Facebook‘s developer platform was a watershed. It gave outside companies unprecedented access to users of a major social network. But the platform, launched in May, has come under increasing scrutiny. Facebook has continued to change the technology and the rules governing what developers can and can’t do on the site. The issue is whether or not Facebook’s platform, including newly-launched “Facebook Connect,” is stable enough to justify other companies making the effort to use it.

Heralded last year as one of the most exciting new areas for investment, the platform has seen its hype diminish, even as it has gradually matured. But third-party applications are doing better than ever, Facebook says — especially apps built to make use of the redesigned interface the company introduced this fall.

Facebook’s efforts to innovate ahead of other web services — rivals such as MySpace and Twitter — have left the platform in a constant state of change. At times, the company has appeared to compromise its goal of offering a “level playing field” to third-party developers in the name of innovation.

For the millions of Facebook users out there, the everyday impact of rule and technology changes is that their favorite apps — whether simple “poking” apps like throwing sheep, or more complex video-viewing apps and casual games — occasionally disappear or become harder to find within the Facebook site interface. On the good side, the spammy apps that send unwanted messages tend to get shut down.

The rule and technology changes make Facebook an increasingly useful service, but the fact that they happen so often and with little warning has created misgivings among some leading developers about working on the platform. For example, recently — and without prior warning — Facebook moved the notifications that apps send users about friends’ activity. These notifications now appear in the toolbar at the bottom of the site but no longer on the homepage. The result is that apps may be harder to find (although Facebook says its internal data shows increasing app usage as a result of the change, because users see the toolbar on every page of the site).

Future ramifications of Facebook’s changes to the platform may be more significant. The company recently rolled out a new service, called Connect, that lets users sign in to other web sites using their Facebook identities, then share information that their friends can see on Facebook. More than 100 partner sites, including CNN, Digg, Citysearch and many others, have been announced as Connect partners. So far, many of them — including Digg — haven’t fully implemented Connect’s features. Adopting Connect opens up a third-party site to becoming dependent on Facebook. The data a site may choose to send Facebook may say a lot about users’ behavior. For example, if Digg lets its users send a link to Facebook about what they’re reading, Facebook can learn what sorts of stories Diggers like to read. What if, one day, Facebook decides to add a news aggregator to the site as a core feature?

Depending on how Facebook manages its platform in the coming months, it could either foster more serious investment from outside companies, or scare them off. The following analysis is based on extensive interviews I’ve had with Facebook executives and numerous third-party developers on the platform. It is an attempt to sort out the present risks and opportunities for outside companies.

Facebook has the leading platform

If you’re a developer and you want to build an application for a social network, Facebook has been and will continue to be the most attractive option for the foreseeable future. Its platform dominates. The “canvas” pages that third-party apps use to operate on the site get a total of more than 500 million page views per day, the company says. That’s almost six times the 85 million daily page views generated by all applications that use the competing OpenSocial standard. In sum, 52,000 Facebook applications create more than 14 billion page views on the site per month. More than 70 percent of people on Facebook use at least one application per month. More than 4,200 applications have more than 10,000 users per month.

OpenSocial, a Google-led initiative to help social networks offer platforms, launched a year ago with the intention of aiding rival social networks like MySpace and hi5. It allows third-party developers to build apps that can be easily implemented on any OpenSocial social network. So far, there have been 7,500 total OpenSocial apps created across more than 20 partner sites. Right now, the initiative is tiny compared to Facebook. [Click through to read rest of story]


Meanwhile, Facebook has quickly grown to reach more than 140 million monthly active users worldwide, according to internal statistics published by the company. Outside analytics firms like comScore invariably show that Facebook’s worldwide traffic overtook MySpace’s this past year to become the largest social network in the world.

When Facebook launched its developer platform, it declared that it would offer a “level playing field” for third parties — which became a key factor in luring companies to embrace Facebook’s platform. “We reserve the right to build anything and compete with any of our applications,” said Zuckerberg at last year’s Web 2.0 conference in San Francisco. “But we intend to do it on equal footing and on fair ground. We have gone a long way with the applications we have now to not take advantage them.”

Zuckerberg, recently reiterated this point to me in an interview:

What we wanted to do was make the platform as neutral as possible — to open up a core set of services, a set of different ways to communicate, like news feeds and [application invite] requests. We wanted to make it so that all apps can do the same set of things.

But things have been complicated

The past year and a half have been a running battle between Facebook’s rule and technology changes and developers’ efforts to work with them — or around them.

