Facebook Buys Its Way Into the Heart of the App World

Ilya Sukhar's startup was only a few months old when Facebook came calling.

Sukhar launched his company, Parse, in August of 2011, offering the world's software developers an easier way of building apps for mobile phones and tablets, and just after the New Year, as coders started kicking the tires on this new-age internet service, he heard from Facebook CEO Mark Zuckerberg and the tech giant's chief technology officer at the time, Bret Taylor. Even then, Sukhar says, the Facebook brain-trust was angling to acquire his tiny company.

By January 2012, Apple had sold about 150 million iPhones -- with its App Store on the verge of delivering its 100 millionth mobile app -- and in some ways, Google's Android mobile operating system had surpassed the iPhone, nabbing over half of the smartphone market. But Facebook, a company that aspires to match the influence of Apple and Google in the tech world, was still struggling to hone its own mobile application, much less find a way of expanding across the rest of the smartphone world. Parse, even in its infancy, was a way of gaining a foothold the company very much needed.

Though Sukhar and his startup shrugged off Facebook's advances that January -- "We were moving fast and still quite small and just wanted to keep going," the Parse CEO remembers -- he kept in touch with the tech giant. Over the next year and change, his cloud service became the backbone for over 60,000 mobile apps, and this past spring, the two companies did indeed come together. Facebook paid a reported $85 million for Parse, bringing Sukhar and his cohorts onto the second floor of a building at the heart of the tech giant's headquarters in Menlo Park, California.

In Facebook's world, $85 million is a small sum, and Parse still has less than 30 employees, but the implications of the deal could be far-reaching. Last week, about 600 software developers gathered at a hotel in downtown San Francisco for a one-day conference dedicated to all things Parse, and with Zuckerberg making a surprise appearance to introduce Sukhar's keynote speech, the event marked a notable turn for the social networking giant.

For years, Zuckerberg has described Facebook not as a mere social network, but as a software "platform" -- something that any business or independent softwre developer could use to remake and enhance any software application, from games to music services to news sites. Originally, this involved companies like gaming outfit Zynga building software than ran atop the Facebook website, but nowadays, the Facebook Platform is something very different. The Facebook Platform is pretty much Parse.

"It has always been an evolution. Platform had one focus when the web was dominant, and these days, it's very much mobile-focused," Sukhar explains during an interview inside Facebook HQ. "I wouldn't say Parse is replacing Platform, but I think that Parse is going to become a pretty pivotal, underlying layer to a lot of the other things that Platform is trying to do these days."

Yes, you can still build software on top of Facebook -- Zynga-style. Each month, more than 260 million people play games from Zynga and other outfits on the site. But in many respects, software developers have moved away from Facebook.com -- and away from the web in general -- towards a new world where applications are specifically built to run on mobile phones and tablets. In buying Parse, Facebook is staking a claim to this new world.

It's a move that will not only continue the expansion of Facebook's reach and influence, but provide the tech giant with a new stream of data it could use to target ads across its services -- its main source of revenue -- or even to guide its path into the future, showing it where the mobile game is going.

You Can Cook Without Building an Oven

You can think of Parse as a service that eliminates the heavy lifting that comes with building a mobile app. In coding an app, you have to build not only the small piece of software that sits on phones -- "the front-end" -- but also a much larger "back-end" system that runs on a bunch of servers somewhere, driving that phone software. Parse takes care of the back-end for you -- or at least most of it.

Just as a chef shouldn't have to grow his own onions and build his own over, Parse believes, a mobile coder shouldn't have to build things like databases and login servers.

What this means is that, each month, Parse sends out billions of push notifications to tens of thousands of mobile applications and receives billions of calls to its applications programming interfaces, or APIs. The service gives Facebook a means of tying its social network into all those applications, but it also provides the company with a window into what tens of thousands of people are doing with their mobile phones.

With the arrival of its new mobile app, Facebook Home, earlier this year, Facebook created what you might call a de facto operating system -- software that could push Android into the background on some phones and make Facebook the primary means of running certain other apps. But Parse is a bigger play. It's a play for any and all apps. It can feed Facebook in a way that's not that far removed from the way Android and its app store feed Google -- at least in theory.

"I think this acquisition just shows how smart Mark Zuckerberg is," says Steve Derico, who runs Bixby Apps -- an outfit that builds mobile apps for big companies such as BMW and Lenovo -- and has tinkered with the Parse service since well before the Facebook acquisition. "Some people are still wondering why Facebook bought Parse, but the answer is data. Data, data, data."

But before we give Zuckerberg too much credit for this steal of an acquisition, it's worth remembering that -- because the deal gives Facebook access to this data, because Facebook could potentially use this data to create apps that compete with developers who builds on Parse -- many coders may be wary of tying into the social network. If Parse is to thrive from inside the Zuckerberg machine, it may have to face down some lingering skepticism.

Building Blocks For Everything

Two years ago, after vowing to launch his own startup, Ilya Sukhar began building a mobile app called Posse, a tool that could share your location in realtime with a bunch of other people. "It let you meet up if you were caravanning out to Tahoe," he says.

He soon realized he was completely unsuited to such a thing -- "I'm an engineer," he says. "I wasn't a guy who was great at thinking about the next social phenomenon" -- but the experience helped seed another idea. Building the app, he remembers, was far more difficult than it should have been, even for a seasoned engineer like himself -- someone who had helped build two previous software startups. He spent all sorts of time coding stuff that would sit on a server somewhere, stuff like databases and push notification engines and offline caches.

