Mozilla's Global Usage Share Now at 1.6 PercentSunday July 27th, 2003thelem writes: "OneStat.com have released their latest browser usage data. It's as expected, IE6 has made big gains at the expense of IE5, Netscape 4 has made losses and Mozilla and Safari have made gains." thelem adds that IE now has a global usage share of 95.4% (up 0.1 percentage points since February), Mozilla 1.6% (up 0.4 percentage points), Netscape Navigator 4.x 0.6% (down 0.4 percentage points), Opera 6.0 0.6% (down 0.1 percentage points) and Safari 0.25% (up 0.14 percentage points). "These figures show similar trends to those being reported by thecounter.com, although thecounter.com puts mozilla usage at around 2.2%." What about NS 6-7x? Netscape 6-7x is considered Mozilla, as well as all Gecko-based browsers. And that is how it should be (inclusion of all Gecko-based browsers in one statistic). A pitty they don't bother to "label" it as such: "Gecko-Based Browsers(Netscape6/7,Mozilla,...)". :-( That dosn't look like how they are counting it. Notice that they have Mozilla at 1.6% under the "popular browsers" and also 1.6% under the "total global usage". Then they have "Netscape Navigator 4" at 0.6% but in the "total global usage" they have a seperate category for Netscape Navigator (presumably all versions) that is at 2.5%. It looks like they left off Netscape 6/7 from the "popular browsers", which if the Global totals are correct means Netscape 6/7 should have around 1.9%. If that is not how they are counting it then they need to clarify. Since IE works only at Windoz platform, the above figures are correct, then it shows that more than 95.4% user is using windoz? It was available for Solaris for a while, too. Many mozilla users are on windows. Me for instance. Hopefully not many of them are on Windows Me. Hey! I'm using Mozilla on Windows ME! its a perfectly fine Operat... oh hold on, I have to reboot, I'll be right back. It's good to see Mozilla usage increasing. It bothers me to no end when people design sites, especially commercial sites, that only work in IE and justify it by saying well, 95% of the people use IE so that is good enough. Would these people would turn away 5 out of every 100 people in a brick-and-mortar store because of who they are? I guess so... but in the real world that could get you in trouble, even if you turned away 1 in every 1000. Oh, and they did leave out NS 6/7 and make it sound like Netscape is losing popularity just because the ancient NS 4 is disappearing. Doing the math: Total Netscape of 2.5% - Netscape 4 of 0.6% = Netscape 6/7 of 1.9% And adding that to Mozilla, at least 3.5% are Mozilla based. I don't know how they are counting other Mozilla based browsers. When you consider how many internet using people there are, that is a lot of people using Mozilla. er... not quite. As the post above mentions, "Mozilla" in these stats already includes Netscape 6/7. So you have 1.6% for Mozilla and Netscape 6/7 combined. 2.5% total for Netscape would be 1.6% of Mozilla/NS6/7, 0.6% Netscape 4, and 0.3% presumably for earlier or unidentified Netscapes (although given rounding errors, that could just be 0.2%). Based on previous stats collections I've seen, that's probably 1.4% of Netscape 6/7 and 0.2% of actual Mozilla (that is, 90% of Gecko-based is Netscape 6/7), but it would be interesting to see current stats on that. Let me preface this by saying that I understand that a monopoly in the browser market would be disastrous for any kind of standards implementation. But when it comes down to sheer economics, if a site builder can *better* serve 95% of his/her audience by quietly sacrificing the remaining 5%, then many site builders are going to do it. And in the cold, hard light of business/money-making who would blame them? As long as the percentage of build time required for cross-browser implementation is higher than the actual market share using that browser, you're going to have people drawing the line close to their main market. And remember that cross browser implementation does not just mean get it working in Mozilla. It also means Opera, Konqueror and Safari. So that 5% time remainder gets broken up into small, but very complicated pieces. All that said, I hope the W3C specs get more specific, and that underdog browser manufacturers get it right and have their day to shine. Let me preface this by saying that I understand that a monopoly in the browser market would be disastrous for any kind of standards implementation. But when it comes down to sheer economics, if a site builder can *better* serve 95% of his/her audience by quietly sacrificing the remaining 5%, then many site builders are going to do it. And in the cold, hard light of business/money-making who would blame them? As long as the percentage of build time required for cross-browser implementation is higher than the actual market share using that browser, you're going to have people drawing the line close to their main market. And remember that cross browser implementation does not just mean get it working in Mozilla. It also means Opera, Konqueror and Safari. So that 5% time remainder gets broken up into small, but very complicated pieces. All that said, I hope the W3C specs get more specific, and that underdog browser manufacturers get it right and have their day to shine. First of all, there are few <strong>practical</strong> things that you can do in IE only that can't be painlessly implemented across platforms. This