OneStat.com Reports Mozilla Usage Share at 1.8%, Heise.de Notes Sharp Increase in Gecko VisitorsTuesday January 20th, 2004kroxx sent us a link to OneStat.com's latest browser usage share figures, which report that Mozilla has a 1.8 percent global usage share. This makes it the fourth most popular browser after Microsoft Internet Explorer versions 6, 5.5 and 5. In total, IE has a global usage share of 94.8 percent, according to the Amsterdam-based website statistics firm. Meanwhile, Gunnar writes: "Heise.de, one of Germany's major computer-related websites, reports that Gecko browsers had a share of almost 30% (29.9%) in January 2004 among their visitors. This is up 10.5% in a year (or a 50% increase)." An English translation of the Heise article is available. Given the total number of internet web users, 1.8% or even anything close to that is a lot of Mozilla users. On most sites I random browse to that have viewable stat trackers the Mozilla/NS pecentage is anywhere from 1 to 5% - and these are just on typical non tech sites. It still boggles my mind that people would design web sites that only work with IE. Turning away even 1% of your customers because of their choice in browser is rock stupid. I can imagine some PHB thinking that viewable by "95%" of the people is good enough without realizing how many people that represents. (yea, 1.8% of browser hits may not exactly match to 1.8% of everybody but would have to be close and just as meaningful) BTW, I guess Onestat is including NS 6/7 as Mozilla? Would you pay double the price for your web site to reach an extra two percent of visitors? You and I might, but it doesn't necessarily make sense economically. Albano If you design your site according to standards, you never have to redo it. If you design an IE-only site, you may have to redo it according to standards in the future if many more people start using Mozilla, Safari, Opera, etc. It seems that with IE-only design you run the risk of having to pay more and reach fewer customers. What kind of stupid business decision is that? By the way, OneStat's estimates of non-IE browser usage is far lower than other stat sites. See http://upsdell.com/BrowserNews/stat_trends.htm for some higher estimates. Additionally, many IE-only sites work only when JavaScript is enabled, meaning that a further 15% of users cannot access the site. As far as I can tell, IE-only design is almost never the result of a deliberate business decision. It's the result of ignorant web designers who think that everyone uses IE and everyone will continue to use IE. It's an assumption that's getting further from reality each day. #25 it doesnt cost double to make cross platform sitesby joschi Wednesday January 21st, 2004 11:47 AM if you hire smart developers who write standards compliant, cross browser code from the start supporting other browsers and platforms comes cheap. "Would you pay double the price for your web site to reach an extra two percent of visitors? You and I might, but it doesn't necessarily make sense economically." You completely missed the point. If you plan on having a lot of customers then 2% can be a lot of people and a lot of money! Even paying double would usually be worth it! And I would seriously question anyone who would charge *DOUBLE* to make a web site that could be viewed in IE and Mozilla. That would imply they are writing it twice which means they are doing something very wrong (either trashy code or just pocketing your money). The document is dated January 19 2003, should it say 2004 or is it one year old? Looks like a typo. It is currently linked from their home page as "latest news". They also say the stats are "since July 2003" so unless they had a time machine they probably meant an article date of 2004. OneStat's percentages do not total 100% and it is not clear that they handle all of the different versions of Gecko properly. Although they report IE versions with some degree of granularity they only report a single "Mozilla" number. It is not clear to me that they report Netscape 6.x, Netscape 7.x, CompuServe 7.x, AOL for Mac OS X, Galeon, Kmeleon, Epiphany, Phoenix/Firebird, Camino or any number of other browsers based upon Gecko. It would be interesting if someone would get one of their "lite" accounts and post the results of the actual web server log vs. their reporting of the browser statistics. We could then visit the page with a variety of user agents and determine if (and how much) they under report total Gecko usage. Personally I believe you can add at least 1% to their so called statistics. I have OneStat on mycroft http://mycroft.mozdev.org/ I don't see the exact logs, but mozdev folks probably have them. Right now they look like that (January 2004) 1. Mozilla 128,937 96.85% 2. Microsoft Internet Explorer 1,902 1.43% 3. Netscape Navigator 1,630 1.22% 4. Opera 240 0.18% 5. Other 215 0.16% 6. Apple 172 0.13% 7. Search engine crawlers 41 0.03% Total 133,137 100.00% So for personal stats, they do reach a total of 100%. Question would be what "other" is, but on my site, it doesn't make 1% as you claim. :-P Ah yes, and you don't see the