Building apps for Facebook: A dance with the devil

Dalton Caldwell caused a stir recently when he posted an open letter to Mark Zuckerberg, accusing top Facebook execs of threatening his latest startup with the Luca Brasi treatment: Sell out to us or we'll crush you.

"I had explicit approval from Facebook to build what I was building," says Caldwell, a software developer who unexpectedly found himself in the crosshairs because an app he was building to run atop Facebook was similar to Facebook's recently unveiled App Center. "They said, 'Sorry, we just need the revenue.'"

Caldwell's public dust-up with Facebook -- which led to VC and Facebook board member Marc Andreessen stepping down from the board of Caldwell's company, Mixed Media Labs -- revealed much more than the personal tensions that occasionally flare between big and small software companies. It also opened a window on the extraordinary power wielded by Facebook and the power imbalance that thousands of developers have to keep in mind when they ally themselves with the biggest social network of them all.

On the one hand, Facebook relies on a legion of third-party developers to invent apps that help keep more than half a billion users returning day after day. But in hitching their success to Facebook's reach, any developer that builds for Facebook's platform -- be it Branchout or Zynga -- can be sure to be in for a bumpy ride.

Just look at what happened to BandPage, a San Francisco-based startup that built a tool for bands to share their music, communicate with fans about tour dates, distribute news -- you name it. Founder J Sider, a longtime band and venue manager who started the business two years ago, decided to build it entirely on Facebook. That's where musicians wanted to reach fans, he concluded, especially in a post-MySpace world, and Facebook didn't offer anything that met their needs.

BandPage scored some big customers, such as Jason Mraz, and it quickly took off. Over the past two years, Bandpage lured about half a million clients, raised about $19 million in venture funding and, as recently as early January, boasted more than 30 million active monthly users across Facebook.

Then Facebook introduced Timeline, and with it decided to kill a marketing feature known as "default landing tabs." This meant that a brand's (or band's) Facebook page could no longer default to a custom app, such as a BandPage. Instead, fans would have to find the band's Facebook page and 'like' the band before getting to where the band -- and BandPage -- wanted users to visit. This redesign hit BandPage like a sudden thunderstorm at an outdoor concert. In a matter of a few months, traffic cratered more than 90 percent, to around 3 million active monthly users.

Sider isn't bitter, however. When he set out, he says, he always viewed Facebook as a launching off point. While riding the Facebook traffic train, he and his team have been building a standalone product, knowing that Facebook could crash his party any day, with little if any warning. In late July, BandPage rolled out BandPage Everywhere, a tool to let musicians manager their music, photos, tour dates, and so on, across the Web and on Facebook from a single dashboard.

"It's one big roller coaster," says Sider of working with Facebook. "It might be one hell of a ride and you might love it. Or you might get off the roller coaster and throw up. We decided to take the ride, and we're going to keep it going."

Facebook's 'black box'
One of the trickiest challenges for Facebook developers is navigating the tweaks Facebook constantly makes to "Edgerank," the algorithm that determines what shows up in the all-important News Feed. What appears in your News Feed might seem like a meritocracy, the result of how many of your friends 'liked' it or commented, but that's not the case. At least not entirely.

Facebook can decide whether an app is seen by thousands or tens of millions of people, in effect controlling just how viral something becomes. Users often think that something is suddenly popular when what's really going on is Facebook is running tests -- always in the name of user experience, but doubtless also with an eye on revenue, especially as a newly public company with a sagging stock price.

The impact can be huge. One week, an app is on fire. The next, it's reduced to ashes.

"It's incredibly frustrating," said one developer, who didn't want to be quoted by name for fear of souring his relationship with Facebook. "We all worry about Facebook making changes, and we all want to figure out how to get in the activity stream more. But much of it is a black box."

Facebook execs are guarded about exactly how the News Feed works. A Facebook spokesperson stressed that "apps grow on Faceboook based on quality and popularity among friends," but he also acknowledged that just what that means is Facebook's secret sauce.

Any active Facebook user sees this in action. Remember the weed-like growth of Zynga's Farmville? Gradually, that petered out -- at least in part, it seems, because Facebook determined that users were feeling spammed and, in response, took a whack at Farmville. The impact was big enough that in Zynga's latest earnings report its execs blamed Zynga's shortfall in part on changes Facebook made to favor new game titles.

