The trend moving forward is that social networking sites are moving towards opening up their systems to third party developers. Facebook, initially aimed at college students, is currently leading the pack by a long shot. This will provide countless opportunities for entrepreneurs to solve a variety of the social web's existing problems: Non-centralized authentication, and excessive data duplication.
Let's talk about two of the biggest problems first.
Excessive Data Duplication
Registration for social networking websites typically involves the need to provide profile information. Often, if you have accounts on multiple social networks, you need to login to each individual website and update each profile whenever you make changes. The data, and its input is duplicated for every social network you are a member of. At various stages, profile data spanning multiple websites will be in an inconsistent state, or even worse, become irrelevant. This 'silo' technique of hoarding data is clearly no good for the user's or for these companies' bottom lines.
Non-Centralized Authentication
We deal with this joy everyday. How many hundreds of website do you have logins and passwords for? If you step back and think for a second, it is almost ridiculous that every single new website demands you to sign up and remember another password. Truth is, anyone who solves this problem is going to change the world. Many have tried (remember MS Passport?), nobody has succeeded as of yet. The push for a universal identify management system is stronger than ever.
The Opportunity
Social networks developing out 3rd-party development hooks presents a great opportunity to solve these aforementioned issues. By using the spirit of the web 2.0 movement and distributed systems, it would be possible to 'glue' all the services together.
For instance, consider two services: a global identity management system (like openID) and an user profile service (a sort of "one profile to rule them all" service). Social networks would now have the possibility to authenticate against the identity management system in order to log a user in.
Next, the 'user profile service' could operate in a number of ways. I envision the user controlling their master profile on the 'user profile service' website, and then having a number of options to 'push' profile data to social networking sites or 'pull' profile data and new fields from them. From a technology perspective this is not hard to accomplish. However, the politics of implementing these two solutions is going to be tough indeed.
Conclusion
Third party development hooks into today's hottest social networking sites is going to present a lot of opportunities to make money and/or change the way the world works. The burden of maintaining countless passwords, and redundant profile data is one of the cumbersome inefficiencies that we may soon be able to eliminate.
