What would be really interesting is if the social networks start using eachother's APIs to import and track friends from rivals. You could add the myspace app to your facebook account, and then you could track your facebook friends while using myspace (vice versa ad infinitum). After the carnage, there would be a single portable social graph. And the sites themselves would be commodities.
Well with what you describe, the graph would become the commodity, the sites themselves would remain unique and competitive, but as interfaces to the graph.
Overall I think your idea sounds like a win for all. Users could have a single presence at the site of their choice, with friends from other sites, new entrants would have a lower barrier of entry, and existing sites are less likely to lose users to the network effect. Of course, the big players would probably have more competition from niche players too, so I guess that would be the downside for Myspace and Facebook.
The fifth paragraph made me think. What if Google, Yahoo or Microsoft started a social network based on the addresses in their webmail apps? It would instantly be bigger than Facebook.
Obviously size isn't the only thing that matters. That graph in the middle of the article is a pretty sick demonstration that MySpace is dominating this game. And yet, nobody's trying to buy it off of News Corp. Valleywag's claimed that Rupert Murdoch is a lot more excited about Facebook than his own purchase. As is the rest of Silicon Valley, hence the valuations that have nothing to do with the user numbers. Or the yearly revenue...
The thing that is most similar between (post-IPO) Google and Facebook is the immense, almost unbelievably universal goodwill held about the company by the public. Goodwill's not actually worth much money in and of itself, but if the entire world wants Facebook to make money, it eventually will. Let's hope that happens before the users leave.
So if it isn't the users and it isn't the revenue, then Facebook's insane valuations are the product of... hype?
Valuations aside, you're right in saying what really matters is monetization and this is where Google shines. I don't care how much information Facebook has about me, I don't go onto Facebook looking to buy things. I go there to see what's going on in my friends' lives. As the article states, ads on Facebook are an annoyance.
Google's in the right market for targeted ads because when we're looking to buy something, we're more open to advertisements and clicking on them. On Facebook, I'd guess that very few, if any, are looking to buy something (and hence Facebook's abysmal click-through rate).
Also, I can't see users leaving Facebook (eventually the time sunk into building connections just entrenches you into the network), but I hope it figures out something quickly, or these valuations might drop.
The most interesting, underutilized and least known aspect of Facebook's platform, imho, is the fact you can export Facebook data to other sites and other communication systems (albeit in a limited way). O'Reilly's been blogging about this lately.
"I do think that there are two views of the social graph, though, and how it gets deployed: there's a platform view in which it can be exploited to build smarter applications on Facebook; there's a deeper view in which the Facebook-discovered social graph can be accessed by other applications elsewhere on the net. Facebook's granular control of what information you reveal and to whom is thus a key part of the platform -- but the question is how far Facebook will go in letting other sites use this information. If Mark's answer is the first, Facebook is ultimately a closed platform; if the latter, it becomes a true open platform and value enabler."
The Economist article also confuses news feeds (your friends actions, on your FB homepage) and mini-feeds (your own actions, on your profile page). Makes you wonder if the author has actually ever been on to Facebook. Perhaps there's less to The Economist's analysis of Facebook than meets the eye?
Actually, Facebook has a very sophisticated infrustructure. Take a look at Thrift, http://developers.facebook.com/thrift/, the open-source project for cross-language development they released earlier this year. Even Google is trapped using about four programming languages (C++, Java, Python, and JavaScript). The latest version of Thrift has interfaces for 10 languages (Cocoa, C++, Erlang, Haskell, Java, OCaml, Perl, PHP, Python, and Ruby).