Facebook has consistently given its own applications special advantages since the platform launched. Its own Photos application, which lets users upload and share photos, competes with a range of third-party apps that offer photo-sharing and other services. Since last fall, for example, Photos has been featured more prominently than rival apps in the company’s news feed. It shows you when friends have uploaded new photos that have nothing to do with you, for example — an action that Facebook doesn’t allow a third party to display in feeds.

Tipping the playing field in favor of Photos is one small example of a platform rule change. The changes to the notifications, mentioned earlier, is another. Here are some other examples. After third parties began offering fan pages for musicians and other public figures last fall, Facebook came out with its own Pages. After third-party applications let you designate who your favorite friends were on your user profile, Facebook introduced that feature. Last month, after other applications began offering ways for users to be automatically notified about friends with upcoming birthdays, Facebook came out with its own feature to do that.

Facebook says developers just get more interested in building category-specific applications when it builds more core features for them to use — like the ones listed above. The company cites comments made publicly by developers of calender-notifications to prove this point, like this statement by Raj Lalwani of SocialCalendar, an app developer affected by the calendar feature change. Lalwani put on a brave face, but his conclusion was more ambivalent: “Constantly evolving, fortunately or unfortunately, is the only way to succeed as a startup in Silicon Valley.”

Privately, we’ve heard some of these calendar-app developers are saying things more along the lines of “time for us to work on other social network platforms.” Meanwhile, a Facebook representative replies that “if your entire business is built around a feature that Facebook can add as one radio button, it’s probably not the best business model.”

Of course, many third-party developers have also been breaking rules. By July, Facebook had disabled or forced rule compliance from more than 1,000 blatantly spammy or otherwise pernicious apps.

When the platform launched. Facebook had straightforward rules in place governing what apps could and couldn’t do, but it had no policy or process for dealing with infringers. Developers, ranging from venture-backed companies to college students in their dorm rooms, quickly figured out they could grow with impunity by using deceptive methods to get users to install apps. One of the more popular methods of deception was for an app to tell a user they had to invite friends to get started. The user’s friends would then receive an email with a link that, when clicked on, automatically installed the app on their profile page.

In the months after the launch, users’ profile pages became cluttered with apps that they hadn’t intended to put there in the first place. User anger began to mount.

Instead of taking direct action against the infringers, Facebook tried to combat them with technology, often disabling Platform features that had been exploited. For example, after apps gamed the news feeds to disproportionately show their users’ activity, Facebook drastically scaled back the ability for apps to appear in news feeds at all. It didn’t start taking direct action against perpetrators until March of 2008.

Because Facebook launched the platform with relatively few rules — and no experience being a platform company — the company’s ongoing changes are like “rebuilding an airplane in mid-flight,” says Mike Schroepfer, vice-president of engineering.

The wrong move risks sending the whole platform into a nosedive, the right move makes the plane go faster.

The battle escalates

Facebook’s efforts to reshape its platform culminated in last summer’s redesign, which all but eliminated one of the most prominent features of the platform up to that point: The profile box for applications. That interface was a simple way to display an application, very similar to the simple widgets that had been on rival social networks like MySpace for years. Facebook disliked the concept because it didn’t promote information-sharing to the degree that its news feeds did. Developers that relied on the profile box to get traffic saw their numbers plummet as the redesign was implemented on the site over August and September.

Zuckerberg, on the redesigns impact on developers:

What we’ve found was that there was some confusion early on. We changed a number of things and iterated a lot after we launched to get things back to a point where baseline levels of [app] activity were the same as before. Apps are slightly less front and center now and somewhat out of the way, that’s fair.

But the company has been clear about its intentions since the platform launched, Zuckerberg says: “We didn’t see it as being a widget platform.”

This is small consolation to companies suffering big losses because of the near-elimination of such an important feature. “To call Facebook a ‘platform’ is an insult to the word platform,” says one developer who asked to remain anonymous. “Platforms are fundamentally supposed to be stable and reliable, and Facebook has been anything but.” In August, Slide vice president Keith Rabois told The New York Times that “no one wants to build something that just becomes a [research and development] company for Facebook.”

Facebook, meanwhile, says that the core application programming interfaces and canvas-page features of the platform have remained stable. It frames the redesign as refocusing the site on user activity streams (the news feeds, the profile feeds, etc.). The redesign has so far been one of the starkest instances of Facebook choosing innovation over stability.

Even as the redesign hid third-party apps, Facebook continued to feature its own. Photos, for example, still automatically shows up on every users’ main profile page — users have to add third-party apps manually to the main profile page, and even then they can only add a few. In fairness, Photos has been a core Facebook feature since before the platform launched. The example nevertheless shows that Facebook cherry-picks areas where it exclusively promotes its own features. Facebook’s argument for doing so is basically that its apps offer core forms of communication, so they need special prominence.