Social reading was a fun one. This spring, we all started to see what our Facebook friends were reading on the Washington Post, Yahoo and elsewhere. One day I peeked at my News Feed to discover Go Daddy founder Bob Parsons and Marc Andreessen had both been reading an article with the headline, "Supermodels without photoshop." Even for Facebook, it felt like over sharing.

Certainly it was more than many Facebook users had thought they'd bargained for. While Facebook won't discuss the specific changes it made and why, the company started tweaking away and, poof, suddenly the traffic to social reading apps dropped off. It's as if Facebook creates and ends fads.

Take the money
I'm not saying that Michael Seibel, the founder and CEO of video-sharing app Socialcam, took the $60 million buyout offer from Autodesk last month because traffic was dropping on Facebook, but that's what was happening. While it's true that a lot of people loved Socialcam and rival Viddy -- Socialcam's app had 16 million downloads on iOS and Android since it launched in 2011 -- it leaned heavily on Facebook for its growth. And the numbers made it look like Socialcam's best days were behind it.

Socialcam is one of many apps that got a huge push by the rollout of Open Graph, in which apps get shared on your Timeline as well as your friends' New Feeds and Tickers. But Socialcam used Facebook aggressively, and its spammy and crafty techniques were criticized often. To watch a Socialcam video that someone shared on Facebook, for instance, you first need to sign up for the app.

The rise and fall on Facebook was spectacular. On April 1st, Socialcam had 570,000 monthly average users, according to AppData. By June, that number had soared 9,952 percent, giving Socialcam average of 57.3 million monthly users. Come August 1st, however, usage had all plummeted -- down to 31.9 million, a decline that likely had to do with Facebook responding to complaints and making tweaks. And the falloff continues on Facebook.

No matter. By that time -- on July 17, to be precise -- Seibel had sold to Autodesk, which says it will let San Francisco-based, four-person Socialcam continue to operate independently.

Working the relationship
Since things change fast on Facebook, app makers need to work hard to keep abreast of what might come. It's not easy, of course, since Facebook keeps its product roadmap and design plans close to its vest.

The relationship is one that Rick Marini, for one, can't let slide. Marini is the founder and CEO of BranchOut, the largest professional network that isn't LinkedIn. BranchOut, which has raised $49 million from big Silicon Valley investors, is entirely built on Facebook, and Marini's team meets with the folks at Facebook weekly to give them feedback and learn what they can about what Branchout could do better.

"We came out of nowhere and now have 30 million users, and there's no way we could have grown that fast without Facebook," says Marini. "Since the platform is always changing, we are always in touch with them. We tell them we see this kind of use behavior after you launched this change, and if things aren't working, they want to know."

Facebook, of course, is rarely out to punish apps on its platform. Its execs are doing what they think is best for Facebook, and if that leads to some pain, that's the risk of attaching yourself to the world's largest social network. It's a symbiotic relationship, and sometimes an uneasy one. But ultimately, Facebook has created a flourishing app economy across all sorts of categories -- shopping, music, reading, games, fitness, cooking, and on and on.

The power that Facebook carries is similar to that of Google with search. When Google makes major changes to its search algorithms, Internet roadkill is sometimes the result. But Google, like Facebook, is doing what it thinks is best: Google wants quality search results (with quality defined by Google), and Facebook wants a quality Facebook social experience (with quality defined by Facebook).

And like Google, Facebook is now a big public company. As such, it's making decisions based not just on user experience. There are also millions of shareholders to satisfy and so there is added pressure to maintain revenue growth . It's that reality that leads Dalton Caldwell to conclude that Facebook, like the ad-driven Twitter, is no longer a true "platform," or what Zuckerberg used to describe as a "social utility."

So instead of working with Facebook -- or selling out to it -- Caldwell launched a Kickstarter-type campaign for App.net, which he describes as a "real time social feed without the ads."

It's gaining traction. Caldwell's been asking people to become members in advance and, remarkably, yesterday he met his $500,000 goal and, as of midday today, with 13 hours left in the campaign, App.net has landed 11,084 backers who have committed $724,650.

"Part of my thinks this is a crazy idea," says Caldwell, who sold the social music startup iMeem to MySpace in 2009. "This is my bold experiment."

It's a sentiment that many current Facebook developers might one day decide to act on as well.