On the first day of each month, the good folks at Net Applications release their NetMarketShare statistics, which measure worldwide usage share of desktop and mobile operating systems and browsers.
This month’s numbers, covering August 2013, contain several eyebrow-raising blips:
With a worldwide installed base of nearly 1.5 billion PCs, each one of those moves represents a shift of tens of millions.
What happened? Did 40 million people suddenly wake up to the fact that support for Windows XP is ending in a matter of months and rush out to replace their outmoded machines with shiny new Windows 8 devices?
Sorry, no.
The real explanation might be more prosaic: Net Applications recently changed its formula for measuring usage. In an undated note labeled “Important methodology change,” the company explains:
This month we start deducting hidden pages from our usage share statistics. Hidden pages are pages that are rendered but never viewed by the user, therefore, they should not be included in usage share data. An example of a hidden page is a page that loads in a background tab upon the launch of the browser and is never made visible.
That note appears to have been added to the site in July, although it's not clear from the note when the new methodology took effect. The change in methodology is intended to help measure actual behavior by PC users rather than background activities performed by browser code. Previously, the company’s stats had excluded prerendered pages, which Chrome uses to load pages in the background as a Google user types a search request. In February 2012, prerendered pages accounted for 4.8 percent of Chrome pages. In June 2013, that number was over 12 percent.
This month’s change goes still further. In its FAQ page, Net Applications notes the impact of hidden pages on the raw page counts from different browser families. These figures were measured in June 2013:
Hidden pages by browser family
If you factor those methodology changes in and assume all other factors are equal, both Firefox and Opera should have seen greater drops in usage than Chrome. Yet the usage share for each of those browsers was up, with only Chrome showing a drop. It will take a few more months of data to confirm whether this is a trend or just a blip.
That methodology change also suggests one possible reason for the apparent sudden drop in Windows XP’s usage. Windows XP users, blocked from using the most recent versions of Internet Explorer, are more likely to use an alternative browser such as Chrome or Firefox, both of which use hidden pages. Windows 8 users, by contrast, are more likely to use the standards-compliant Internet Explorer 10 (and IE 11 in Windows 8.1).
A few more nuggets worth noting from this month’s report:
All of these numbers need to be placed in the context of a shifting market, of course. Worldwide, consumers and businesses are buying fewer traditional desktop and portable PCs and spending more on tablets and other mobile devices. But with an installed base of 1.5 billion or so PCs it's still a force to be reckoned with.