make it easier for you to get them as a client knowing that you were collecting their data, they had seen what you had done with their data, and so they trusted it? That they trusted it more, also?
David: Yes, I mean, I wonder if we could ever get the majors without that. I mean, I don’t know – I would like to think that we would win them over, but certainly the idea that we were powering Orbitz and already pricing their stuff properly for consumers, you know, that’s kind of like, they get to try before they buy. It was a huge benefit to us to be able to do that. And Orbitz, we gave them incredibly, incredibly good deals, so its not like they gave us that privilege for free.
Andrew: Did you try to sell to others before Orbitz, and get turned away?
David: Yes, we had a really long interaction with Delta, starting probably even 97-98. And there, it just wasn’t the right time for Delta to use an external system. We weren’t really ready in the sense that we weren’t at that minimum viable product level. And so they would say you’re pricing all of this wrong, and they’re right – we were pricing a lot of stuff wrong and it took a long time for us to code in all of the rules and then, all of the little edge cases.
One thing I like to talk about is these edge cases. Let’s say the fare requires a Saturday stay over and you cross the international date line, and it was Friday on both, so you basically left Friday and arrived Friday, well, did you have a Saturday stay, or not? So there’s all these, when you get into real world, and time zones, and time, and things like that you get a lot of crazy edge cases and it takes a long time to figure out the industry’s answers to those, because it’s not like anybody wrote those things down.
So, yes, we had an early engagement with Delta that ended up failing, essentially, and just incredibly painful, and all the usual things that one could say about that, but we learned a lot from it, but it would have been much better for us if we had gotten them as a customer.
Andrew: I just realized I forgot to say that my sponsor, or to give the sponsor spot to Scott Edward Walker of Walker Corporate Law. I say all of the time that entrepreneurs who get started should get their company in order before they get into anything serious. Don’t save that until later. I talked to entrepreneurs, in fact, just before we talked: a half-hour before I finished a conversation with an entrepreneur whose co-founders just left her, right as she was getting going. Situations like that come up – you need to be prepared, and the way to prepare is by talking to a lawyer early on, setting up the paperwork early on, and allowing yourselves to when the problem happens, to not just deal with just the problem, not both the problems and the legal issues that go along with how do you separate, how do you make sure that the person you now hate doesn’t own half of your company for the rest of your life – all of those things come up.
Talk to someone who has seen them many, many times – that’s the person who I’m recommending you talk to – Scott Edward Walker of Walker Corporate Law. Frankly, he’ll help you get started, but that’s not where he focuses – his big focus is on those big transactions, the buyout, the sale, the raising money, that’s where he comes through for you, that’s what he’s experienced in, but of course he wants to help you get started if that’s where you are. Go check him out at walkercorporatelaw.com. David, by the way, do you prefer to be called David or Dave?
David: Dave’s fine, either way.
Andrew: Was it awkward when I called you so formally at the top of the interview, David?
David: No, no. Not awkward. Really, it doesn’t matter to me.
Andrew: You know, I don’t know the guy well enough yet. I saw your Executive Assistant refer to you as Dave, I said alright, that’s for her. I haven’t yet gotten to know him on that level. So Dave, it seems like everything is just falling into place for you. Your Dad set you up, right, by showing you how to set up a computer, frankly, it all is possible. But helped you learned to code, then you created a great game, then you got into this. Was there any challenge along the way, or was it all as easy and simple and straightforward as it seems?
David: Certainly. Don’t forget that that was 1977, too. Very few people had access to a computer. I remember I heard Max Levchin, I think he was on NPR, talking about this. There was a whole generation of us who just lucked in to having access to these devices so early in our lives and early in the timeline of this stuff. That was incredibly lucky, fortuitous and all good.
I will say there’s been a series of really difficult challenges. For example with Crash we had no idea that that was going to be Sony’s mascot game. And if it hadn’t been we would have sold 100,000 copies instead of tens of millions. It’s just as simple as that. Right? In every one of these cases where it’s been successful there’s been a right-place right-time element to it. In every one of these there’s just been tremendous amounts of slogging. That’s the thing that people don’t see.
