Statcounter is tracking browser usage roughly by percentage of requests. They do some corrections, for instance they try to account for Chrome's prefetching.
More people (or more computers) might use IE as their default browser, but if Chrome is responsible for more activity (requests), what's the more popular browser?
Does anyone know if the prefetching/prerendering Chrome does is a significant portion of web traffic? Because that could easily skew request-based analytics.
Headers wouldn't work, because a prefetch often turns into a page view.
Edit: their methodology is public. They use the aforementioned API, and headers for the two other browsers that prerender (Firefox and Safari, who rarely prerender because pages must explicitly configure it). Pages pre-rendered and discarded would add 1.3% to Chrome page views if they weren't discounted. Here is the StatCounter FAQ: http://gs.statcounter.com/faq#prerendering
It didn't seem to make any non-marginal effect on the graph. I wonder if they can accurately account for the pre-rendering.
Of course, pre-rendering might just be that good, that it's always hitting the right pages.
In the most general of ways, the browser more people use.
Which in this case, via Akamai stats, is clearly IE.
It depends.
Are you selling ads by pagehits? Then the former. Are you selling user info by number of users? Then the latter.
(And aside from ads and personal info, browsers don't make money in a direct way.)
Not entirely true. The Chrome Webstore makes some money, though admittedly not much.