And how are they hitting that person with a tracking pixel and associating it with the pseudo-account?
It's still creepy, but it's likely more about determining knowledge of the friend graph (i.e. suggesting 2nd degree friends via a connecting pseudo-account) than about ad targeting.
Though yes, as soon as that pseudo-account could be tied to an actual account, Facebook could use passively gathered info to target you.
The creepiest part about Facebook is the sheer volume of facial data paired with social connection data. With access to that, even if you have never been online in your entire life, there's a good chance I could take your driver's license photo and know who you associate with.
> And how are they hitting that person with a tracking pixel and associating it with the pseudo-account?
That's the easy part. There are companies that specialize in this sort of thing, as well as in merging profiles from several devices (pc, tablet, smartphone), you can bet that if two-bit advertising technology companies know about these tricks that Facebook does too, and they probably know a trick or two that has not become mainstream yet.
If you want an explicit explanation of how such a link could be made it's a hard choice: too many possibilities.
I've done technical due diligence on about 10 advertising technology companies, node identification in a graph gets easier through two things:
- more known nodes in the graph (Facebook has many)
- more activity by the unknown node (just wait and track)
Sooner or later there is a moment where just for an instant that node can be strongly associated with a real world ID, for instance, a contact in someone's address book, a tag in a photograph and some shared online activity or something as simple as a phone call. At that point it is game over, the contact can now be associated with the device ID on the other side for instance through some running app.
Apropos games, many games monetize by embedding a library supplied by an advertising technology company that wishes to gain access to devices without waiting for the user to visit a website. These libraries leak information all over the net.
Is one example by FB, there are many more and some of those require permissions that make no sense at first sight until you realize what is happening under water. If you ever wondered why some shitty game requires access to your contacts, location and other interesting bits of data this is it.
It is very hard to stay off the radar of the likes of Google and Facebook, I have a pretty good idea of how this stuff works in the background and I have no clue how I could not leak enough bits for those two companies to tie my online activity to my real world identity in a single profile.
It's still creepy, but it's likely more about determining knowledge of the friend graph (i.e. suggesting 2nd degree friends via a connecting pseudo-account) than about ad targeting.
Though yes, as soon as that pseudo-account could be tied to an actual account, Facebook could use passively gathered info to target you.
The creepiest part about Facebook is the sheer volume of facial data paired with social connection data. With access to that, even if you have never been online in your entire life, there's a good chance I could take your driver's license photo and know who you associate with.