On Starting and Scaling Dropbox (YC W07)

by Drew Houston

Drew is the founder and CEO of Dropbox.

Rejected once before, Drew eventually went through Y Combinator in the Winter of 2007 with Dropbox and has since grown it to over 500 million users today.


Sam: Hi. I'm Sam Altman. This is "How to Build the Future." Our guest today is Drew Houston. Drew, thank you for taking the time.

Drew: Thank you for having me.

Sam: So you were in Y Combinator with Dropbox in the summer of 2007.

Drew: That's right.

Sam: How did you come up with the idea and end up at YC in the first place?

Drew: Well, I had the idea...it was borne out of frustration. I kept forgetting my thumb drive. The sort of origin story of the company was when I was riding the Chinatown bus from Boston to New York and forgot my thumb drive and couldn't get anything done on the trip. And then I had another problem, which is I was working on my first company, which we can talk about, and I had a real pain point around collaborating with my co-founder around the things we were working on. But it was personal frustration. I had hacked together all these scripts to try to fix it, but then that bus ride was the last straw.

Sam: You had mentioned in our first conversation you had applied to YC before.

Drew: We were rejected.

Sam: You were rejected?

Drew: For the first...

Sam: That was either a big mistake or a really good call.

Drew: It was a good call. My first company was called Accolade. It was doing online SAT prep, because we realized back in the day the SAT was changing, and no one had really built a good online course. So, started that company, applied to YC, rejected largely because my co-founder and I were both part time on it. A couple other reasons, market size, probably some other stuff.

Sam: Did you interview, or were you rejected without an interview?

Drew: We were rejected without an interview.

Sam: Wow. I don't think I knew that. And you came back the next batch?

Drew: I got a little more feedback about what happened, and some of the concerns were the ones I said about how my co-founder and I didn't know each other super well, neither of us was full-time on it, some market issues.

Sam: What happened with that co-founder?

Drew: We kept working on the company until I started Dropbox, and then we sort of drifted apart.

Sam: You applied to YC as a single founder and then met Arash right after, right?

Drew: Yes.

Sam: Very quickly after. Can you talk about that story?

Drew: It's interesting, usually these co-founder stories start with...

Sam: "My childhood best friend..."

Drew: "We played t-ball together..." No, Arash and I probably knew each other for maybe two or three hours before...

Sam: Wow!

Drew: ...we threw in together. Sam: What gave you the confidence to be okay with doing that?

Drew: Well, it's funny. I applied as a single founder to YC, and I got an email from Paul saying, "Hey, idea seems interesting, but you really need to find a co-founder." Only problem was, application deadline was like three weeks away, so that's like saying, "Oh, you need to get married in the next three weeks." What happened was, I... Kyle Vogt, who's one of the founders of Justin.tv, which turned into Twitch, and then started Cruise.

Sam: So far the only YC founder with 2 billion [inaudible 00:02:54].

Drew: Yeah, so Kyle and I knew each other from MIT, and we were both in the Entrepreneurs Club together, and he knew Arash. They were floormates in East Campus at MIT, they were both from Kansas. Arash was having a pretty similar experience to one that I was having, which is we had friends who had dropped out of school, moved from Boston, settled in San Francisco. Suddenly their companies are raising millions of dollars, taking off, so I think both of us had a little bit of cabin fever.

Sam: It's frustrating to watch that happen.

Drew: Yeah. It's motivating.

Sam: It is. So you literally, two or three hours you had spent together.

Drew: Yeah. We got introduced. Kyle sent, or just told...I was complaining to Kyle, "Is there anyone great that you know who could be a natural co-founder?" Because all my friends, or the folks I was closer to, just for whatever reason, the timing wasn't right, weren't ready to make the jump, whatever it was. And then Kyle suggested Arash. I probably have the email somewhere where Arash said, "Hey, let's meet up." We hung out in the Student Center for a couple hours, and he dropped out of school the next day. It was pretty wild.

Sam: I have to add this to Kyle's list of special services at YC. I didn't even know this. So you guys just started working on it immediately, talked about the equity, and went forward?

