Of the five companies, Yahoo appears to have been the favorite of NSA analysts. The Sunnyvale company received national security requests for "content" from at least 30,000 users. Microsoft fielded requests affecting at least 15,000 accounts, while Google requests affect 9,000 user accounts. Facebook said that its requests affect at least 5000 users. LinkedIn said it received fewer than 250 national security requests.
The companies may also have received requests for "non-content" information, such as data about the times, senders, and recipients of e-mail messages. Yahoo, Google, Facebook, and Microsoft all said they received fewer than 1,000 of these in the first half of 2013.
The data come with several caveats. The companies are barred from disclosing the precise numbers of national security requests or users affected, and must disclose the figures in ranges of 1,000 instead. The government also requires a six-month delay before reporting figures about national security surveillance, which is why Monday's reports only cover the first half of 2013.
Some companies took the opportunity to release several years worth of data. Google's data, for example, shows that national security requests affected fewer than 3,000 user accounts in the first half of 2009. By the second half of 2012, the figure had soared to more than 12,000 accounts. Data requests to Microsoft affected at least 11,000 user accounts in late 2011, a figure that grew to 15,000 for the first half of 2013.
The companies emphasize that the fact that the government requests information about an account doesn't necessarily mean that the request was granted. Microsoft, for example, said that it has "successfully challenged requests in court, and we will continue to contest orders that we believe lack legal validity. "
Several of the companies said they plan to continue pushing for greater transparency. Google, which has long led the pack on this kind of transparency effort, said it plans to push for the right to disclose "the precise numbers and types of requests we receive."
But as my colleague Andrea Peterson wrote last week, it's not clear that this kind of piecemeal disclosure is the most effective way for the public to keep tabs on government's surveillance activities. It would be simpler for everyone if the government itself published comprehensive statistics about the types of requests it performed. That would not only save individual tech companies the trouble of compiling reports, it would also give the public greater confidence that they were getting the whole picture.