Andrew: What do you mean by, “Slogging?”
David: Just thousands of hours of doing boring stuff. With, for example, the airline data, you would think about something as simple as time there’s obviously a flight departure time. Well that’s local time. So then there’s a time zone thing, where it says how many hours off of GMT it is.
Andrew: I see. So what you’re saying is it’s not so much that you got a client who turned around and cut off relations with you, or something dramatic happened where you got everything up and running, then the whole business came crashing down. It’s not that kind of challenge. The challenge of the slog, of sit down there, you have a difficult problem, don’t get up until you do it, until you solve it. And then when it seems like the solution is great, but it turns out not to be great, you have to go back and do the work. It’s do the work type of challenge.
David: Yeah and that’s partly because of the domains I’ve worked in. Both travel and email, they’re characterized by a lot of this data and detail and slogging. I’ll just give you a sense of this with travel. Aviation Taxes. You know when you look at your ticket you paid a 9-11 fee and you pay a departure fee and there’s all these little taxes, at ITA we had to create a tax language. And that file of tax rules for aviation taxes for the world was 100 thousand lines. Someone type in, a bunch of people typed in 100 thousand lines of, essentially, tax code to properly encode all the aviation taxes. And, by the way, if every country in the world changes their aviation taxes once a year, you’re looking at a change or two a day. So you’re not just doing it once, it’s constantly updated.
And similarly with email there’s 20, 30, 40 thousand mail servers. And they all behave slightly differently. Mostly, they behave the same, but there are little quirks. So you’ve got to slog through figuring out what are quirks.
So it’s a lot of this kind of manual data entry stuff I think that a lot of developers, or just people in general who work on hard problems, they can feel like it’s beneath them it’s boring. Every single thing I’ve worked on involved a lot of that.
Even with Crash, I remember manually tuning every single turtle in the game. The turtles that move around, someone had to go in and tune every turtle in the whole game.
Andrew: What does tuning a turtle mean?
David: It means the turtle walks across the path. And it has to be timed right so that the game play works well and it’s fun, and it’s not too hard, and the camera’s not in the wrong place for you to see the obstacle, and all these little values you set for each turtle in the whole game.
Andrew: Sometimes I feel like all that king of slogging, isn’t that what successful people just give their interns to do? Don’t you just say to that intern, here’s the tax law. Go do it?
David: I don’t think so. I think most of the people who are hugely successful have done tremendous amounts of slogging. I think that it’s not something that the media focuses on, and it’s not something people talk about. I know that the early guru people did a tremendous amount of slogging. They were working hundred hour weeks just like we were. Every one of these things that’s substantive requires a lot of slogging. And I don’t think it’s easy to delegate it. Because, imagine that tax file, maybe you can give that to an intern, but it’s not, you kind of have to have the domain knowledge. Right? You can’t.
Andrew: Yes, I was thinking, maybe you could give that to an intern, but if things screw up, oh boy, that is horrible.
David: You’re right.
Andrew: You don’t want to screw up taxes.
David: Yeah, you really don’t want to screw up taxes.
Andrew: You were COO until 2007, then after that you became a board member, or you stayed just a board member, right?
David: That’s right. Yeah.
Andrew: Why did you take a step back from day to day operations?
David: Obviously part of it was just burn out. I and other people on the senior management team had been doing multiple jobs each until we raised money. We raised a bunch of money in 2006 and we brought in a lot of management help. And that was a challenge in some ways because we had to figure out how do we transfer all this knowledge to the new people, how do we transfer the kind of cultural values to the new people. But obviously doing one job was much easier than doing five.
Also I did a lot of travel and technical sales where I was trying to convince people to give us data that we needed. It was just a lot of hard work, and there’s an element of burn out to it. So there was an opportunity when we raised money for me to take a step back and not take a break so much, but just figure out what I wanted to do next.
Andrew: And so you were thinking of what to do next before the sale to Google, which happened, I think it was three years later, right?
