Twitter are fairly transparent in their policies. You should read this if you want to get a sense for how a large social network can operate with "best practices" -- at a level that's practical anyway -- with respect to protecting user information: https://support.twitter.com/articles/41949-guidelines-for-la....

Of particular interest is their link to http://www.law.cornell.edu/uscode/text/18/2705, the part of the Stored Communications Act that allows a gov't entity to prevent the network from telling the end user about investigations under certain rules.

You should also take a look at the EFF's best practices stuff, such as https://www.eff.org/wp/osp (dated).

The article only covers the police. I wonder what the potential differences there are that could be in effect for other agencies in the US or international forces. Is there still a real black back door for the right level of clearance ?

Well other international forces could break in to those back doors. Or maybe facebook could gladly allow them in. As they say nobody is above the law!

"Police need a subpoena in order to access "basic subscriber records" including name, length of service, credit card information, email address(es), and a recent login/logout IP address(es). For the next level of information, including message headers and some IP addresses, a court order is required"

Those are the "basic" records?

"For the real bounty -- messages, photos, videos, wall posts, and location information -- a search warrant is required."

Wondering why 'credit card information' is considered a basic record!