The New, Need-Driven Browser Choice Model
It used to be relatively easy for people to pick which browser to use, and one of the reasons that Microsoft's Internet Explorer still has dominant browser share is simpy that it came bundled with millions of copies of Windows on new computers. Firefox, more than any other browser of the last several years, made it a bit more complicated to decide on a browser. If you placed value on its extensions, which are hugely useful, it made more sense to use than Internet Explorer. Now, though, we're moving toward a browser landscape where you can't make a good decision on what to use without considering what you do online.
Web app performance is all well and good, but what about those oh-so-useful Firefox extensions? I know I can't get rid of Firefox because of them. From AdBlock Plus to iMacros, I'm just very used to how my favorite extensions work, and, in particular, I must have iMacros to streamline tasks I do every day. This is a case where performance isn't everything, and the particular tasks I perform dictate the fact that even if I use other browsers, which I do, I can't dump Firefox.
As far as pure performance goes, even there the decision-making process isn't all that simple anymore. Just take a gander at Lifehacker's recent arrays of benchmark tests involving popular browsers. The open source browsers, Firefox and Chrome, do very well in several tests but they don't win at everything. Internet Explorer wins at a speed test involving launching the browser when eight tabs are open. Opera wins at cold launches, but not all launches.
And that's the bottom line. The right response to someone who asks "which browser should I use?" is to ask what the person does online.