Drew: Pretty much. We met up in the Student Center at MIT. He was already pretty excited about the concept of starting a company, of doing YC, and we hit it off. I wish I could say that, "Oh yeah, here are the 17 things I was looking for on this 7-point scale, for..." But no, I was like, "This kid's clearly smart. A little quiet, but very thoughtful," and over time it became clear that this was a really good, good fit.

Sam: How did you guys validate the need for Dropbox? I've heard stories, although it may be apocryphal, that you posted on Reddit, ask people if they wanted something like Dropbox.

Drew: The first real test was, I made a demo video. It was initially just to get into YC, because the whole Y Combinator application process was pretty similar to the world where I came from, which was college admissions and test prep. One of the things you can do to give yourself an edge in admissions is just do something different. I knew everybody was going to have a demo link in their application or whatever that would be, but I thought that Dropbox was something that would really appeal to the Hacker News audience. It was really to get into YC, where I made this three-minute demo video with some music equipment that I had from my band, and I put it on Hacker News. It stayed at the top of Hacker News for two days. I got the thing I was looking for, which was the email from Paul. The thing I didn't want was the, "You also need a co-founder in three weeks," but that was the first...I think we did that at least twice. First to get into YC, then to build some momentum around the company.

Sam: There's this infamous comment, either on that post or a future one you made, probably that one, where someone's like, "I think that's the stupidest thing I've ever heard of. Don't these people know that Wget exists? You can just use that. No one needs a company for this." We send that around to other founders when they get demotivated, because people shit on their idea on Hacker News or Reddit, or something else.

Drew: Yeah, we were getting it from all different directions.

Sam: We can jump to that, because I think that's a really interesting point. One of the things I respect most about you is you have just kept doing your thing no matter what noise there is in the ecosystem about "Dropbox this" or "Dropbox that," whether it's like, "I don't need this because I can use Wget," or, "This company's overvalued," or whatever. And now you're just this behemoth, and everyone's like, "Well, we were wrong about Dropbox." How do you handle just constant negativity from people saying, "This isn't going to work," or anything like that? Drew: I think the first thing is you have to recognize that that is always going to happen. We experienced that from the day we founded the company, actually before that, and all the way to now. When you zoom out, you realize that a company like Apple, quarter after quarter, can post the best, the most revenue, most profit of any company in business history...

Sam: "It's over, it's dead. Company sucks."

Drew: ...and the next day, everybody's like, "But they're really dead now." So you have to appreciate that that's human nature and that's just going to happen. I think where it gets a little harder is, you have to simultaneously have thick skin and be able to tune things like that out, but you also have to have kind of thin skin. If your customers aren't happy or your team's unhappy, you have to respond to that. So that's a weird dynamic that you have to handle. But I think it starts with just getting some perspective that that's always going to happen.

Sam: How do you decide when you need to address something with the customers or team versus just let it go on its normal course?

Drew: I'll give you more in the fundraising context. In addition to people on Hacker News, a bunch of the comments were negative, but actually a bunch were positive, too.

Sam: [crosstalk 00:08:08] positive.

Drew: We got a lot of pushback from investors saying, "Hey, look, Google's going to do this, it's a commodity." Online storage was kind of the startup cliché of its time the way that Bitcoin or other things have been, and frankly I agreed with them. I was like, "Yeah, Google can probably do this. Yeah, we don't really have a lot of structural advantages. Yeah, we're not really sure how we're going to make money. We don't even have users." That's all true, but you have to realize that's not causal. Just because other folks have done something, or done it in a way that you don't think is good, doesn't mean it's inherently a bad idea. Let me put another way, I would ask investors, "Do you use this? Yes, there are 50 of these things out there, but do you use any of them?" They would say no, and I'd say, "Isn't that interesting?"

Sam: Was this the biggest problem that you found in early days was just the level of competition or presumed competition in the market?

Drew: Competition didn't bother me that much because our sites were also a lot lower in the beginning. Really, I had a target customer of one, which was really myself, and victory, for me, would be never having to carry on a thumb drive. So that was the starting point.