breakup in versions for IE or others. For that, you would need a paying account. I notice in their pressbox that they stop mentioning any browser as soon as it gets below the 0.5% mark. Exception here is safari, which they have been mentioning for the last 2 reports and only now it's appearing in the list with 0.48%. I've set up a "lite" counter at http://stats.bclary.com/ and will run some tests on different user agents etc to see what the reports say about alternative gecko user agents. I do have access to my raw logs and we can settle this once and for all. :-) This site is not yet visible due to DNS, but check back tomorrow. The site is now available and it appears that onestat does correctly report all of the Gecko user agents in the mozilla evangelism sidebar as Gecko and reports Opera with an IE ua as Opera. I take back my comment about onestat underreporting Gecko. Here are the initial logs and reports. There is some disagreement in the logs and the report. It does appear that Netscape 6.x/7.x are reported as Netscape Navigator and not as Mozilla. x.x.x.x - - [20/Jan/2004:23:28:42 -0800] "GET / HTTP/1.1" 200 - "-" "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7a) Gecko/20040114 Firebird/0.8.0+ (MozJF)" x.x.x.x - - [20/Jan/2004:23:28:42 -0800] "GET /favicon.ico HTTP/1.1" 404 - "-" "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7a) Gecko/20040114 Firebird/0.8.0+ (MozJF)" x.x.x.x - - [20/Jan/2004:23:29:06 -0800] "GET / HTTP/1.1" 200 - "http://www.mozillazine.org/talkback.html?article=4199" "Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.6b) Gecko/20031216 Firebird/0.7+" x.x.x.x - - [20/Jan/2004:23:29:06 -0800] "GET /favicon.ico HTTP/1.1" 404 - "-" "Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.6b) Gecko/20031216 Firebird/0.7+" x.x.x.x - - [20/Jan/2004:23:30:09 -0800] "GET / HTTP/1.1" 200 - "-" "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7a) Gecko/20040121" x.x.x.x - - [20/Jan/2004:23:30:46 -0800] "GET / HTTP/1.1" 200 1533 "-" "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7a) Gecko/20040121" x.x.x.x - - [20/Jan/2004:23:31:14 -0800] "GET / HTTP/1.1" 200 1533 "-" "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031007" x.x.x.x - - [20/Jan/2004:23:31:19 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/5.0 (Macintosh; U; PPC Mac OS X; en-US; rv:1.5) Gecko/20031007" x.x.x.x - - [20/Jan/2004:23:31:26 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.5) Gecko/20031007" x.x.x.x - - [20/Jan/2004:23:31:37 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.4) Gecko/20030624" x.x.x.x - - [20/Jan/2004:23:31:45 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/5.0 (Macintosh; U; PPC Mac OS X; en-US; rv:1.4) Gecko/20030624" x.x.x.x - - [20/Jan/2004:23:31:53 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030624" x.x.x.x - - [20/Jan/2004:23:32:01 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.4) Gecko/20030624 Netscape/7.1" x.x.x.x - - [20/Jan/2004:23:32:10 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.0.0) Gecko/20020628" x.x.x.x - - [20/Jan/2004:23:32:18 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/5.0 (Macintosh; U; PPC Mac OS X; en-US; rv:1.0.1) Gecko/20020730 AOL/7.0" x.x.x.x - - [20/Jan/2004:23:32:31 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.0.2) Gecko/20030208 Netscape/7.02" x.x.x.x - - [20/Jan/2004:23:32:34 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:0.9.4.1) Gecko/20020508 Netscape6/6.2.3" x.x.x.x - - [20/Jan/2004:23:32:45 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:0.9.4.2) Gecko/20020502 CS 2000 7.0/7.0" x.x.x.x - - [20/Jan/2004:23:32:57 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1)" x.x.x.x - - [20/Jan/2004:23:33:08 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/4.0 (compatible; MSIE 6.0; AOL 8.0; Windows NT 5.1)" x.x.x.x - - [20/Jan/2004:23:33:19 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/4.0 (compatible; MSIE 5.5; Windows 98)" x.x.x.x - - [20/Jan/2004:23:33:26 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/4.0 (compatible; MSIE 5.5; AOL 7.0; Windows NT 5.1)" x.x.x.x - - [20/Jan/2004:23:33:34 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/4.0 (compatible; MSIE 5.5; AOL 6.0; Windows NT 5.1)" x.x.x.x - - [20/Jan/2004:23:33:44 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/4.0 (compatible; MSIE 5.0; Windows 98)" x.x.x.x - - [20/Jan/2004:23:33:51 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/4.0 (compatible; MSIE 5.0; Mac_PowerPC)" x.x.x.x - - [20/Jan/2004:23:33:56 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/4.0 (compatible; MSIE 5.0; AOL 5.0; Windows 95)" x.x.x.x - - [20/Jan/2004:23:34:03 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/4.0 (compatible; MSIE 4.5; Mac_PowerPC)" x.x.x.x - - [20/Jan/2004:23:34:17 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/4.0 (compatible; MSIE 4.01; Windows 95)" x.x.x.x - - [20/Jan/2004:23:34:29 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/4.0 (compatible; MSIE 4.01; AOL 5.0; Mac_PPC)" x.x.x.x - - [20/Jan/2004:23:34:36 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/4.78 [en] (Win98; U)" x.x.x.x - - [20/Jan/2004:23:34:45 -0800] "GET / HTTP/1.1" 200 1532 "-" "Mozilla/4.78 (Macintosh; U; PPC)" x.x.x.x - - [20/Jan/2004:23:36:24 -0800] "GET / HTTP/1.1" 200 1533 "-" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1) Opera 7.23 [en]" x.x.x.x - - [20/Jan/2004:23:39:45 -0800] "GET / HTTP/1.1" 200 1533 "http://www.mozillazine.org/talkback.html?article=4199" "Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7a) Gecko/20031220 Firebird/0.7+" x.x.x.x - - [20/Jan/2004:23:39:51 -0800] "GET /favicon.ico HTTP/1.1" 404 - "-" "Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7a) Gecko/20031220 Firebird/0.7+" 1. Windows 17 56.67% 2. Apple Macintosh 9 30.00% 3. Linux 4 13.33% Total 30 100.00% 1. Mozilla 12 40.00% 2. Microsoft Internet Explorer 11 36.67% 3. Netscape Navigator 6 20.00% 4. Opera 1 3.33% Total 30 100.00% The last time they released statistics (http://www.onestat.com/html/aboutus_pressbox23.html ) they listed the manufacturers seperatly. These were their stats: BROWSERS: Mozilla 1.6% Netscape Navigator 4 0.6%. BROWSER MAKERS: Netscape Navigator 2.5% Mozilla 1.6% The Mozillas match up, but the Netscapes don't, suggesting that they have recorded 1.9% Netscape 6/7 usage that is not shown. Assuming this has not changed that would make market share 3.5%, which is more like what I have read elsewhere. I took a look at the Google Zeitgeist a little while back. This is what I came up with: http://www.grack.com/news/NewGoogleZeitgeist.html In short: MSIE: ~90% Mozilla: ~5% Internet Explorer (all versions) fell from 95.4% (July 2003 http://www.onestat.com/html/aboutus_pressbox23.html ) to 94.8% now. On the other hand, mozilla climbed from 1.6 then to 1.8 now, which is only 0.2 percent. So most of that 0.6 percent did not go to gecko browsers (seems safari took the majority, and opera and konqueror divided the rest). It's good news for web standards compliance, but it's not necessarily good news for mozilla. Still, one of the goals of the mozilla project was to get web designers to build for standards rather than for specific browsers, so indirectly I suppose it is good for mozilla. #13 OneStat = more or less worthless, but good at PRby leafdigital Wednesday January 21st, 2004 5:24 AM They track only sites using their stat counter, right? Past experience would suggest that this introduces extreme bias (probably toward amateur sites and away from professional ones). Frankly I would ignore their numbers. The best publicly available stats are from Google Zeitgeist (Google have a sufficiently huge and varied audience) which suggest Mozilla is a bit under 5%. I checked a couple of sites (not Mozilla or web technology related) using my own log analysis and these give similar results. Try your own non-browser/tech-related sites (i.e. those for which you have no reason to expect readers are more likely to be browser-savvy) and see what you get... my guess is that for most sites your numbers will be similar to Google's, although obviously there will also be wide variation. (Developers or managers of existing sites shouldn't be looking at numbers even from Google, never mind OneStat, when making browser support/testing decisions; they should be looking at the stats from their own site.) --sam #15 Re: OneStat = more or less worthless, but good at PRby janahan Wednesday January 21st, 2004 6:14 AM Not really in regards to google... The problem with Google is that it is pretty much the default search engine on Mozilla Based browsers (search from URL bar). Whereas IE defaults to MSN search. This is likely in itself that woudl skew results towards a mozilla bias. However, one can also argue that this shoudl give a greater than 5% share. That depends on if they are using page impressions (likely) or unique IPs. If you use the search bar, then you don't go to the google homepage, so mozilla has one less page impression per search. It is also likely you are better at searching the net, so won't look at as many pages of result or try combinations of keywords as IE users. However, mozilla users will probably make more searches than IE users and will probably surf the net for longer. BTW When telling other people to look at their own statistics, get them to make sure their site works in all browsers first. If it doesn't, then people using browsers it doesn't work for will get to the homepage and give up. #18 Re: OneStat = more or less worthless, but good atby jgraham Wednesday January 21st, 2004 7:26 AM Google stats are just as dubious as I've pointed out both here and the forums before. Although they probably have small errors from the cross section of the population which uses the site (i.e. google has a varied audience), there are many other possible systematic errors that would affect the results, depending largely on exactly what consitutes a 'hit' on google (although other factors, such as the ways tabs affect browsing, could also be important). In combination with other stats, they give a ballpark figure that Mozilla useage is <10% and probably <5%, but I'm very sceptical of any claims that are more accurate than that. It would be interesting to see Mozilla usage