Dollar Value of a Datamined "Free" Service User?

[grarpamp]

Many companies do not charge users of their services.
They say things like "It's free and always will be."

Pick any "free" service... Facebook, Google, Yahoo, Match, Reddit,
the list goes on and on.

Only the smallest of free services can be supported by someone's
personal funds. So for larger services it's well understood that if
their users aren't paying them, the money must come from somewhere.

The old model answer of former years was advertising through page
ads, clickstreams, referrals and those sorts of obvious and observable
things. Combined with burning off stock market offerings.

The new model answer includes all the former ways, plus much deeper
profiling and analysis of users, their pictures, messages, interests,
locations, etc... all done on the secret unobservable corporate
backend, and selling off that metadata to the highest bidders.

And perhaps more dollars from simply lying, violating their publicly
posted hole ridden privacy policies, and selling directly identifiable
user data as well. Or say taking immunity from prosecution from
having given/sold it away to cozy governments.

All of which have monetary values associated with them.

Companies also hide their revenue, costs, and their "active in last
three months" vs "dead user" numbers.

Some usable public statements and government/market filings are
available. And datacenter and employee costs can be approximated.

There is current value analysis, such as in selling what's hot by
this months clicks. And future/lifetime value analysis, such as
Target datamining for pregnant women to own. There are also dataset
specific values, such as names and addresses vs political alignments.

The values could be represented in $dollars per month per user.


Anyway, question...

Anyone have links to studies made within the last five years that
have attempted to calculate the dollar value of datamined users?