Mozilla's Global Usage Share Continues to Rise
Monday September 30th, 2002
According to OneStat.com, the global usage share of Mozilla 1.0 has doubled to 0.8 percent, making it the Web's seventh most popular browser. Netscape 7.0 has also seen its usage share increase, rising from 0.3 percent to 0.5 percent. However, Microsoft Internet Explorer still dominates the browser market with a total global usage share of 94.9 percent.
Thanks to Martin for the news.
#1 The usage share of Gecko-based browsers
Monday September 30th, 2002 7:18 AM
... is 0.5% (Mozilla 1) + 0.8% (Netscape 7) + ? % (Netscape 6, Chimera et. al) > 1.3%.
#6 Re: The usage share of Gecko-based browsers
Monday September 30th, 2002 9:38 AM
Yep, according to Chuck Upsdell's Browser Stats at <http://www.upsdell.com/BrowserNews/stat.htm> Gecko-based browsers have a 1.5% usage share.
#12 more specific stats would be nice
Monday September 30th, 2002 10:56 AM
I have a nedstat counter on my own sites. For my weblog it indicates 14.7% marketshare for mozilla based browsers. This can be explained from the fact that my site is likely to attract the geeky kind of visitor due to my own background.
It would be nice to know what the figures are for categories of sites rather than just all websites all over the planet. I suspect that armed with this knowledge, many e-commerce sites, for instance, would be more eager to support Mozilla since the kind of people running Mozilla are early adopters and as such relatively active on the web. Just like only a few % of kazaa users contribute content, I suspect that only a relatively small group of active web users is driving ecommerce sales.
#2 Innacurate IE share stats due to UA spoofing . . .
Monday September 30th, 2002 7:59 AM
94.5% global share for IE?!? Yeah, right! Now, takeaway all those that are spoofing their browser user agent string, like any of those Opera users that haven't bothered to change the default UA settings in their browser, and you'll most likely see much different stats. If it weren't for people letting their non-IE browser spoof as a version of IE, then the overall market share for Microsoft's lame-assed browsers would probably be at least 5% less than current stats.
#3 Re: Innacurate IE share stats due to UA spoofing .
Monday September 30th, 2002 8:19 AM
Maybe the people gathering the statistics should check something other than the user agent string.
#11 "Something else" - Like what?
Monday September 30th, 2002 10:51 AM
Tanyel: "Maybe the people gathering the statistics should check something other than the user agent string."
And what would that "something else" be then?
#18 Re: Something else - Like what?
Monday September 30th, 2002 3:02 PM
#22 Re: Re: Something else - Like what?
Monday September 30th, 2002 9:20 PM
I wish you had an accurate way to determine the sex of the people you talk about.
#30 Client-side JS
Tuesday October 1st, 2002 8:45 AM
One way is to HTTP POST the user-agent data to the server. Another is to access a logger CGI from an IFRAME.
"One way is to HTTP POST the user-agent data to the server. Another is to access a logger CGI from an IFRAME."
#7 Re: Innacurate IE share stats due to UA spoofing . . .
Monday September 30th, 2002 9:39 AM
Opera is easily detectable no matter what it is set to identify as, it always appends "Opera" onto the end of the UA, so set to IE5 it sends "Mozilla/4.0 (compatible; MSIE 5.0; Windows 98) Opera 6.05 [en]".
Even if Opera isn't properly detected, that will reduce the IE share by all of half a percent.
#28 Easy to detect, but some analysers don't
Tuesday October 1st, 2002 4:42 AM
Last I checked the source code, the popular 'Analog' log analyser at least does not correctly detect Opera. (Can we say 'lame'?) I hope that has been fixed by now.
You're right that it makes bugger all difference in any case. I'd be pretty shocked if the 'real' share for IE turned out less than 85%, or the 'real' share for Gecko turned out more than 3%.
I don't know why anybody is bothering to dispute these numbers - even though they are probably inaccurate, there's no question that IE already won the browser war, and Mozilla will not (in the foreseeable future) make up anything more than a niche. On the desktop (and excluding its use in AOL), eventually reaching 5% share would be a great success for Gecko - 10% would be an astonishing victory.
Mozilla might have a larger role as a component in other systems. (Non-Windows desktops, set-top boxes, etc.)