Sam: How long did it take you from the day you had the idea for Dropbox until you had a product where you...until the last time you used a thumb drive?

Drew: Probably about...I'd say a couple months. Maybe six weeks, six to eight weeks. When it's early, you can make massive progress like that, but then it took a long time to make it bulletproof and sort of safe for the world to use.

Sam: What were the big initial problems that you faced? Or what were the, as you think back towards the few times Dropbox almost died? What actually went really wrong?

Drew: I think the early problems, when they were real problems, were just keeping up with growth. We had this idea to build a prototype, the prototype worked well. We made this video, and we put it on Digg and Reddit, and that just blew up, which I can talk about. But then you need to build a company, because you start a product, then you're like, "Hey, we have a product, now we need users. Okay, let's go figure out how to get users. Okay." So you're...

Sam: They always leave this part out of the movie of startups, you know.

Drew: Yeah. Since you're on this treadmill, every day you have to do everything you were doing yesterday just as well or better, and then you have a bunch of new stuff to deal with. So it's like, "Okay, well we need to build the product, we need engineers. Okay, wait, before we can get engineers, we need to pay them. I guess that means we need to fundraise. Okay, we need to fundraise, and then we need a lawyer and an accountant." So just the circle keeps getting bigger. You deal with that by dividing and conquering. You hire people, but then you have a new problem, which is you're hiring people to do something you don't know anything about. Suddenly everybody's running off in different directions. And so I think the hard things were really keeping up with that. The rate at which things were breaking was faster than we could fix things. That was a really hard thing to get used to, because you're used to, especially if you were used to getting good grades in school and things like that... Ben Horowitz said this well, he said, "The average grade for running a company is a 22 out of 100." When you're used to getting As, that's a humbling thing.

You have to simultaneously keep two things in mind. Which is on one hand, you really should aim for perfection and try to make things as good as they can be, and when problems arise, not tolerate them. But at the same time, you've got to recognize things...it's like trying to bake a cake in five minutes. There's just going to be stuff on the walls. It's going to be a little screwed up, no matter how good you are. And what you really need to do... And probably even harder is, your job as CEO changes every six months, every year, every couple years. Nobody tells you that. In fact, some of the things that made you good at the prior stage become a hindrance for the next stage. For example, I wrote most of the first version of Dropbox, and certainly Arash and I did together, but there just became too much stuff to do, and I had to stop coding.

Sam: Any of your codes still running in production?

Drew: Yes.

Sam: Very impressive.

Drew: I think they've done some analysis. There's still stuff in there, and a lot of how the core product works and the protocol has been pretty stable, which has been cool.

Sam: Although there's so much advice online about how to start a product, there's actually very little advice about how to transition to a company. Anything else that either is particularly good advice that you heard or advice you now give to entrepreneurs who are trying to build the next Dropbox about how to successfully transition from an engineer to someone running a large company?

Drew: Yeah. The first thing that would've been helpful to know is first of all, nobody's born a CEO. You learn it. The challenge is, you just don't know what your blind spots are. You know how to build...at least in my case, I grew up programming, I knew I could make the code work. We knew we could get people on Hacker News excited about it. We knew we could make our early users happy. But if you want to bring a company to sale, the chessboard is a lot bigger than just building a good product. You have to be able to make smart and fast decisions about what markets do we go into, our business models, competition, strategy. How do you build a team, how do you run a team? That's pretty tough when you've never done any of it, and you don't really have a lot of context on those things, don't even know those are important.

Sam: How did you learn those things?

Drew: I think the first thing you need is, on the one hand, like a healthy paranoia around... Probably one of the most useful questions that I like to ask myself, and still do, is, "One year from now, two years from now, five years from now, when I look back on today, what will I wish I had been learning, or what would I wish I had been doing?" And so I think really taking it on yourself to be systematic about how do I train myself, what am I going to need to know? They say you go from coder to psychiatrist, which I've found is totally true, or politician, in the job, and then there are a lot of different points in between.

