This was first intended as a reply to the Firefox developer mailing list, but then I decided to post it here first.
Personally, there have so far been two gripes with Firefox that I've taken some issue with:
* One was the disabling of on-demand loading of pinned tabs after session restore (between versions 9–11);
* The other is the current brouhaha over Adobe's Flash crashing the plugin container process, which is really not the fault of Mozilla. (more below)
Sometimes it's not users leaving Firefox, but some of them starting to use Chrome as their very first browser. Well, Chrome coming around is a good thing, because this gives people more choice as to which browser they want to use, as Chrome and Firefox both possess unique and attractive features that meet their users' different needs.
The current situation with Flash crashing the plugin container in Firefox is coincidentally a good cause for moving to HTML5 audio and HTML5 video, specifically Ogg Vorbis, Ogg Theora, and WebM, which are free and especially license-free formats.
YouTube's work in converting most of its videos to WebM reduces the immediate requirement and sometimes unpleasant chore of installing Flash on Linux, thereby increasing adoption of Linux, as other sites will hopefully follow suit in adopting free formats.
If we exclude the Summer low and the current Flash issue, then the next reason behind a reduction in Firefox usage could be the choice of format in sites using HTML5 video — most users tend to choose the browser that plays back whatever their favoured media site offers, with variations (mobile/desktop) of Chrome being in a rather advantageous situation right now, as it has built-in support for Flash.
Yet the situation with HTML5 video seems to be split right now along the lines of which HTML5 codecs are supported by which groups of browsers: Safari and IE vs. Opera, Chrome, Firefox and its derivatives.
The choice of YouTube and DailyMotion to offer videos in license-free formats is highly commendable. Now, if YouTube could actually stream high-profile events using HTML5/WebM in addition to Flash...
Desktop to mobile/tablet
Yet another reason in reduction of Firefox market share could just as well be the transition of people's major computing devices from desktops (including notebooks) to hand-helds (smartphones, tablets), nearly all of which currently have WebKit as their main rendering engine (in the form of either Safari or Chrome). I do not know if there has been a separate browser market share comparison for just desktop computers, because I understand that general tallies have usually encompassed both desktop and mobile spaces, with mobile being the separate segment.
Ultimately, as Mozilla and then Firefox were introduced, it was hoped that the browser market would eventually take the shape that it of recent times has started to form (at least worldwide) — in that no one browser would completely rule the market to be in its singularity the one to hold back innovation, and the one to pose itself in unintended consequence a widespread vector for malicious attacks.
So, in conclusion, the situation in my humble and perhaps half-informed opinion, is quite a bit more mixed with regard to what may be the possible causes of Firefox browser market share reduction this Spring and Summer.