by region (e.g. Europe, North America, Asisa,...) . I have noticed that major German sites often report a relatively high Mozilla usage. Heise at 30% is a good example. Even better (since it's completey non-techie) is Spiegel, Germany's major news site, which reports a Mozilla share of somehwere around 20-25%. Slowly, but surely, we are chipping away at IE's share. This is encouraging. :) Mozilla usage at heise.de was actually up 10.5 percentage *points*, which is an increase of around 50%. (Ok, that *was* picky.) --sam You are, of course, correct. But then again, that's what I submitted. MOst have gotten lost somewhere ;-) #16 Re: OneStat = more or less worthless, but good atby corwin Wednesday January 21st, 2004 6:34 AM and one could argue that the most popular IE extension is the google toolbar :-) As eiseli said, the really interesting fact here is that MSIE has decreased in share. The Onestat story makes a big deal about the increase in MSIE 6, but ignores the fact that overall MSIE share decreased for probably the first time since they began cutting off Netscape's "air supply". The updated stats at http://stats.bclary.com/ show pretty conclusively that Onestat counts Netscape 6.x/7.x as part of Netscape Navigator and not as Mozilla and that Onestat under-reports "other" user agents. Looking at the code of onestat's detection is it clear they will not correctly handle Gecko user agents which are reporting themselves as MSIE, and will not handle Gecko user agents which block images from their domain. Considering that Netscape 7.x had over 25 million downloads as of July 2003, that international Netscape sites still list Netscape 6.x/7.x and many people do not upgrade as quickly as you might think, I believe you can without a doubt conclude that Onestat's statistics should report over 2% global Gecko usage and perhaps significantly more. I have always considered the "magic number" for Gecko's share to be 5%. Once web sites begin to see 5% of their users using Gecko, they will not be able to ignore Gecko support. Increasing real and perceived Gecko share however is problematic for several reasons. One problem is due to the number of different Gecko browsers. Onestat actually does a better job with most Gecko user agents than other reporting tools which fragment Gecko's share amongst the various distributions. Onestat's main fault is in its reporting of Netscape 6/7 as Netscape Navigator rather than Gecko. If you use a web site analytic program and it does not accurately report Gecko usage, complain to the author/vendor and ask them to do a better job. Another problem area can be found by looking at sites such as the newly redesigned MSNBC (especially after they moved their domain to MSN.com) where you can see many problems in Gecko. It is unfortunate that a so-called #1 News site would neglect non-MSIE browsers, but it is not surprising. Another problem is due to ad agencies such as DoubleClick which improperly create IFRAME embedded advertisements using Flash without WMODE for Gecko. The resulting degraded experience for Gecko users makes it look like a Gecko problem instead of a problem with DoubleClick's ad code. If you want to help Mozilla and its entire family of Gecko-based browsers you should encourage everyone you know to use Mozilla or a Gecko-based browser, make sure to not configure Mozilla in such a way as to under report Gecko with these types of browser detection schemes, constructively complain to sites which do not support Mozilla and Gecko. For example, do not simply complain that Flash Ads on MSNBC are broken. Instead figure out the basic problem and explain to them why their code is broken and how to fix it. Although they have not fixed the problem as of yet, they did respond positively to my constructive and detailed analysis of the problem with *their site and their ad supplier*. Be a good Mozilla Community citizen and get involved in Mozilla Tech Evangelism and contribute to Mozilla's success by effectively communicating with web sites about how to improve their support for Gecko in particular and all standards-compliant browsers in general. you have to cut off 2,2% from that statistik, since they count Safari as Gecko (i wrote them by mail), this will be fixed in the next version of the stats I have a small site and I now have almost 15% Gecko share. This is up from 1% last year. We are growing by leaps and bounds. Gecko is hungry for butterfly. OneStat does not mention Netscape in their recent statistics. As one can see in older statistics of this strange company, they do not count Netscape branded Gecko browsers as Mozilla, they count them as Netcsape (I can see it in my own Onestat statistics). And what's now the global Netscape market share? Nowhere? Nobody uses Netscape anymore? Or are they just Anti-Netscape? And more strange: They just recognized 12 months after its release Safari, but they still do not know the existance of Konqueror - they count it as Mozilla. Can a company, which claims to provide accurate and detailed web site statistics, be taken serious, when it obviously does not monitor the relevant market? |