The Photos app is illustrative of how the company makes product decisions, says Schroepfer. He joined the company this summer from Mozilla, the nonprofit that builds the open-source Firefox web browser — also a platform for third-party developers. When Facebook launched in 2004, it originally offered only profile photos, but users kept uploading so many photos that it began offering a more general photo-sharing service, Shroepfer explains. It’s now the largest photo-sharing service in the world and a key way that users interact with the site.

Facebook’s overall goal is to help people share information with each other more easily. User data and user interactions form the center of Facebook’s value. In Facebook’s view, farming out such services to third parties risks making Facebook irrelevant and destroys the cohesiveness of the user experience on the site.

To some developers, though, every instance where Facebook gives its own apps a special advantage seems like a subjective move that contradicts the platform’s founding philosophy.

Rule changes aren’t the only thing that rub. This summer, Facebook announced a new program called Great Apps, and another called Verified Apps. The programs are intended to reward apps that Facebook considers exemplary. Great Apps get special benefits. They receive more prominence in news feeds than other apps, for example, and can have users send and receive more invites per day. So far, the company has only designated two Great Apps: The fundraising app Causes, and music app iLike.

Last month, Facebook began deploying the wider-ranging program called Verified Apps. Interested developers can pay a small fee and apply for this designation; if they receive it, they’ll also get more prominence in news feeds and other benefits similar to Great Apps. Details are still coming out about how Verified Apps is different from Great Apps. Presumably, Verified Apps won’t get as many benefits.

Some developers have grumbled that Great Apps is a sort of welfare program for apps that aren’t otherwise getting traffic on their own merits. Facebook says these new programs are intended to incentivize good behavior; they are ways for the company to start playing good cop. Both programs are still young, so it’s hard to judge their significance yet.

The strange case of Facebook Music

It’s one thing for Facebook to tweak the rules to prevent user abuse and block spam, it’s another to attack the core feature of a leading third-party developer. Which is why many of us who cover Facebook closely were surprised to hear that the company was looking at offering its own music service.

ILike beat out a number of competitors shortly after Facebook’s platform launched to become the main third-party music provider on Facebook. Over the last year and a half, iLike has expanded from simple beginnings to offer a broad suite of music features somewhat analogous to those of Myspace Music. It began offering full-song streaming music through a deal with music company Rhapsody, for example, earlier this year. ILike chief executive Ali Partovi has cited his company’s efforts as an example of the platform’s level playing field.

However, only weeks after Facebook designated iLike as one of the two Great Apps, Facebook’s business development team began talking with both iLike and a number of competing online music companies about its own plans for offering a streaming music service. Facebook’s idea is to offer many of the most popular features already offered by iLike, including the ability to create a playlist, see friends’ playlists, import playlists from iTunes to share with friends, share music in the newsfeed, and let other apps play music through the Facebook app. The main difference: Facebook’s implementation would have special advantages not afforded to iLike and other third-party apps.

Facebook has been considering the idea of either integrating iLike more deeply or building its own features. However, it’s still speculation to say what Facebook is or isn’t doing, Zuckerberg says:

It’s to be expected that we’ll have conversations about different things. We haven’t done anything with music. It’s not clear that we will. The reason we’re considering it is only because there’s so much complexity around music licensing and rights. It’s not even that we want to build something ourselves. We’ve talked to a lot of music providers about how to do rights work, about how we could use our leverage to help them get rights.

The iLike app is restricted by record-label licensing deals, so Facebook users can only play up to 25 songs per month before they get only 30-second clips (otherwise, users have to pay for unlimited access). Facebook would surely prefer to give its users unlimited and free streaming music, comparable to MySpace Music. But there are practical issues with Facebook’s plans for licensing music. Perhaps the largest is that music licenses typically only last a year or two; when a license runs out, the music labels will just demand more money.

But the philosophical issue is this: When the platform launched, Zuckerberg basically said that Facebook wouldn’t be focusing on specific media applications like music:

We’re not really a media company. To the extent that someone is building an application that is focused on specific media, that is unlikely to be something we would want to go after… The types of problems we solve, are all deeply technical problems.

Since then, apparently, Facebook has redefined what it considers core features, and now music could be one of them. The risks this poses to iLike, which relied on Facebook’s word, are being interpeted by some developers and Connect partners as a warning about what Facebook may do in the future.

