Facebook Will Die Within Three Years, Blogger

Has Facebook jumped the shark? Although I believe that Facebook is a valuable tool for certain communication purposes, it is difficult not to conclude that Facebook's model for what social computing should be if forward-incompatible. Although it was an early innovator in social networking, Facebook is now the anchor that resists the social web's movement toward a less centralized architecture. The harder that Facebook resists the tide, the faster it will sink. Unless Facebook drastically changes its approach to social web services (and soon), I predict that it will be dead within three years.

All of this should come as no surprise. The internet has always been, at least at its core, a decentralized architecture in which no single entity can exert control over the system as a whole. The internet as a technology enables a free flow of information directed by the decisions of users and not by corporate of governmental edict. Time and time again, the services that understand and embrace this paradigm have prevailed in the long term over short-sighted businesses that attempt to make a quick buck by seeking to control the flow.

The kind of control that I am referring to is not that of censorship or of copyright protection or of any of things that we traditionally associate with control over information (though these concerns are also salient to a discussion of the emerging social web). Instead, I am referring to the attempt by web service providers to build silos and walled gardens rather than creating open APIs and data architectures. Social networks that try to fence in their users always fail in the long run while social networks that allow their users to expose data and functionality (on their own terms, of course) to other services always win in the long run.

Facebook understands this; but rather than accept the rule and adapt its business model accordingly, Facebook has tried a variety of clever tricks to try to break the rule. Take, for example, the Facebook Application Platform and Facebook Connect. Ostensibly, these services seek to enable users to expose data and functionality (on the user's terms) to third party developers who can integrate new functionality to the Facebook experience. But therein lies the problem: Facebook sees itself as the unifying force that ties together disparate social networking services rather than as a mere social networking service. Consequently, the scope and nature of services that can be built is limited by Facebook's own barriers. Further, the circumstances in which third party services can read or write information is dictated by Facebook. Most importantly, Facebook does not intend to expose its data and functionality for display on another site, but rather intends for itself to be the forum in which data and functionality are exposed.

There is nothing wrong with seeking to be the place where disparate services are tied together. In fact, sites like FriendFeed do precisely that and do it well. The problem is that Facebook is trying to do too much. It wants to be both the patches and the stitching (or the paper and the glue, if you prefer). By simultaneously creating the environment in which services come together while also to providing those services itself, Facebook limits its users' ability to choose alternate services to fulfill a role similar to that of Facebook's core functionality. For example, it is difficult for a Facebook user to choose an alternate photo sharing provider in a way that integrated properly into the rest of the Facebook experience. Facebook will not allow seamless integration in this respect because Facebook's own photo sharing service is of strategic business importance to the company. They do not want to allow competition in this area, so they don't.

The emerging picture of the web looks very different from that painted by sites like Facebook. In the new web, service providers do not seek to do too much. Instead, encapsulate functionality into manageable bits that can be seamlessly integrated with other services and other service providers. Examples such as Twitter, Flickr, and Revver abound. These bite-sized services allow their users to expose data and functionality on the user's terms in order to integrate with other social web services. Breaking the Facebook paradigm, these sites do not really intend to be the user's landing page. Instead, they assume and encourage that their users will find creative ways to assemble various social web services into the location of their choosing (this location becoming that user's identity on the web).

Think of an apple orchard. In the new social web, micro-services like Twitter are the apples. Sites like FriendFeed are baskets. Each user chooses her own basket, and then fills it with the apples of her choice. Advanced users might even make their own baskets (that is, integrate social networking services via server-side scripting into one's own website, as I do).

This model offers users several advantages over the traditional walled garden approach. First, users have greater choice in selecting service providers. Don't like how Facebook status works? Use Twitter instead! Don't like that? Try Brightkite. Second, users have more fine-grained control over the disclosure and exposure of their personal information. Since service providers are not seeking to be everything for everyone, users can set privacy policy on a smaller scale within each service rather than on a larger scale through a site like Facebook. Moreover, users are empowered to choose service providers based on the privacy options available to them. Third, users can have redundant social web services in the event of companies going out of business. The recent news that Google is shutting down several of its social networking services highlights the need for social network users to have a "Plan B". In a decentralized model of social networking, users can have completely redundant services that can be swapped in on a moment's notice (for example, my micro blog lives on WordPress, Twitter, Identi.ca, Brightkite, Facebook, and elsewhere).

There are also disadvantages to decentralized social networking. In particular, it is notoriously difficult to monetize social networks already. By moving toward a more decentralized structure, making a profit by offering social web services will become even more difficult (especially considering that for any given service, there is always a free alternative). The idea that social networking can be profitable on a data mining + advertising model is long dead; companies like Facebook waste their time and resources attempting to realize that dream.

Despite the lack of profitability, social networking will flourish under the new paradigm. The development costs associated with building smaller services is lower than the cost associated with trying in vain to build a comprehensive walled garden. In any event, most of the development and innovation will come from hobbyists rather than professionals, from open source rather than proprietary software. Incidentally, the future of content management systems like Drupal and WordPress lies in providing users with the server-side tools required to tie together other social web services. In other words, "content management" will become less about content creation and more about content gathering.

As Mitch Kapor has argued, architecture is politics. The architecture of the internet is that of a decentralized information system. Consequently, the services that flourish online are those that embrace decentralization. This is a very good thing for users because:

The decentralized architecture of the Internet minimizes the role of central authorities and maximizes the ability of any participant to offer or receive any information or service and to develop new capabilities and services.

Sites like Facebook are doomed unless they radically alter their business and development models to reflect the needs, interests, and capabilities of internet users. There is room for Facebook to move away from providing service and toward assembling the services of others in a single location. Frankly, I think that this is the only way that Facebook will survive the coming revolution in social computing.

By Devin Johnston's blog, used under Creative Commons.