Sam: Did you mostly figure this out yourself, or did you ask people, like, "What am I going to need to be doing? What will I wish I had done in a year?" Or did you just...

Drew: A little of everything. Reading was probably the single most valuable thing that I did.

Sam: Anything you'd recommend other people read?

Drew: Yeah. I think my favorite book on management is "High Output Management" by Andy Grove, which is kind of the Bible of how do you scale an organization. I have a very long list of recommendations. "The Effective Executive" by Peter Drucker is really good. But basically what you're talking about is, you need to have good judgment about a lot of different things. What that really is saying is, how do you gain wisdom really quickly? Last book recommendation, "Poor Charlie's Almanack" by Charlie Munger. This is probably the best book on that that I've read, which is really about how do you build a bunch of mental models to break down the complexity of the world and know when to use each. Reading is part of it, but then also having...you sort of think about that one year, two year, five year...having friends that are one step ahead, two steps ahead, five years ahead, and just sort of looking at the deltas between them. I remember I had a coffee meeting set up at Sequoia. Our first venture investors were really helpful in connecting me with other entrepreneurs. I remember racing out of the office, running down to Starbucks, having a seat, and start peppering him with questions. I asked, "How do you spend your time? What's important to you?" He's like, "Well, it's really important that we have a great mission as a company." I'm like, "Okay." And he's like, "And it's really important that we have great values and a great culture."

And I'm starting to tune out, because I'm remembering which bugs I didn't fix. I'm like, "Did I finish that before I committed the code?" And I'm like, "What the hell is this guy talking about? None of this has anything to do with our customers, getting code out the door." But then you start hearing that 5 or 6 or 10 times, and you realize, "Okay, the only one here who's confused is me." Then you realize, especially if you're an engineer, you're going from designing a system of code and algorithms, which is very orderly and discrete, to a system of people, and things like culture, values, all kinds of other things, are the programming that makes that work.

Sam: I think everyone's confused when they hear that for the first time, and they tune out and they think about code. Have you found any way to get entrepreneurs to internalize that earlier? I never have, so I'm curious if you've figured out a way to...it may just be something you have to live.

Drew: I think you really need that portfolio of different ways that you can train yourself. Reading's one, mentorship or getting advice from people at different stages is another. I like to study the history of business and, really, excellence. And so learning about, okay, well it's one thing to know how Google was built or how Facebook was built, but then you can learn a lot by understanding, how does a consumer packaged goods company run? Like how does Swiffer...how do you come up with a Swiffer? How do you think of a competition there?

Sam: I've certainly learned the most from studying businesses that are not in the technology industry, and I don't think people ever get enough of that.

Drew: You want to go beyond business, too. So how would Phil Jackson, with the Bulls, the Lakers, the Knicks, so many championships, how do you build a culture on a team like that? How do governments solve big problems? Then you can also lean on your investors and the Board. For example, Condoleezza Rice is one of our board members. She's seen scale on all kinds of different environments, from running the State Department, being Secretary of State, it's an 85,000 person organization. She's provost of Stanford, that's a 20,000 person organization, and so on. You want to triangulate from a bunch of different sources, which helps you internalize these lessons more than if you just read like a blog post about it.

Sam: The most interesting meeting I had in terms of thinking about how to build YC in the last year was going to the Vatican and listening to how they talk about, like...they think on thousand-year time frames, and about a billion people, and it's a very different perspective, but it was really good.

If you were going to start a new company today, what do you think the biggest opportunities are?

Drew: I think that the whole AI, machine learning, machine intelligence is going to be transformative, especially when you zoom out. I think things like, when paired with things like robotics, machines are going to take on a new role in our lives, which will have a bunch of really interesting positive and negative effects, which we can get into. I'm not as worried about the robots killing everyone, taking over the world, but we can talk about that later.

Sam: There's a lot of good and bad short of [inaudible 00:19:56] could happen, I think.

Drew: But I think it's a really exciting time. In our world, we talk about Dropbox of all things, from keeping files in sync to keeping teams in sync, which is a little bit of a departure. Because when we started, we were like, "All right, our goal is to make it so nobody has to carry a thumb drive around again. Life would be a lot better if your stuff was in the cloud." Today, that's largely solved.

