> Are you sure? Google has a search and location data, Google email is on one side if not both of most email chains, receipts, etc.
Absolutely not discounting all of that; there are some pretty strong parallels between the data you can infer on a person based on their gmail graph versus their fb messenger graph, for example.
Meta has a lot more "self-revealed-preference" data, though. Which of $thesePosts did you engage with? Does this pattern continue if we mix in $someOtherVariable? How long did you dwell on that one post before like/dis-like? Are you more likely to come back and spend time here if we tell you that $thisPerson has commented on your post instead of $thisOtherPerson? ... etc.
I think only YouTube serves as a plausible source of dwell time in the "at what part of $thisVideo did you click the like button" sense. If you don't use YouTube or use it signed out then it's (slightly) harder for google to attribute your actions to you. Facebook doesn't really have _anything_ that can be accessed without logging in.
Depends on what part of EU you are in. From older HN threads I learned that in some countries people still pay for sending text messages, so I guess that makes it more likely that some app can become popular there.
WhatsApp also demands full access to your address book, this means that Meta has a nearly global FoaF graph for anyone that has ever installed WhatsApp.
Between Google Search, AdSense, Google Analytics, Google Tag Manager, and various other APIs, Google has all the pieces in place to get a pretty good overview of what users are doing online. (And that's not even considering that the user's likely to be logged into a Google browser that's uploading their browsing history wholesale.)
Legally speaking, Google's not supposed to be correlating most of this data. But, as we know well, that doesn't mean they aren't doing it.
Absolutely not discounting all of that; there are some pretty strong parallels between the data you can infer on a person based on their gmail graph versus their fb messenger graph, for example.
Meta has a lot more "self-revealed-preference" data, though. Which of $thesePosts did you engage with? Does this pattern continue if we mix in $someOtherVariable? How long did you dwell on that one post before like/dis-like? Are you more likely to come back and spend time here if we tell you that $thisPerson has commented on your post instead of $thisOtherPerson? ... etc.
I think only YouTube serves as a plausible source of dwell time in the "at what part of $thisVideo did you click the like button" sense. If you don't use YouTube or use it signed out then it's (slightly) harder for google to attribute your actions to you. Facebook doesn't really have _anything_ that can be accessed without logging in.