David: Yeah. So the Google sale was 2010. Then we went through this whole process with the Department of Justice for almost a year. So it didn’t really close until 2011.
Andrew: I remember that. The first thing I remember thinking when the sale happened was, “Who are these guys? You mean Orbitz doesn’t just go to the airlines and do it all themselves?
David: Right. Nobody knew who we were.
David: We were pretty well known in the travel industry, but certainly the high profile that Google had made people realize, “Well who are these guys? How can something I’ve never heard of be worth 700 million dollars?”
Andrew: And impact so much of the stuff that I use, and I still didn’t know that it existed.
David: Yeah, we maintained a really low profile because we were mostly focused on selling stuff to businesses. Our customers were airlines and travel intermediaries. To a large extent they didn’t really want us to have a high profile. It wasn’t beneficial to them. So we kept a pretty low profile.
Andrew: I see. Why did you guys sell?
David: There’s a point in every business where you can double down on the risk, you can exit, or you can raise money. Those are your 3 options, generally. With any business there’s a point where you look at it and say, “For us to take this to the next level, what will we have to do?”
Around 2005 we realized for us to take it to the next level and, long term, defend ourself against competition from the Sabres of the world, we would need to build our own reservation system and start hosting airlines. We would have to not just be a search provider. We would have to provide the whole platform for a major airline. So that’s the reason we raised money, because we wanted to start this reservation system effort and get to parity.
And obviously parity’s the first step. And then you go beyond there. We thought we’ll bring web technology and 21st century thinking to this reservation problem as well. And indeed we did. Ultimately it became very difficult to get a customer and keep them in that space.
At some point we thought we could keep trying to do this reservation system thing, and then have the risk of the attack from below. Other guys were not sitting still. They were making their search engines better. They would under price us. And, of course, the problem that you have if you’re a new entrant like we were, and you have only part of the offering, is that the other guys can bundle. They can say, “Search is just free, as long as you host on me,” right?
So we started to see very strong signals from our customers that they were getting that kind of pitch. In other words, that our competitors were starting to compete on the basis of lower distribution fees or lower hosing fees. We don’t have those things so we had no way to compete with that. And I don’t think that was necessarily fatal but it’s the kind of calculation you have to make when you look at the state of the business and say, “Do I exit now, do I partner with somebody or do I double down and try to go after these other competitors?” At some point you make the decision one way or the other, and that’s what you do.
Andrew: So you did it. You sold to Google. You rocked our world for a while on the outside. And obviously your life was changed because of it. And then I wondered, why email. And it had something to do with your inbox when you were at ITA.
David: Yeah, I mean I was getting 800 to 1000 emails a day even as early as 2004. I was kind of on the leading edge of this inbox overload problem. I also have a background in computational linguistics. So, I’d done work on text understanding which seemed highly relevant to the inbox overload problem, the idea that obviously a human is not going to be able to deal with a thousand messages a day. It’s just preposterous that a person’s going to be able to do that.
It was pretty obvious even back then that you would need to have some AI or some kind of agent that would work on your behalf to triage a lot of it for you. At the same time, I saw nobody doing anything like that. No one was even thinking about it. The client tools that people had, Outlook and Lotus Notes, Gmail was starting to get traction then, they were still kind of the way they were in the ’90s. They hadn’t changed much, yet the volume was just massively increasing.
If I looked at the addressable market again you’re talking four billion people use email. I mean it’s just absolutely ubiquitous. So, it was very compelling to me. At the same time I realized that it was a really, really hard problem. The thing that I really tried to gauge, and whether I was right or not we’ll see, was can a startup actually do something to fix this or is it too hard.
Andrew: But weren’t you thinking… I’m looking at the first version. The first version was Cloud enabled desktop app for all your mail, and it was for Mac only. I would be worried that Google would jump in there and make it all browser based. You weren’t scared of Google?
David: Well, of course you’re always thinking about… I don’t know if scared is the right word. But, you’re always thinking about the competitors.
One of the interesting things, though, is that mail solutions by the big companies were already ubiquitous and they still are. It wasn’t like we were going to be surprised and oh my God now Google has mail. I mean they’ve always had it, as long as we’ve been in the business.
