Great piece. Not so sure about the network effect and the applicability of Metcalfe's Law of data tho. Perhaps the challenge is with the conflated way in which most people think about "value" and who is the recipient of the value. At Gartner we ascribe three kinds of value to information 1) its realized value from actually deploying it, 2) its probable (accounting asset) value that reflects the likelihood it will be used, and 3) its potential value if it were deployed across all relevant business processes. We do this to help clients understand and measure the value gap between their underutilized "dark" data and what its information value potential is. Yes, I have developed models to quantify all this.
For Facebook, its realized value of data is the $3.7B of revenue, well because all they do is sell data (more or less). They're a pure info-based business. But it's information's probable value (in accounting terms) is the gap between its recognized $7B in assets and its investor-expected ability to monetize this info (i.e. its current $63B market cap), or $56B. Who knows for sure whether FB will actually be able to monetize its data more or less than this.
The WSJ published a piece of mine pre-IPO in which I discussed this and computed the value of an account at $81. (WSJ-To Facebook You're Worth $80.95). Interestingly, researchers at the University of Vienna saw this and contacted me that they'd done a study showing that users threatened with FB deleting their data, value it at $12 on average. So yes, a user's data has more value to FB and to other users than to the user him/herself.
But this isn't a network effect at all. A network effect says the value of a phone (or one's FB account) grows the more other people have a phone (or a FB account). The data is produced by using the account (like a phone call). An odd thing about data is that the cumulative realized value of a piece of data grows the more its used/shared/copied, but the data's market value diminishes because it has become less scarce.
Ultimately in your argument I think you're first using data in the singular (piece of data) and then in the plural (data set), which somewhat confuses things. So yes, the aggregate realized value from a piece of data grows the more it is shared, but its value to FB or to any given consumer of it is actually reduced. That is, I'd pay less for data that everyone else already has.
In short, we all need to be careful when we use words like "value" (which kind of value and to whom?) and "data" (a single piece of data, all of its copies, or a data set including that piece of data?) in the same breath.
Keep up the great ideas and writing!
And for more on the topic of information economics, or "infonomics", see recent pieces also in Forbes and the Financial Times:
Forbes - Infonomics-The Practice of Information Economics
Financial Times interview - Extracting Value from Information (free registration req'd)
Doug Laney, VP Research, Gartner, @doug_laney