New NetApplications Market Share figures
The marketshare figures have been missing while NetApplications rework them to weight them by country for a more accurate global picture. Explanation here.
What can anybody find that’s new and interesting here? At first glance:
Google global share is actually falling, purely because Baidu is rising fast in China.
Mac share is much smaller than the unweighted (US biassed) previous figures. iPod touch even more so.
In worldwide browser versions, 8 year old IE6 is still the market leader - (sorry, web designers everywhere).
Even country-adjusted, Symbian only generates half the traffic of iPhone. That’s after scaling back US (virtually zero share) and scaling in China and India, where Nokia has high share and iPhone low, and the phone is the only web access for many users.
Unfortunately mobile stats are now corralled off in a thousand dollar upgrade area, apart from Operating System Market Share, which leaks a little mobile data.[ Edited: 01 August 2009 05:56 AM by sleepytoo ]
From my Sunday piece on the Net Applications numbers:
“...despite their flaws, we tracked the net metrics firm?s reports because their sample size was relatively large ? some 160 million visits per month ? and because they offered regular snapshots of broad market trends. They revealed, for example, the rise of Firefox, the decline of Internet Explorer, the failure of Windows Vista to catch on. Their reports were consistent, dependable, and free.
Net Applications? reports are still free, at least some of them (mobile market share data now costs an arm and a leg). But their dependability ? and perhaps their credibility ? just took a huge hit.”
I don’t really see that the dependability of Net Applications stats has taken a hit. They are still are a consistent sample from month to month and year to year; but not a random sample. What’s taken a huge hit is the naive interpretations of the figures that were sometimes made in the past. They never were absolute indicators of market share, but the changes from month to month and year to year always were meaningful and could be used with other sources of data to make reliable inferences. The new adjustments make them much more usable. Even now, their sampling is clearly far from random in each country, but at least they have adjusted for sampling rates by country.[ Edited: 03 August 2009 06:56 AM by sleepytoo ]