I didn't see any indication as to their methodology for how they're calculating these stats. It's probably been said a zillion times before, but if it's JavaScript-based you can safely assume that they are incorrect as it's likely there is a large intersection between people browsing with an ad-blocker, private browsing, or with content blocker and those people using Firefox.
Even if the measure is somewhat flawed it's hard to argue that drop isn't related to a drop in the underlying number of users.
Mozilla is awesome, but growing Firefox is an uphill struggle.