build time percentage difference is not as large as many people like to make it out to be. If the site is properly designed according to standards, it will gracefully degrade into any web browser. Developers just have to lose the attitude that your website will have to look precisely the same in all web browsers. The problem is that a lot of developers are either behind the curve or too lazy to figure out how to do these things. This handcuffs their operations when they get a large client who has a lot of Mac or UNIX systems. There are a gazillion reasons to not design exclusively for IE, and over the next few years, the people with the mentality outlined in the parent will learn them the hard way. It's called a false economy. As far as I can tell from the various standards redesigns which have gone on (e.g. devedge.netscape.com, wired.com etc...) the average page size drops and the page is easier for Google et al to index (I could have sworn I read something to that effect at Meyerweb.com anyway, can't find it now though...). To put it another way: If you design for the standards (XHTML,CSS) you design for all browsers, and by proper use of standards you save money on bandwidth (you need to spend less on pipes and web servers) and increase you business visibility (through better search engine categorising). If what I've said so far is accurate then those companies shoving out IE-specific designs are doing themselves a grave injury. N.B. I am changing this from an argument about supporting Mozilla/Opera/Safari vs IE to an argument about supporting standards - the browser support is then implied (pages written for standards will display in NS4.x and be perfectly usable, they just might not be pretty). Another argument: I find using Mozilla and DOM Inspector, JS Console and JS Debugger when designing pages speeds up the development process, and all the developers I've shown these tools to find them useful too. Faster development is money saved hence developing with IE only is a false economy. Admittedly I'm talking about coding HTML by hand, developers using WYSIWYG environments aren't going to worry about that sort of thing. Of course on Mozillazine we're preaching to the converted... I agree that standards compliance shoud be one of the high priorities in any site production. The benefits you've outlined are all extremely compelling reasons to switch to XHTML/CSS coding methods and hopefully site builders will start to understand this in the near future. "If you design for the standards (XHTML,CSS) you design for all browsers ..." But, and this issue comes up every time, coding to standards does not guarantee any sort of consistency in browsers. As an example, Netscape 7 has a bug in which LABEL objects will *disappear* from the screen if positioned absolutely. This obviously has a critical impact on the usability of forms. Now the question is no longer "Do we support standards?", rather "Do we support this browser and its bugs and therefore modify our coding practices?". Further down this path, the question arises "How much time do we spend customising our coding practices to suit browser A?". Given that time is not an unlimited resource for most production teams, you then have to weight up how much time you give to each of the other browsers and their bugs and you are going to opt for the greatest cost/benefit. This is COLD, HARD ECONOMY. It's fairly safe to say that most teams will test for the major browsers (touch wood), but with a 90% market share in one direction, unfortunately not all major browsers are equally "major". > only work in IE and justify it by saying well, 95% of the people use IE so that is good enough To fight this argument you can say "you just have to write it standard and will lose no audience. No 5%, none" > To fight this argument you can say "you just have to write it standard and will lose no audience. No 5%, none" Except if they use a browser that doesn't properly support the standards. Read my post above for the Netscape 7 label bug as an example. I think if you didn't bother to spend the time customising your code to accomodate browser idiosyncosies (not sure if that is spelt correctly) you'd end up with a form with no labels. See how many users stick around then. I'm not saying standards don't rock my world, for they most surely do, but implementing them in imperfect browsers takes time, and time costs money. Sooner or later a decision gets made, unfortunately, and usually the majority wins. Here are the numbers for my site (ms developer related). They are kind of odd in that I don't normally see this
Here are number from my site (ms developer related). They are sort of odd, in that I don't normally see such a high share of NS 4.x. But Nav 5.x (which is what mozilla reports to be) has been steady at 4-10% range over the last few months. Browser Percentage IE 6.x 64.36% IE 5.x 17.73% Navigator 4.x 6.22% Navigator 5.x 5.00% Other 4.96% Navigator 3.x 0.91% Opera 0.38% Navigator 2.x 0.24% IE 4.x 0.10% Navigator 6.x 0.08% Would you provide your site address and the country where you are located at? Also, sorry about picking on your previous thread, as I didn't notice "take 2." Who is "onestat"? Is an authorized way of counting? No. It seems they don't even know that Konqueror says "I'm IE" and also Opera does normally. Also Mozilla if you want to. So what are they counting? They are counting user agents across a whole number of sites. I'm not sure about Konq, but even