How much plausible deniability of all of this stuff that can be happily plugged into the Google, etc, APIs and tools should we extend to the big players? I'm sure there are plenty of people at all those companies who know that the data connection integrations they provide aren't only getting first-party, originally-sourced data. Google's ad platform doesn't need to explicitly get their hands dirty tying all the threads together for you or maintain a singular massive database of everything, they just need to supply enough hooks to let all the OTHER companies do it. Which is probably good in that everyone's ad-hoc attempt is probably less-accurate than Google could do on their own... but all that data is still floating around and it all originally got connected to use to target ads in these systems.
EDIT: And if you extend it to the publishing arms (e.g. Youtube or MSN or whatever) than I'd bet many of the big players in ad serving have other departments that are running integrations with some of those data aggregation platforms. They know how the game works for sure.