Sam: Mission accomplished.

Drew: Yeah, mission accomplished. But along the way it opened our eyes to a bigger problem when we realized that, for a lot of our happiest customers, it wasn't really the storage that people were buying, it was the sharing. They were really using Dropbox to get their team on the same page. What that meant a shift, which we can talk about, but a couple years ago, we were like, "Okay, of all the things..." We're sort of blessed and cursed with this original product, because we struggled in the beginning. People were like, "Well, what's Dropbox?" Or, "What's Dropbox for?" The answer, when you think about it, "What's a computer for? What's a file for?" So we ended up really supporting all these different use cases, but at the same time, it was very clear even in the beginning that the role of files in people's lives are going to change. In the beginning, yeah, maybe you'd have text files and mp3s in your Dropbox, but in the world of iTunes and Evernote and Spotify and Netflix, there's not a file to be seen. Tech just moves quickly, so you always have to keep evolving with the patterns of what's changing.

Sam: Was this a strategic shift that you consciously made? Did you say to the whole company some day at an all- hands, "We used to be a file company, and now files are dead and we're going to do this other thing"?

Drew: We did expand from...I would say there was an expansion and then a contraction, which I can talk about. We knew that the role of files was going to change. Files were never going to go away, but they were going to change, and we thought about, "Okay, what place do we really occupy in people's lives?" There are a couple of big use cases. One was private photo sharing, another was collaborating with a team on files, then the third was, we had a platform...or we do, we have an ecosystem of all these developers. So we expanded to say, "Okay, we're going to build a separate thing for photos, because that's kind of a different product. We're going to do a lot more on collaboration and so on." That was when we started to buy. We bought Mailbox, we started building something called Carousel, which is our photo app. Then in 2014, we stopped doing those things.

Sam: How'd you decide to stop doing them?

Drew: It was tough. I think what we found was, first of all, I think it's easy to make strategies that look great on paper. When you think about your overall market, you want to hold onto as many use cases as possible. So many of our users were so happy with things like sharing photos, and we could do such a better job there. But at the same time, you look at photos and collaboration, it's very different products, very different ideal business model, very different competition, just pretty everything is different.

And so we found ourselves sort of straddling these two worlds, which made it very hard A) to build the best focus product, and B) just made it hard for people to understand what we were doing. And as much as like, "Oh yeah, we were out in front of all this," competition sets in. I think there's just sort of a sense of, "We're doing a lot of stuff, it's not really working, and people are just getting agitated." Then I re-read a book called "Only the Paranoid Survive" by Andy Grove, which describes how actually Intel wasn't always a microprocessor company. They actually started in memory, and they were getting crushed by the Japanese. He talks about the story of that transition from memory to microprocessors.

And as I was reading it, my stomach started churning, because I'm like, "Oh my god, this strategic inflection point he's talking about, that's what we're in." There's a part of the book where it says, "CEOs like to keep their options open." What you really need to do when you're in the middle of one of these inflection points is, "Put all your eggs in one basket, and watch that basket," quoting Mark Twain.

Sam: That is so true.

Drew: I went into the office the next week, and we were like, "We're not doing Carousel anymore, we're not doing Mailbox anymore."

Sam: Was there resistance, or did the team get behind you quickly on that?

Drew: A little of both, I think. You find that, when your company gets bigger, sort of everybody knows, it's not a secret that some of these things aren't working the way that you want them to. So in the sense, to one set of people it's a relief, and frankly they're like, "What took you so long? This was obvious. I told you so." It's very disappointing to the people that are working on it. It's disappointing to me. Here are two products that I really loved, and it's embarrassing, it's publicly embarrassing. Now something that changed is that a lot of problems in your company, nobody really cares about them except for the people in the company, but when you go on a big upswing like the one we had, suddenly everything you do has a lot more, the peanut gallery's a lot bigger. And the press, and just the folks can be a lot...there's a whole machine that will build you up, and then...

Sam: They love to tear you down.