Of course it’s daunting. But, it’s kind of like when we looked at travel and said Sabre is, like, a multi-billion dollar market cap company with lots of really smart people. We didn’t view that as a hopeless fool’s errand to try to compete with them.
At the same time we also weren’t naive like well this is going to be easy. I think with a space like email the risk is much greater and the reward is much greater in the sense that it’s closer to a core capability for someone like a Google or an Apple. They care a lot about it, probably not as much as they care about things like their core business. But, they care about it. So, clearly it’s something highly hypercompetitive.
The other thing that was clear even when I just looked at the space in 2007 was these guys have worked really hard to devalue the products to essentially zero for consumers. That’s another barrier. I mean if you can’t charge for something, that’s challenging.
But, I also thought that if you could move the needle at all for four billion consumers that would be really, really cool. Like, you would have a lot of opportunity to make an impact that way, and that’s kind of what I wanted…
Andrew: I feel like it’s one of the biggest problems we have in common, all of us who are online. It’s the amount of email that we get in. I tried for about seven months to just stop checking email, and I said that’s it.
David: Good luck.
Andrew: I actually had a responder that said I have a team of people at Mixergy who will respond.
Andrew: I can’t touch this anymore. If you need any help go to mixergy.com/contact. It didn’t work. There were too many things that I needed to break that rule for.
David: Yeah, that’s too hard. Email is just so ubiquitous. It’s the only thing that really works between arbitrary people. Nobody has to be signed up. Everybody has a mail account. That ubiquity was really appealing to me, and also just the level of brokenness and frustration was really compelling to me.
Andrew: Here’s the thing. If we just say email I don’t think people are going to get the significance of it. If we say just lump it together… I’m opening up my phone right now. I think there’s something special about this. Then I’ll talk about something that I think the audience is going to be worried about, too.
What’s special about it is the swipe gestures to get to what you need? What’s special about it is the ability to respond. When I say respond faster it’s not just that the respond button is closer. It’s what I say to respond. It’s shortcuts.
David: One click reply, and things like you could unsubscribe in one click from a mailing list and Inky knows what the mailing list mail is versus mail from your friends. This idea that the mail that comes in is actually categorized for you I think is a huge improvement…
Andrew: Categorized in what way? What kind of categories?
David: Well, for example, social. If it’s a social message you can look at your filtered inbox and not see that stuff. And then when you want to look at the social messages, go into the social smart view. So Inky moves things around for you and categorizes them so that you can choose to see them filtered out, or not. But the point is that Inky [SP] helps you by filtering out what the intent of a message is, and what kind of message it is.
Andrew: Isn’t that something now that Gmail has started doing?
David: They have started doing it, and one thing that we like to say is that we work well with Gmail, and we also happen to work well with everything else. So if you’re one of the three and a half billion people that has mail somewhere other than Gmail, we can give you the same kind of classification ability for your accounts too.
And also, we have a different approach. For example, we separate out your packages, so you can search for #package and see all your packages. Or you can go into the package smart view and it will say for each one, the package tracking. So it goes and finds the package tracking status for you on the shipper’s web site, and puts it right in there.
It’s true that the intent of Gmail’s tabs is similar. I think the devil is in the details at some level, though.
Andrew: It is a very Mac-like experience. Here’s the one issue I have, and I want to talk to you about it. Security. I worry about giving my email to another company. And I know, from talking to the team internally here, that you’ve thought of that.
David: This is a really fascinating issue because I, almost out of a sense of obligation, baked into the whole plan, preservation of privacy. And the interesting thing, from a market or entrepreneurial stand point is that I’m not sure that it actually matters for market share. I’m not sure that the vast majority of people will vote with their feet for something that preserves their privacy.
But what we do is that Inky processes all your mail on your device. So your device is either an iPhone or an iPad or it’s a Windows desktop, or it’s a Mac desktop. But it’s your device using your network card to talk to your mail server using your CPU to process it. We never see the mail. The mail is never touching our system and that’s just baked in to the architecture. And we did that precisely because I don’t want to have the ability for any of my employees ever to be able to read people’s email. I just don’t want it.
