I find it hard to believe that Apple didn't know. I would expect any device with a in-house app installed to be reporting metadata about said app back to Apple. Comparing the number of installs of a particuler app + the distribution of IP addresses/locations where this app was installed vs that of a normal in-house Facebook app would likely show a discrepency.
Reasons for the contrary (Apple could not have known) my premise is incorrect, the metadata is insufficient, Facebook or in-house apps in general are much more widely use than presumed and often deployed from a public facing servers resulting in a location distribution too similar to the actual US population distribution, etc.
Facebook has offices all over the world and employees testing with devices on home networks all over the world (not to mention developers giving builds to family and friends, potentially).
I would hardly expect Apple to be scrutinizing the install patterns of an enterprise app. Those resources would be better spent improving the App Store review process and TestFlight.
Reasons for the contrary (Apple could not have known) my premise is incorrect, the metadata is insufficient, Facebook or in-house apps in general are much more widely use than presumed and often deployed from a public facing servers resulting in a location distribution too similar to the actual US population distribution, etc.