In these types of discussions, you want to separate what is theoretically possible in certain instances in a lab from what is practicable from a business standpoint. One need only look back a couple of years to, I think, America Online, which inadvertently released some search data. It was such a large file that a reporter was able to identify a number of individuals from a list that numbered, I think, in the millions. That was with a fair amount of work. Since then, I think the industry has taken additional steps to make it even more difficult to potentially identify a particular person. In technology, I suppose one can never say it's impossible, but in the case of BlueKai and the way our data systems are set up, I think we're very close.
On December 11th, 2012. See this statement in context.