"There are all these things underneath the hood of these apps that just don't matter to the end user and should be common building blocks -- but, surprisingly, they weren't," he says.

Shortly thereafter, Paul Graham, the head honcho at startup accelerator Y Combinator, introduced Sukhar to another engineer, former Google man Kevin Lacker, and the two met up at Four Barrel, a coffee shop near Sukhar's San Francisco apartment. Lacker drank coffee and Sukhar sipped water -- he doesn't do coffee -- but they agreed the problem needed solving, and they set out to create a cloud service that would offer those common building blocks.

When the pair officially joined Y Combinator that summer, they roped two other engineering types into the project -- James Yu and Tikhon Bernstam, both veterans of the document sharing service Scribd -- and by the end of the summer, the foursome had launched a beta version of the service. This was Parse.

The basic notion was hardly revolutionary. For years, we'd heard a similar pitch from the likes of Google and Microsoft and Heroku, companies that offered cloud services specifically designed to take the pain out of building back-end infrastructure. But Parse was different in that it focused squarely on mobile applications, which turned out to be the future of software development.

"Mobile development will just turn into...development," says Lacker, who brings the company not only a certain technical experience, but a dry sense of humor and a wonderfully detached view of the software world. "In 1994, you would be like: 'I'm working on Internet Software.' But in the 2001, you wouldn't say that. You just say: 'I'm a software engineer."

Parse is still very young, but it has clearly found a place in the world. After just two short years, it's running over 100,000 mobile applications, including apps from names like The Food Network and Sesame Street and Ferrari -- not mention the way it so swiftly caught the attention of one of Silicon Valley's biggest players.

Rebuilding the Coliseum

On the second floor of Building 16 at Facebook HQ, where Sukhar, Lacker, and Yu and the rest Parse team now run the service, there's a poster of the Coliseum that reads: "Rome wasn't built in a day. But they didn't have Parse."

For many who use the service, this is a more than just a good tag line. "We used Parse to show what our app could do," says Alex Fajkowski, a contract developer who stumbled onto the service and used it to build the Food Network's mobile app. "But the way it works is that if you build an app with it, there's little else you have to do. It's good to go. It scales up for you."

Jeff Tanneebaum, a developer and venture capitalist who attended the Parse conference last week, agrees. "I've build many apps on top of it, just because it's so easy to do," he says. "The bottom line is that it's user-friendly. The design of Parse -- the care with which they've built it and its APIs -- makes it so easy to understand."

Sukhar attributes much of this to the talents of James Yu, who once served as a hardcore signal-processing engineer at Intel, but developed a taste for interface design at Scribd and, with his hipsterish black horn rims, served the point man for the interface that fronts Parse.

But there's a flip side to the Parse way of doing things. As Google and Microsoft realized with their cloud services Google App Engine and Windows Azure, many developers want control over back-end development -- or at least some control. "It's not for everyone," Steve Derico says of Parse, and as it turns out, despite his long association with the company, it's not even for him. He doesn't use the service in building mobile apps through his own outfit, Bixby Apps. He builds the back-ends himself.

At last week's conference, in an effort to address this other audience, Parse unveiled tools that let developers run additional code on its servers. The trick will be finding the sweet spot between simplicity and control -- not the mention reassuring the world's developers that Facebook isn't somehow a threat to them.

At least for the moment, Facebook is giving Parse all the freedom it needs to expand its reach. That's in the company's best interests. Though it has moved into Facebook HQ, Parse largely operates on its own, and its service still runs on Amazon's cloud service -- not in Facebook's data centers. "If you look at our product," Sukhar says. "It hasn't changed at all since we got here. There's no extra emphasis on Facebook. We still have Twitter in there."

But in the long term, Facebook sees the service as a way of plugging a world of mobile applications into its social network, through its log-in service, which lets you long into any application with your Facebook username and password, and through other APIs that trade information with Zukerberg's machine. "The Facebook Platform has completely evolved over the past six years to offering a set of services to build and grow mobile apps, and Parse is basically the 'build' pillar of that strategy," says Mike Vernal, the Facebook vice president of engineering who oversees the Facebook Platform.

The arrangement works, Sukhar explains, because both Facebook and Parse are interested in mobile applications that run on all mobile devices, from iPhones to Androids phones and beyond. "Parse is a very cross-platform approach to development," he says. "There are not a lot of companies thinking that way, and Facebook is one of them....Parse couldn't be Parse inside many other companies."

Vernal adds, however, that the Facebook Platform extends beyond Parse. A big part of the new Platform, he says, is the company's efforts to promote mobile applications across it social network and third-party services and generate revenue for developers. "Growing your app is still one of the key values we can offer developers," he says. "We're trying to make it so that developers can build apps can reach everyone in the world for whom that app is relevant."

In any event, as the old Facebook Platform settles into little more than a place to play games, the company has fashioned a very different Platform for the age of mobile computing. Like the old, the new may struggle at times to win the approval of more than just a limited group of developers. But Zuckerberg and company are at least taking the right tack.

"You have to build an app to have any use for [the Facebook Platform]," Sukhar says. "You can't buy ads, you can't integrate log-in, you can't share something with Facebook -- until you have an app."