when Opera is pretending to be IE, it still adds an "Opera" on the end, so it can be (and is by onestat) counted. But some people do spoof user agents, and different people visit different sites, so counting user-agents at a bunch of sites is always going to be just that, counting user-agents at a bunch of sites. If you're maintaining a website, you should do your own counting of user-agents to find out about the audience for your particular site. And if you're trying to work out which browser is most popular, you can take just about any statistic you like and turn it into a news story... It's not even the fact that people spoof user-agent strings. What about browsers that, say, listen to the HTTP specification wrt. the back button? The back button isn't supposed to hit the server again, but show exactly what the user last saw. I know for a fact that Internet Explorer goes back to the server (not sure about others). This means that any browser with this behaviour is bound to have numbers that are inflated by the lack of standards compliance. What about browsers that cache poorly? What about all the other ways in which hits don't correlate to page views? The basic design of HTTP doesn't allow for reliable statistics of this nature. Even if it did, it means nothing for any particular site owner, as demographics can vary widely between different sites, different countries, and so on. Don't put too much faith in these kinds of statistics. Usualy you count visitors to a site, if the same ip with the same user_agent has visited any page in the site within a defined period of time (say 15mins) the ip os counted as the same visitor. But that is flawed too, in different ways. For example, where I work we have 4 staff, all of us have the same IP and same user-agent. So if two of us visit a site at similar times, we'd only count once. ISPs with web proxies can have hundreds of users coming from the same IP. Conversely, what if one user views a few pages, goes away to make a coffee, comes back 16 mins later (assuming your count is using 15 mins) and clicks another link, they will be counted as two visitors. You can take it a step further and use cookies to track people, but then you lose the people that block cookies. There's no good way of counting... I wonder why are the figures on http://www.upsdell.com/BrowserNews/stat.htm so much higher? Forget about the stats at thecounter.com. Their numbers have not changed since mid May. Good figures, though the actual figures may be slightly more, however, its wont be much different. On my own web site for example, i get a LARGE number of Mobile Phones hitting the site, but Moz based browsers stands at 7%, Mobile Phones are at 6% with Ericssons at 5% and Nokia at 1%. But the point is, my web site is more "enthusiast" oriented. ANyway, back to my point, as a web developer, i dont see any issues coding for both browsers. If a web developer uses STRICT HTML 4.01, even IE6 will have "better" standards support. Although arguably Moz has the better support of the two, Most users will never realise that IE is not displaying the page as the designer designed (myself), and woudl not notice, unless they are using Moz, and if they are using moz, why would they be using IE anyway! Not knowing how to Code to stanards is not an excuse for ignoring browsers. And althoguh i am not ahppy abotu IE's overall support of certain things (PNG, etc), its more than adequete to code to one browser these days. I dont code to NS4 anyway (never did), and any "differences in IE and Moz, i can mask in the Stylesheets anyway. As for "users", well more and more people are seeing the benifits of Moz based browsers. My father was firmly in the IE camp, but got stung by popups sooo badly, and few times ended up with Bonzai Buddy, and Gator, unintentionally. IN the end, to help him. I installed and set up and configured Netscape 7.1, selling it for the pop up blocker. However, he got used to it, and when he discovered tabbed browsing, and other stuff, he TOTALLY hates going back to IE. Ppl will realise slowly. Site designers will realise that NORMAL people will be using Moz, not just geeks, and thigns will change. slowly but surely. One good example of a site which has changed, is SonyEricssons web site. I used to complain that it didnt support netscape browsers well. Not only have they fixed support, but the new "Sony Nav bar" takes on the look of the Moz/NS Modern Theme, if either is used. Ok it sucks if u have a different theme, but hey its a start! http://www.sonyericsson.com/ I know that all such statistics are inherently flawed, but a though that just came to mind. Would it be fair to say that thecounter etc would be mainly used on sites that are smaller general/IE/windows content, and not so much on LINUX related pages. I say this as AFAIK thecounter use an image counter(?) which would tend not to be used on more "geeky" LINUX related sites and also bigger site, who would instead use proper log analysis. This IMO will cause a drastic swing to counting only at "My Webpage that is 'best viewed in IE5.5.005 @ 1024*768 resolution'" sites, hence meaning that these results are worth less than nothing. ... usually don't look at sites of these people. onestat in itself has many large customers, like Lycos, Air France and Fortis. See http://www.onestat.com/html/aboutus.html . I just checked the statistics for my own site and I was pleased to see that mozilla + opera account for nearly 35% now. That is a