Analysis: “Manifest Destiny, Facebook Style”

Harvard’s student newspaper, The Crimson, penned an editorial in March, 2004 called “Manifest Destiny, Facebook Style” that ridiculed Columbia University students for protesting Facebook’s invasion — Columbia’s engineering school had its own, less developed social network at the time, called SEASCommunity.com. It was one of the earlier social networks to get steamrolled by Facebook.

Why has Facebook been so successful?

Because Zuckerberg and his company have, so far, been the best at building features that users want. By this point, there’s a very long track record showing this. Facebook designed the site so people could only see profile information about their friends or people in their geographic vicinity, starting with college campuses. Users became convinced that the site was a safe place for sharing real information about themselves and living their real lives online. Over the years, it grew to high schools, companies, and eventually opened up to the world. Then it launched the platform. The entire time, it has stayed focused on helping people share information about themselves. It came out with news feeds before any major rival, it came out with status updates so friends could tell each other what they were up to — around the same time as micro-messaging rival Twitter.

Today, the company is keeping this focus on communication, even as it continues to try to figure out how to run its platform. If Facebook fails to be the best way to share information, it could fail just like the other social networks it has bested. Zuckerberg knows this, and it’s why Facebook tried to buy Twitter:

I view it as a big part of Facebook’s role in this ecosystem [the web] to continue pushing [information sharing], to continue making it so that more people around the world can bring all of this information [from Facebook] to these different places. You know, there’s a rate at which this will happen, and if we’re not on the edge of pushing this out, then we’re not doing our job.

If applications on Facebook’s platform encroach what it considers key areas — like Photos — Facebook has shown it will tip the playing field to its own favor. This likely explains why Facebook could potentially compete against iLike by giving itself special advantages. Facebook knows that music is a great way to keep users engaged on the site, and it knows it has built a site that could make it very easy to share music.

But wasn’t the “Great Apps” program intended to do exactly that, offering the chosen applications (iLike and Causes) the opportunity for deeper integration and enhanced promotion so that they’d be surfaced to more users? And what’s to prevent Facebook from applying the same logic to invade any other business in the future?

Says another upset developer who wishes to remain anonymous: “What’s at stake is nothing short of an identity crisis for Facebook: Does it intend to be a technology platform like Windows and Google, or yet another vertically-focused media property like Yahoo or CNET?” Facebook insists it is staying a platform company.

But the changes also raise questions about Connect. What if Facebook decides to create its own local-review feature one day, to compete with Citysearch? — Facebook Pages already offers a sort of review-style service, after all. What if Facebook decides to create its own news aggregator feature one day to compete with Digg? Why should these companies trust Facebook to identify users on their sites, and create traffic on Facebook instead of on their own sites?

It’s understandable that Facebook feels it must focus on innovation at the expense of some applications. But its actions result in a very real cost for the company: Negative public perception. The company hopes to develop new ways of making money through traffic generated by third-party apps. These apps generate billions of extra page views — interaction that makes the site more interesting for users — and lead to more opportunities for Facebook to make money through advertising. The company has considered ideas like developing a payments service to let apps offer ways to buy and sell goods within the site. It may one day try to monetize Connect by helping other sites target ads to Facebook users who visit them. The company isn’t saying much on these fronts, but it’s actively testing out a variety of ways to serve ads on the site that emphasize making it easier for advertisers to reach Facebook users through sharing features.

Developers know Facebook is valuable; Facebook knows its platform is valuable. And Facebook understands the issues. The numbers show why developers are still building apps — as Zuckerberg says, they get social context:

They can get sharing, through users’ experiences on Facebook. Developers can choose to invest more to take advantage of this. But we’re not necessarily asking large companies to invest millions of dollars the first time they decide to experiment with the platform. The platform is still young. We’re open about that. There’s a reason that more than 10,000 apps have been built. Value is being given. We’re going to be investing in it more and more.

The problem is that trust, once lost, is hard to regain. Smaller developers will no doubt continue to pay what some call a “tax” of making constant changes to their applications. But will larger developers stay? And will other web sites fully integrate Connect — not all of them are yet, although the service has only been rolled out for a little over a week. As yet another anonymous developer put it, “they can’t have their platform and eat it, too.” Google has launched its competitor to Connect, called Friend Connect. MySpace is working on its own.

That’s Facebook’s gamble. If it continues to be the most popular social network in the world, and if its constant changes result in a better experience for its users, developers will continue to build for it. This end will justify the company’s means.

[Old Facebook profile screenshot via Maxopedia; “Manifest Destiny” painting screenshot via History on the Net]