As the world goes mobile, that by itself is disquieting for Mozilla.
The current mobile situation is:
* iPhone: can't run Firefox there at all, really. There's Firefox for iOS based on WebKit like everything else on iOS, but I don't even know that StatCounter would count it as "Firefox".
* Android: Google for a long time had (and in many cases continues to have, as far as I know) agreements with OEMs that forbid a default browser other than Chrome. In some cases those agreements forbid preinstalling a non-Chrome browser at all, even as non-default. So you have to rely on people downloading an extra browser from the app store, and people don't do that much.
In addition to all that, "mobile" sites tended to be written to "WebKit", not "standards", for a while for various historical reasons. Unfortunately, at this point that's pretty deeply ingrained and people continue to do that, albeit with "Blink or WebKit" replacing "WebKit", which entrenches the problem.
* it could come with an alternative app store
* maps can be used as a web app (how is that different from Firefox OS?)
But yes it would look more like Amazon's offering than Google's. Look where microg is now, though. That is without Mozilla's backing.
Not as snappy or polished as chrome on android to be honest.
I personally switched to it because I felt I give enough data to Google as it is. Probably not a great reason, because they still get a lot of data from me.
Note that the denominator for market share as measured by StatCounter is "page views", not "browser users".
So for example, say you have 4.1 billion users, who use their laptop and phone evenly for their web browsing. On their phone, none of them use Firefox. All of them load the same number of pages every day. In this scenario, 4% of page views corresponds to 8% of users using Firefox on laptop.
Note that these are not realistic assumptions, of course: people vary widely in terms of how many pages they load, vary widely in terms of how their web activity is split between mobile and non-mobile, etc. For example, in 2010, I would bet that the average Firefox user used the web more than the average IE user, so the StatCounter number for Firefox was inflated compared to the actual fraction of users using Firefox.
If you want to measure actual number of Firefox users, you need to use a data service that tries to measure users instead of page views (i.e. not StatCounter).