large change and must be something of the last few months. I don't check my statistics that often. It used to be something less than 5%. Obviously I'm in an environment (university) where people are more likely to pick up Mozilla so it is by no means representative for the whole community. However this does mean I will be a bit more aggressive in using stuff that doesn't work properly in IE. I saw asa using some pretty nifty rounded borders for instance and I most definately am fed up with being unable to use position: fixed, png alpha channel, etc. Perhaps Mozilla should consider a UNDERDOG edition of it's browser in honour of that great comic tv show... Never underestimate your adversary hé! If Mozilla is a good as it really is then it will continue to make waves. Just like any good service, when you're really good, you don't need any publicity. When people post in this forum, I think that it shows an enormous interest. When was the last time you saw an IE forum with avid fans? Perhaps Mozilla should consider a UNDERDOG edition of it's browser in honour of that great comic tv show... Never underestimate your adversary hé! If Mozilla is a good as it really is then it will continue to make waves. Just like any good service, when you're really good, you don't need any publicity. When people post in this forum, I think that it shows an enormous interest. When was the last time you saw an IE forum with avid fans? Most people don't know about Moz or the benefits it can provide them. There's a great page that explains the benefits of Firebird: http://www.mozilla.org/products/firebird/why/ Unfortunately now one has to explain why one is advocating a 0.6 release, or why people should use this page to understand why they should download Moz 1.4, or ... which is difficult when I don't particularly understand the roadmap myself. If Mozilla's usage was at 1.2% in Feb. and is at 1.6% now, that's a 33% improvement over a 5 month period! Pretty good for something that wasn't really marketed or considered for end users by AOL and that most users don't even know about. With the Mozilla Foundation beginning a marketing campaign, our growth should accelerate. If we could just get universities and public schools to start installing Mozilla, that would get a lot of new users used to Mozilla. You can find monthly updated global browser usage on this french site www.estat.fr Go to http://www.estat.fr/actualites.php and check topic 'Derniers Panoramas'. Gecko familly has 2.57 %. They still don't the difference between the members of the growing familly ... In order for Mozilla to get a large market share, we really need AOL to release a new version of its client that is Gecko-based. Not only will that greatly increase Gecko's market share, it will also make sure that designers can't "code to IE," which will in turn increase Mozilla's share even more. MS Internet Explorer 56.3 % Mozilla 23.1 % Netscape 12.6 % Opera 2.8 % Konqueror 2.6 % Unknown 1.6 % Safari 0.1 % WinAmp (media player) 0.1 % Galeon 0.1 % Phoenix 0 % Others 0 % I dont have an exact comparison with previous months (updated analysis software) but the Moz % has been slowly creeping up from www.davidgoodwin.net We ran stats on two sites here recently. One is my site leafdigital.com (main traffic related to Windows software and *ahem* elf porn) and the other one was britgo.org (traffic related to players of a Japanese board game). Neither of these sites should show strong Mozilla bias. They're small sites - mine has a bit over 100,000 hits per month including image hits etc, britgo.org has a few more - but large enough to get reasonable statistics. And the interesting thing is, *both* our sites showed similar Gecko numbers at around 5-7%. We did the tests using a perl script I wrote (and Tim Hunt modified) which is very basic, but we know (and tested) that it accurately reports the major browsers - IE6, 5.5, 5.0, other, MacIE, Gecko, Opera (even if pretending to be IE), NN4. All other browser families are negligible, and yes we verified this by examining the remaining user agents. According to the trends, Gecko usage on my site is already nine times higher than NN4 usage, and five times higher than IE/Mac. What's more, Gecko will overtake IE 5.5 by about October - so any web developers who consider old pre-6 IE versions should definitely consider Gecko as well. In summary: make sure to get the real deal by running stats on your own site using a log analyser that actually works. Don't believe the bullshit 'global stats' counter sites. (For example, at least until recently even the venerable Analog did not correctly detect Opera...) --sam by 'get the real deal' I am saying you as a web developer need to run browser stats on your own site before making decisions based on browser usage. And if you're trying to convince a friend or colleague to test on mozilla then you need to run the stats on their actual site for them. don't swallow the global bullshit, the only real deal is the real stats you run on your real site (using a reliable analyser that actually understands the 'Gecko' concept). By the way, some other notes about my site; WinIE still held 90% of share on my site, but Gecko at 6% was clearly the number 2 browser engine. (Opera is #3, 3%). Also, all non-browser results (search robots etc.) are excluded from these percentage share counts, we verified that all significant undetected strings were robots and excluded them from the count intentionally. --sam |