Drew: ...they love to tear you down. Again, even if I understand that and can put that in perspective, it's very hard for the team. Everybody loves the idea of focus, but what focus actually means is shooting stuff you love and turning down things you know you can make work. That's part of the cyclical nature of it. There'll be times when you expand and times when you contract. Sam: Speaking of the expansion, one other thing I remember...I didn't know you really well in the early days of Dropbox, but I remember being impressed by just how much you got done. Sort of like, Dropbox was really taking off and just seemed to ramp up to where everyone was using it in a one-year period. How'd you manage your time? What did your days look like when you went from 7 users to 7 million?

Drew: Sure. The team...I'm thinking of "7 to 7 million." We were probably fewer than, under 50 people. So I think one of the...probably a few things that have been long misunderstood or not clearly understood about our space, which is first of all, Dropbox, part of the reason why we had so little competition for a while was, Dropbox looks easy, but it's actually really hard to do. And I knew this, because I wanted to just use the other products. Like if any of them had worked, I wouldn't be here. But it's actually really technically hard. That said, there were a lot of competitors who built things that probably were reasonably functional, but they didn't get distribution right, and they didn't get virility right. I had been paying attention when...2007 was when the Facebook platform started to take off, and a lot of the gaming, and you had these companies which were going from 0 to 100 in records. New land speed records were set all the time.

And a lot of that was through virility. Dave McClure had a great model called "Startup Metrics for Pirates: AARRR." I won't spend too long on it, but basically, "Acquisition, Activation, Revenue, Retention, and Referral." That becomes, this was how we managed the company. Users come in, "Oh my god, very few of the folks who signed up, four out of five who sign up don't put a file in their Dropbox, don't install the client. What is going on?" We actually brought people in off of Craigslist and watched them install the product. Then we showed the video in real-time to our whole team in the other room. We had the person and a PM in one room, the rest of the team in the other room. We're like, "Hey, just sit down. Here's Dropbox. Go from here to sharing a file." Zero of the five people succeeded. Zero of the five even came close.

But really paying attention to all the activations is one example, but all the steps in the viral engine, tuning that as much as possible. Coming up with things like the referral program, which had this two-sided incentive where, if I tell you about Dropbox, you get free space, I get free space. That drove 30% of our signups for a while. Share folder's inherently viral, another 20%. So half of our signups were viral. That was something that wasn't well understood before that time, and it was inspired by some of the stuff I saw in the late '90s, like PayPal...or maybe early 2000s. PayPal had an incentive referral bonus. I had the idea for the Digg and Reddit video based on a book called "Guerilla Marketing," which was like, how do you do marketing and get users when you have no money? So a lot of these things came together at that time, 2007, 2008, or a lot of the dots were connected from things that I had read over the last several years.

Sam: I guess if you were less than 50 people, you didn't spend too much of your time managing and thinking about the organization, then.

Drew: When you're like 20 people, you can just sort of manage by intuition. It's manageable because everybody fits in a room, everybody knows what everybody else is doing. But that transition to, now when you say 30, 40, 50 people, is very difficult.

Sam: Okay, last question. I won't ask you when you're going public, but how do you think about...what's the decision framework for staying private or going public, and how do you think about that?

Drew: There's nothing that magical about it. It's primarily a financing decision. It really depends on do you need capital, your employers and investors need liquidity, there's some other benefits around branding or having currency for acquisitions, so on. What's happened over the last 5-10 years is the private markets have found ways to solve most of those problems. Going public is a way to solve all of them. However, then there is, operationally, it's tough for a couple reasons. One is there's a lot of overhead with compliance and regulatory stuff that really...where you have a lot of people in your company spending time on things that don't really have anything to do with your customers or your products, so it really is overhead. There are other elements of discipline and so on that are good, but it is a lot of overhead. The stock market tends to be manic-depressive, so they're either like way more enthusiastic than they should be about your company or way less, which creates these very...has a lot of impact on morale. But I think at scale, most tech companies are public, and that's the path we're charting.

Sam: Great. Well, thank you very much for taking the time.

Drew: Thanks for having me.