A bewildering number of services let computer and smart-phone users store and share files in the Internet’s cloud. But one file-hosting service in particular has evoked the kind of devotion ordinarily accorded social-networking services or beloved hardware manufacturers: Dropbox, the product of a startup founded in 2007 by MIT computer science students Drew Houston and Arash Ferdowsi. The service lets people use almost any computing device to store files in folders in the cloud as thoughtlessly as they store files in folders in their device’s memory. Achieving that simplicity of use—something Houston calls “an illusion”—is very difficult, because it forces the company to wrestle with all the variants of the major operating systems, four Internet browsers, and any number of network file systems. No other service supports so many different systems. More than 50 million people around the world have been beguiled by Dropbox, which is free to many users. The company’s robust growth, together with revenue from the fraction who pay for extra storage and options, has been rewarded by a valuation that various reports place as high as $4 billion. Technology Review’s editor in chief, Jason Pontin, spoke to Houston, the chief executive.
TR: Why did you want to start a company in a field—Internet file hosting—where there were so many competitors? I count as many as 15, including Apple’s new iCloud service.
Houston: For me, it goes all the way back to MIT, where there is a campus network called Athena. You can sit down at any of thousands of workstations and your whole environment follows you around: not just your files but where your icons were on your desktop. Then I left and discovered that no one had really built that for the rest of the world.
Why was Athena so attractive?
You never had to sprint across campus to pick up some paper. You didn’t have to worry about backing anything up, because the whole point was the network was ubiquitous and it was taken care of for you.
It seemed clear to me that in the future [in the wider world] someone was going to take care of this stuff. There were lots of technologies that purported to do so, but when you used them you found they only bit off a little bit of the problem. You’d install one application to back your stuff up; you’d install another app to sync things between computers—and you’d sign up for a Web service to upload your files. I just thought: computers can do a better job than people of remembering all these details.
Tell me the requisite founder’s tale.
The breaking point for me was a bus ride. I went down to Boston’s South Station to ride the Chinatown bus to New York. I was thrilled to open my laptop and have four hours where I could finally get some work done. But I had that sinking feeling that something was wrong, and I started feeling in my back pocket for my thumb drive, and of course I could just see it sitting on my desk at home. So I sulked for about 10 or 15 minutes and then opened up the [text] editor and wrote some code that I thought would solve the problem. And I met up with Arash through a mutual friend at MIT, and he decided to drop out with a semester left, and we went to California and got to work.
How does the apparent simplicity of Dropbox’s user experience emerge from the complexity you must manage?
We want you to have your stuff with you wherever you are, and that requires that we remove anything that gets in the way. There are technical hurdles that we’ve had to overcome to provide the illusion that everything is in one place, that it just sits there, and that getting it is reliable, fast, and secure.
Achieving that experience is not simple: we have a polished exterior, but there’s this jungle of different operating systems [with which we must work] and even gnarlier stuff like operating-system bugs and incompatibilities. It’s a hostile environment: we macheted our way through that jungle of problems. It was a bunch of us spending big chunks of our 20s chasing down these obscure compatibility issues.
Can you give me an example?
On the Mac, when you look at the Finder, your Dropbox folder has this little green check on it that indicates that your files are in sync. That little piece of visual feedback was really important to us. And to do that without access to Apple’s source code, we had to reverse-engineer how the Finder works and find the little piece of code that draws the icons and perform open-heart surgery upon it. And then you had to do that on [all the different versions of MacOS, such as] Tiger and Leopard and Snow Leopard and Lion, and also on the Power PC and on Intel and in 32 bits and 64 bits …
Excellence is the sum of 100 or 1,000 of these little details. We care deeply about making something that’s excellent from an engineering standpoint even though other companies might decide that, say, Linux support is not an economical thing to devote resources to.
Your company has worried over perfecting version control. One option in Dropbox, called “Packrat,” even allows users to save every version they’ve ever made of a file. Why do you have this emphasis?
Since computers have existed, every user has had this feeling that they’re one click or keystroke away from disaster. We were trying to imagine, “What if you were to build a universal Undo button?” Building a universal undo turned out to mean that we had to invent our own file system, but the way we designed it made it pretty easy to record past versions of files and keep them around.
What has been the biggest challenge of scaling to reach so many users so quickly?
It’s easy to make a solution that works 80 or 90 percent of the time, or even 99 percent of the time. But sooner or later, if the day comes where you’re about to present before an audience and the PowerPoint is not there, you’ll stop using the service—and you’ll tell all your friends what a terrible experience you’ve had.
You talk very winningly about reliability and trustworthiness. But what happened last year when all accounts on Dropbox could be accessed, however briefly, without passwords?
In short, there was a code update that was bad. It wasn’t caught by the mechanisms that are meant to catch such things. You can imagine that was pretty much the nightmare scenario for us. In response, obviously, we did all kinds of work to make sure that kind of thing never happened again.
And you wrote to each of the 54 users who were affected, gave them your cell phone number, and personally apologized.
The number was somewhere in that neighborhood. But yeah, I did.
Your business model is what’s called “freemium.” When I sign up, I get two gigabytes of data storage free. For more storage, and for some options, I must pay [$10 a month for 50 gigs or $20 for 100, although users can be given up to eight additional gigabytes for referring new customers to Dropbox]. Do you really believe that enough people will find two gigs constraining? I’ve read that 96 percent of your users pay nothing at all.
In the literal sense it’s just more space, but from an experiential standpoint the real value is having all your stuff in your dropbox instead of, say, only your documents. You can have your whole life in there—with you, wherever you are.
Will Dropbox one day become more than a network for file sharing?
Absolutely. The explosion of mobile devices means that the world needs an elegant solution for the new problems people have. It needs a fabric that ties together all of their devices, services, and apps. Even though today people may think of Dropbox as a magic folder on their desktop, what we’re really excited about is the opportunity to make all this other stuff you use better. We envision little Dropbox icons everywhere, analogous to the Facebook icons you see everywhere. When you take a picture, it should save your photo to your dropbox; and when you make a to-do list on your iPhone, it should save the list to your dropbox. Any app or device should be able to plug into Dropbox and have access to all your stuff, because that’s where it resides.
This future Dropbox is an example of an overused word: a “platform” with which many software developers and hardware manufacturers will work.
Over the next years you’re going to see that the value that comes out of Dropbox is more and more the stuff that other people build. Whether it’s your TV or your camera or the apps on your phone, we want to make it easy for anything that consumes or creates data to be able to plug in. What we’re really trying to build is the Internet’s file system.
We’re far from that. It’s a mess now, right?
Yes. Think of the idea of the connected home. I just moved into a new apartment, and I have this new audiovisual equipment, and the TV has Wi-Fi and the receiver has an Internet jack on the back—but the downside is that I feel like I have 10 ways to watch Netflix badly.
Everything is jockeying to be at the center of the universe at the expense of the user experience. We think we have a lot to contribute here.
Well, we wish you well. But doesn’t competition from Apple’s iCloud service give you pause? There is a company that also sweats the smallest technical details.
I think they’ve demonstrated that they fundamentally care about making the Apple experience really good, but they don’t pay nearly the same attention to other platforms. Even if you’re an Apple user, what happens when you need to share with someone who has an Android phone or you have to work with someone who has a Windows PC?