Latest Stats as at 9/5/2011. ============================= These figures cover the 15-day period 16/4/2011 to 30/4/2011. ================================ ============================ Introduction ============ During this period, the hosting account got hacked and malicious iframe tags were inserted into many files. Out of 15 days of log files, only 6 days were any good, and 9 days of data were discarded as suspect. The huge drop in the sample size has destroyed the accuracy of the operating system and browser measurements for this period. On a lighter note, the log files were smaller for the days after the Royal Wedding. Either the surfers were sleeping off their post-celebration hangovers, or they were busy reading the "12 page Fairytale Wedding Souvenir" editions of the daily papers, or perhaps just watching endless repeats and frock analysis on TV. Operating System Market Share ============================== For the period 16th to 30th April 2011, these percentage figures are based on visits, with earlier figures in brackets:- XP 32.2 (37.4, 39.2, 40.0, 38.9, 39.9, 39.3, 38.2 ,38.2, 41.7, 42.5) Windows-7 27.6 (25.8, 26.9, 25.7, 24.8, 23.0, 22.5, 22.0, 22.6) Vista 14.4 (13.6, 13.2, 13.4, 13.6, 13.9, 15.5, 15.5, 15.9, 15.9) Other Windows 1.6 (1.9, 1.7, 1.7, 1.6, 1.8, 1.9, 2.3, 2.0, 2.4, 2.6) Mac 12.1 (11.1, 10.4, 10.5, 10.6, 10.7, 10.2, 10.6, 10.4, 10.6) iPhone etc 6.9 (5.5, 4.8, 5.1, 5.5, 5.4, 5.9, 6.3, 6.4, 4.2, 3.7) Android 1.4 (1.0, 0.8, 0.8, earlier share was in Other Mobiles) Other Mobiles 3.1 (2.9, 2.2, 2.3, 4.1, 4.4, 4.1, 4.2, 3.7, 2.7, 2.4) Linux 0.8 (0.8, 0.8, 0.6, 0.8, 0.9, 0.7, 0.8, 0.9, 1.0, 0.8) The total sample size above was only 5014 visits by humans. The above figures show a massive drop in Windows XP, a small drop in Other Windows, Linux steady and a rise in everything else. These figures are not realistic, especially since the latest results for the first six days of May are much nearer to earlier results. Browser Breakup =============== Percentages of visits for the period 16th to 30th April 2011:- IE 43.0 (49.1, 52.4, 49.5, 48.3, 48.0, 47.2, 48.9, 45.2, 52.1) FF 22.2 (20.5, 19.8, 20.6, 20.9, 22.4, 22.2, 21.1, 24.2, 22.2) Safari 16.8 (15.1, 13.9, 14.5, 14.6, 13.4, 14.2, 15.1, 14.7, 12.6) Chrome 13.9 (11.7, 10.9, 11.5, 11.5, 10.5, 11.7, 10.5, 10.3, 9.5) Opera 2.0 (1.8, 1.4, 2.2, 2.3, 2.7, 2.1, 1.7, 2.0, 1.7, 2.0, 1.8) All Others 2.0 (1.8, 1.6, 1.7, 2.3, 3.0, 2.6, 2.6, 3.5, 2.1, 1.8) Only six days of data in the figures for this fortnight, and the results are not consistent with previous figures. The large measured drop in IE is unlikely. The sample size was only 5111 visits by humans. The Safari result includes iPhones. Search Engines Share ==================== Percentage breakup of visits coming via search engines for April 16th to 30th, 2011:- Google 90.5 (89.5, 90.2, 91.4, 90.6, 89.7, 89.9, 89.8, 89.9) Bing 3.4 (4.3, 4.1, 3.5, 3.2, 4.0, 4.3, 3.3, 3.2, 4.0, 3.8) Yahoo 3.1 (3.2, 3.1, 2.7, 3.2, 3.4, 3.4, 3.6, 4.1, 4.0, 3.1) All others 3.0 (2.9, 2.5, 2.4, 2.8, 2.9, 2.5, 3.3, 2.8, 2.9) The sample size above was only 3232 searches. =============================== =============================== Search Engine Wars ================== This fortnight, with the hosting account being hacked, has led to an interesting discovery about the search engines. Please read on. Google has put in place technology to exclude hacked sites from its searches. It's pretty brutal. If a site, as scanned by Google, is found to be hacked in any way, the entire site is excluded. About 80% of my searches involved sites where the index.htm files had had malicious iframe tags added. Google excluded all pictures from these sites in its image search results even though none of the picture files were faulty. Any page at all from these sites produced a note saying "this site may harm your computer" and the intermediate message screen had no link forward to the suspect page. In other words traffic via Google searches dropped to about 20% of the previous level within a single hour. In my first analysis I could see the overall traffic had dropped (it was pretty obvious) but I did not realise the cause was that searches via Google had disappeared. Meanwhile searches from Bing and Yahoo continued as normal, as they do not (yet) have the ability to automatically ban hacked sites. Put bluntly, Google is vastly superior for this reason alone. As Google claims, they put their users first. Quite by accident, I analysed the search engine breakup for the whole 15 days including the bad days. The figures were Google 80.5%, Bing 7.2%, Yahoo 6.2% and Others 6.2%. These are the sorts of results we see published all the time by internet analysis companies who take huge samples of the network. Yikes, if there are any hacked sites out there contributing to these statistics, the Google results will be lower! Oh dear, if Microsoft and Yahoo implement security checking to remove hacked sites, their measured market share will drop. Getting a hacked site back on the air is not easy, but Google provides a sensible list of steps to follow. The final step appears to be that you must request a "review" using Google Webmaster Tools. So I had to join GWT, learn what it did and how to drive it, fix all the hacking and then request this automated review. Once I requested a review, everything went back to normal within an hour as if nothing had ever happened. It does appear that Google's crawlers continued to visit while the site was hacked, but did not respond to the faulty files being fixed. Right now it seems that the only way to get a hacked site back into Google is to join Google Webmaster Tools and request a review for each hacked site. =================================== ===============================