A website called Hacker Target has posted a roundup of cute graphs with usage statistics for WordPress themes, plugins, and hosting for the top 500,000 on the internet. The results are fun to look at, but you should consider them as an interesting note, with a big asterisk.
One reason for the asterisk is clear in their opening paragraph:
Results in the charts below are derived from examining the Alexa top 500K sites. WordPress sites were identified through discovery of
/wp-content/plugins/and https://cdn.poststatus.com/ or
/wp-content/themes/in the HTML source of the page.
As many folks will know, the existence of
wp-content is not a definitive identifier for anything, as it and other directories are easy to change. So including these in the search query of the source makes it difficult to use as a reliable source.
Additionally, some of the top websites in the world utilize WordPress in the backend, but the frontend is displayed completely independently. Mashable is a good example of this. Browsing the source code of Mashable doesn't revel WordPress at all.
Even further, many huge websites use WordPress in part of their application, but not necessarily on the home page. Big newspapers come to mind.
Regarding hosting, I have a feeling it's doing a good job of identifying the underlying hosting, but not really the “brands” we know. For instance, Unified Layer is the far-away leader in the list. But what is that? Ah, owned by EIG, the owner of Hostgator, Bluehost, and scores more. That pattern can be attributed to many of the technologies listed in the hosting section. Just because it says it runs on AWS doesn't mean it's not on Pagely, and because it says Rackspace doesn't mean it's not on Pressable, etc.
So, have fun looking at these stats. I did enjoy it. But don't get too excited about their accuracy.