Facebook just announced Facebook Connect. This is potentially huge - I think this is what could truly make them into an indispensable Internet infrastructure layer: the main provider of Identity. If executed well, they could entrench themselves into the online fabric and become the best way for applications to authenticate users and gain access to their personal and social information.
This has obviously been tried before. Microsoft (Passport/Live ID), Yahoo (BBAuth) and Google all have APIs that allow developers to leverage (to slightly different degrees) the large number of user accounts they already have. There are also more open efforts like OpenID that have similar developer-facing goals while trying to decentralize the control of identities. All these efforts have had very limited traction, probably for one or more of the following reasons:
1) Not enough value for the developer - Sure, they will save some coding time, but there is a perception that they will lose control over their users and it is someone else that will ultimately own that relationship. Plus, there isn't much use for these accounts beyond authentication. This is definitely the case with MS/Yahoo/Google.
2) No critical mass - There is little reason for a developer to use such a system if he won't benefit from a pre-existing critical mass of user accounts. This may not be the case for the big guys, but it definitely is for OpenID.
3) Bad UX - Unless the authentication process, and getting the ID in the first place, is easy and seamless, most developers won't risk requiring their users to get such an ID. So far, none of the above have hit on both counts. As far as I know, using systems like Live ID is not completely seamless. And getting an OpenID is not something that ordinary users know how to do.
So we are left with each application out there coding their own identity system, requiring users to memorize new IDs/passwords and re-enter their personal/social information. Facebook, however, has a real shot at changing this. Until now Facebook was just another app/walled-garden running the risk of losing its buzz and seeing users leave for the next fad. But Connect can change the game they are playing completely. Why could it work? Addressing the same three points as above:
1) Value to developers - Facebook has more explicit info on their users than almost anyone else (you could argue Google knows almost as much - but it is mostly implicit). They know their names, locations, work history, education history, likes and dislikes. Importantly, they also know who their friends are. There is enough value here that I think many developers will think twice about not leveraging this information.
2) Critical mass - Facebook may not have as many user accounts as Yahoo or Microsoft (roughly 500M vs.100M?), but they are within striking distance. Furthermore, it's likely that a much greater percentage of Facebook accounts are active and actually represent real people and have real information associated with them. I don't think critical mass is a concern.
3) UX - This is where they could stumble. They need to come up with an elegant and seamless way for developers to weave Facebook identities into their apps. My guess is they can do this.
How would this change the game for Facebook? By turning into an app infrastructure layer, they would not have to worry as much about driving page views on Facebook.com. Most innovation on the Internet is going to happen off Facebook anyway and they can't stay on top forever. With Connect they could just focus on having the best identity system out there. This is probably composed of two parts:
1) The set of APIs that let developers leverage Facebook IDs and all the information they contain
2) Tools for end users to create those identities and populate them with their personal and social information (essentially what Facebook.com is today)
Now for the hard question: how do they monetize this? They are already having trouble monetizing their own page views... how are they going to monetize something that end users don't actually see even if it becomes wildly successful?
The first, obvious way is to continue to rely on ads on Facebook.com. They are going to keep on working at this and may actually figure out a way to bring eCPMs up (perhaps by monetizing influence - I've written a post on this before). But there is nothing new with this option and is in fact a bit of a cop out since I wrote above that FB could stop worrying about driving Facebook.com page views.
The second way is to monetize third party inventory (the same post also talks about this). This makes even more sense in scenarios where Facebook is providing the identity infrastructure for a site. Why couldn't Facebook require developers using Connect to also use "Facebook AdSense"? It complicates things a little bit and not all developers may be willing to do this. How about a freemium identity system? Developers get access to more info if they place Facebook ads on their site... It's worth a try. Again, I really think Facebook needs its version of AdSense.
Third, there will probably be other ways to monetize this that I can't think of right now (talk about a cop out...). But I am confident that, one way or another, an Internet-wide identity provider would be very valuable. And I think Facebook has a shot at becoming this provider with Connect.
Finally - I have to squeeze this in here. I think it's somewhat sad that Microsoft (my ex-employer) didn't do this ages ago. They had the vision - hell, they came out with Passport ages ago. They also had a lot of personal information (through sign-up requirements and other properties) and social information (through Hotmail and Messenger). It's unfortunate they haven't been able to capitalize on this.