22nd November 2004 | 21:34
posted by: blufive in stats_weenie
OneStat.com have done another of their occasional press releases giving details of their stats, and this one shows some dramatic numbers.
Just as a reminder, here's a table I put together in June 2004, showing details from the last three press releases.
|Browser||28 May 2004||19 January 2004||28 July 2003|
*The July 2003 figure is for Opera 6.x
Now, here are the numbers from their latest release:
|Browser||22 November 2004|
*OneStat actually gave separate figures for Mozilla Firefox "0.1" (2.8%, presumably they actually meant 0.10, which was the internal code number of the 1.0 Preview Release) and Mozilla 1.0 (1.8%) - I've combined them into one total for this table.
IE6 has rocketed from 69% to 81% in 5 months. Compared to the previous 11 months' growth (66% to 69%) that is just incredible. It's not all good news for Microsoft though - IE5 and 5.5 have fallen from 24% to 8% in the same period. Again, on past trends, that is a huge fall.
For every 2% gained by IE6, IE5.x has lost 3%. The main benficiary of the discrepancy is Firefox. It's not clear whether the May 2004 figures included Firefox in the "Mozilla" number or not. In any case, these figures suggest that the total usage of mozilla-based browsers has at least doubled, maybe even tripled, in the last 5 months. It's not all Mozilla, either - Opera and Safari have both gained ground, too.
To be blunt, I'm doubtful about the accuracy of these numbers. The gains made by Mozilla-based browsers are just plausible (especially with all the press attention they've been receiving lately) but the changes in usage for IE6/5.5/5 are so vast (17% of all IE users upgrading or changing to alternative browsers in the last 5 months?) that I have to suspect there is some systemic factor involved.[update]
Most other sources show trends in these directions, but nothing on this scale. Considering figures for IE only, OneStat's last set of figures was lagging behind the trend seen elsewhere - showing a much higher ratio of IE5 to IE6 than other sources. Now, it has leapfrogged most other mass-market sources, and shows a lower IE5/IE6 ratio than they do.
I can see a couple of likely explanations: either OneStat has changed their browser detection/counting methodology, or their data sample has undergone a substantial demographic change in the last 5 months.
So, while these figures are good news for Firefox fans, they are undoubtedly compromised by the question marks over the IE figures.