Facebook has over a billion daily active users, so 100 million hours divided over 1 billion users is 0.1 hours/user which is 6 minutes per user. Seems reasonable. Of course, there are lots of people who don't watch any videos and on the flip side there are a lot of people who watch a lot of videos on facebook. Edit : as pointed out below, the autoplaying videos might skew the numbers quite a bit.
I'm not sure if there's any standards between platforms for these things that allow you to compare though. I'd say for example that you should exclude watches that last less than 5s or so. YouTube and Netflix may not have thought to do it because it doesn't make much sense to them but Facebook really needs to since I assume most of their video watches are automatic (accidental) while scrolling through the feed.
It does matter to Netflix. They don't publish their numbers and just use them for internal metrics so you can bet that they are honest with themselves about their numbers.
given a lot of auto-play video in my own feed, I'm presuming it's not all actually 'watched'. if they'd give a separate number on videos 'listened' to (where I actually unmute the audio), I'd take that number more seriously.