Andrew: That’s one of the parts I didn’t know. I didn’t know Inky before.
David: It’s pretty new. We’ve been on desktops for a while, but the mobile version on iOS has only been out for a few weeks. And clearly all the interest around applications is now centered around mobile. We had a desktop version out, but no one was paying much attention. It was kind of good because it let us iterate in plain view with a bunch of loyal users to get that core engine right.
Andrew: I hate that I do, but I actually do still care about desktop. I still need to be able to check email, only email, on desk top. Everything else I just love to do on my phone. But I see the product. It looks beautiful. It looks like you guys have spent a lot of time on it. I didn’t know about it until Steve Young told me about it. He said, “Look, you’ve got to check this out.” And then he made the introduction to you. I had this issue with security, and I like that the email is all processed on my phone. It used to be that that’s the way it was. If you wanted to install anything from Eudora all the way up to mail, it was just processing on your desk.
And then all these other people wanted to process your email in the cloud. I don’t know how other people do it. They give access to a third party, and then another and then another. I still worry about that.
David: Well, I think that you are legitimately worried about it. And it’s not just the company, it’s law enforcement, any entity that has the ability to get access to the data. If the data is stored somewhere in the cloud they can get it. I think in the post Snowden era there’s a whole complex set of issues.
Andrew: Yes. And you’re right that for some reason people don’t care about it. I don’t know why they care about it when it comes to a search engine but not when it comes to email. They care about it when it comes to a search engine but go look in the app store for email that promises security. Or go look in the app store for a notepad that promises encryption, true encryption. It doesn’t exist.
David: Yeah, most of things don’t get any traction because no one cares. It’s not clear to me whether no one cares or there just isn’t a viable alternative yet. And the other thing I’ve observed about security, it’s fascinating. There’s all kinds of egghead details about how we do the security. We use this technique called zero knowledge proofs and my academic advisor at MIT literally invented this stuff and so I really wanted to use it in this context. And it allows us to store data on your behalf, encrypted in such a way that we can’t decrypt it. And it’s really mind-bending and people have a lot of . . .
Andrew: Do you do that? Does data come off my phone and into your system, and you encrypt?
David: We do for your credentials for your accounts because we want it to be the case that you set up Inky once and then anytime you install it on a new device whether it’s a desktop, or a mobile, or whatever, it knows about all your accounts. And so we want to store your account credentials, but we want to encrypt store them encrypted so that we can’t read them. And so it’s kind of counter intuitive that we can do that, but we actually can do that. But one of the things that I’ve found that’s fascinating about this security stuff is the people who care about security ironically they’re never happy with any of the answers you give them either. So the service is to try to cater to the security conscious.
Andrew: I understand.
David: They don’t win the security conscious people over either because there’s always this kind of sense that well that’s not good enough.
Andrew: You know what? That’s a good point. You tell them that this app doesn’t take the data off your computer and send it into the cloud at all. They say, “Well, what if another app is on their computer?” As soon as you’re on the computer some other app could take the data and you prove to them that that’s not possible and they say, “What if someone’s looking over your shoulder?” You’re right. I get that.
David: Right and we end up with this kind of, well, unless it’s open sourced I don’t trust anything and then I don’t know what to say to that.
David: I’m not going to make it open sourced.
Andrew: And I wouldn’t blame you if you took away that security feature. I know that I care about it. I’d love other people to care about it and I accept that they just don’t. The only [??] . . .
David: We’re not going to take it away because I care about it personally and I think it’s a public good. I think privacy is a public good. What I’m saying is that may just be from a purely Machiavellin business standpoint. That may just be stupid. We may just be doing that and it may actually be hurting us where I’m just going to do it anyway because I think it’s important, but the irony is it doesn’t win me any points with the security conscious anyway. So I’m not sure.
Andrew: You know what thought, Gabriel from Duck Duck Go . . .
David: [??] feel better, nobody else.
Andrew: I got emails from Gabriel from Duck Duck Go going back years and years back when some of this stuff just felt like a quicksodic [SP] adventure to try to fight Google on security issues. And today, he’s making a dent. He’s getting a piece of the market, not a huge one yet, but it shows that maybe raising awareness of the issue and fighting for it could help.
David: Yeah and I think also people have to develop trust in your company and your brand.
Andrew: And invest.
David: And it’s obvious that they don’t know who you are in the begging, they’re not necessarily going to trust you. So there’s a hurdle you have to get over where you can say things to people and they just intrinsically believe them at some level as opposed to you’re just some company they never heard of. I think they intrinsically disbelieve you.
Andrew: Yes, yes.
David: What’s the catch? One of the things we get all the time which is just frustrating and funny. People say, “I’m not going to trust your app because it’s free.” Right? And so then there’s a catch. But if we ask you to pay for it, you wouldn’t pay for it.
Andrew: Right and you wouldn’t.
Andrew: How are you going to make money if the app is free?
David: The goal for us is to take this same core engine and do things for small businesses and com- . . . And also even consumers I think. Consumers that are high power users of email, I think, our hope is eventually people will understand the benefit and will actually pay. But if you’re the new entrant and you’re trying to attack from below and the price is zero, there’s no below. And so you can’t go out of the gate with that. So ultimately I think we’ll find ways to monetize it, charging people for it, but you can’t lead with that. With consumers you’ve got to get consumer traction first and then have people spread the word about it.
And we’ll also do things like, you see it in the app stores with games. All the games are free and there’s in app purchases. So we’ll do stuff like that too. We’ll have things that you can buy in the app that tailor to certain audiences.
Andrew: I can see that. All right, and even though security is in there I don’t think anyone else cares about it, and that’s not the reason to go in and check out the app. It’s all the functionality that makes the nightmare that is email, that tames it, that makes it more bearable, that makes it easier to deal with, and it’s all available at . . . I love this domain. It’s just Inky, I-N-K-Y.com.
David: Yeah, and the Inky name was a long time coming. We worked forever it seems on trying to come up with a name that was memorable, easy to pronounce, easy to spell, we could get the .com, we could trademark the name for, you know, in our space and I also kind of like the idea of writing or ink or there is something involving communication, the communication spanning the millennia. So I’m really proud of the name. I think people initially find it odd, like why is there a weird octopus, what does it have to do with mail.
Andrew: I thought that was because of ink in the content.
David: It is, it is and I think ultimately people will get that. If you look at something like mint.com, it’s similar. First people are like, “What does mint have to do with anything?” And it was like, “Oh, mint, like a mint, money.” And then the logo is a little sprig of mint. You get it over time, but it has to kind of bake for a while in people’s minds and I think Inky has that character. And I’m really proud of the name and the brand and [??].
Andrew: Well, it must cost you a lot to get Inky.com. That’s a great domain.
David: It didn’t actually, it didn’t.
Andrew: It didn’t?
David: It was one of those weird things where we lucked out and it was like 2500 bucks, believe it or not.
Andrew: Oh, wow! Yeah, actually I see.
Andrew: It was just parking there. Someone parked it at GoDaddy for a while.,
David: Yeah, it was the guy who had a print shop called Inky Do and he had it on Sido, one of the domain brokers and I emailed him and said, “We’re willing to buy it for your ask.” And he said, “Well, shoo, if you pay me today.”
David: It’s yours. And so I like literally paid part of the money that day.
Andrew: It’s totally worth it. That’s a great domain name.
David: Yeah, and people have asked us to buy it for 100 grand. So just the domain alone, I think, we really lucked out on that.
Andrew: I see him. He’s a guy in Illinois.
Andrew: Well, thank you so much for coming on here. Congratulations on all your success with ITA and on Inky, and all the success with Inky, and for not spying on our email.
David: Hey, thanks, my pleasure to be here and I will never allow any spying on you. And that is my [??].
Andrew: I appreciate it. All right, well thanks so much. Thank you all for being a part of it. Bye guys.