#31 Mozilla may be more expensive for STBs than Opera
Tuesday October 1st, 2002 8:47 AM
Mozilla as a component? Please. Opera is more attractive to STB makers because the price of licensing Opera may be less than the price of the larger RAM bank, the faster CPU, and the fan fan for the faster CPU that Mozilla requires.
#4 Where do these stats come from?
by SubtleRebel <firstname.lastname@example.org>
Monday September 30th, 2002 8:30 AM
Who is using OneStat.com? Their analysis of "global" usage is based only on the data gathered from their client's websites. Who are those clients? If those clients are writing IE biased websites then it is likely to bias the statistics. Maybe Mozillazine.org and Mozilla.org ought to become clients of OneStat.com and see how that might affect their numbers.
#13 Re: Where do these stats come from?
Monday September 30th, 2002 11:50 AM
Right. Mycroft <http://mycroft.mozdev.org/> has usually between 80% and 95% of Mozilla-users, this can drop down to 75% when a new Mozilla release comes out and we get folks from Slashdot ;-) But we are the sites that can make a difference in the statistics. I mean it. I just registered, I hope lots of other Mozilla "biased" sites will do the same. Let's show them the power of the red dino ;-)
#15 Where and How to register my site?
Monday September 30th, 2002 12:06 PM
Good idea! Where and How can I register my site?
#26 Re: Where do these stats come from?
Monday September 30th, 2002 11:26 PM
And the reason to try to (on purpose) make the stats biased to show Mozilla usage would be....? What? So it gives an artificial good feeling when you think that Mozilla is used by more people than it is? Why not try to get as accurate stats as possible? That's not going to happen if you deliberately sign up shitloads of Mozilla-centric (such as Mozilla.org and MozilaZine.org) sites to the stat gathering site. All it's going to do is cause results that are no longer usable.
#35 Re: Re: Where do these stats come from?
by SubtleRebel <email@example.com>
Wednesday October 2nd, 2002 10:01 PM
My point was that there is no reason to believe that the current OneStat.com stats are accurate.
If the current stats are partially derived from IE biased sites then adding Mozilla biased sites would only help to make the overall results more accurate.
Just out of curiousity though, how are these stats ever going to be usable for anything? What good does it do to say that these stats show that Mozilla has 2% marketshare? Would it be useful if it indicated that Mozilla had 5% marketshare? The stats are going to have to show a lot more before they are able to have any useful purpose. Having inaccurate stats that indicate 20% marketshare for Mozilla is a lot more useful than innacurate stats that indicate that Mozilla has less than 1%.
#36 Re: Re: Re: Where do these stats come from?
Wednesday October 2nd, 2002 11:36 PM
"My point was that there is no reason to believe that the current OneStat.com stats are accurate.
If the current stats are partially derived from IE biased sites then adding Mozilla biased sites would only help to make the overall results more accurate."
Is there a reason to believe they are not? They sound very plausible and reflective of what I've seen "out there" myself.
"Just out of curiousity though, how are these stats ever going to be usable for anything?"
Tracking usage and growth as one metric (feedback is another) to see user satisfaction and perceived quality and value, to help determine what needs to be done in the project. For example, if feedback is all positive but there's no growth in usage, it could be that PR is the biggest problem.
But then again, I wasn't the one who posted the article. Perhaps you should ask Martin why it's important?
#37 Reasons to suspect inaccuracy
by SubtleRebel <firstname.lastname@example.org>
Thursday October 3rd, 2002 11:38 AM
"Is there a reason to believe they are not?"
Yes there are many reasons to believe they are not accurate. One reason is that there are millions of IE biased sites and it would be rather odd if none of them used OneStat.com, but there are very few Mozilla biased sites and to my knowledge none of them use OneStat.com to keep track of stats. Then there is the fact that not too long ago, the OneStat.com website itself is IE biased; Mozilla based browsers were not even able to navigate the site properly. Another reason is that these numbers do not reflect what I see in web log files for general purpose websites; for example, I help maintain a tourism website that gets several thousand visitors every day and it shows close to 7% of its visitors are using Mozilla based browsers and less than 2% using Netscape 4.x; off the top of my head I do not know the figures for the various versions of IE.
"Tracking usage and growth as one metric (feedback is another) to see user satisfaction and perceived quality and value, to help determine what needs to be done in the project. For example, if feedback is all positive but there's no growth in usage, it could be that PR is the biggest problem."
If the stats are inaccurate then they are useless as any kind of metric. If the stats are biased then they will never properly reflect growth. However, the biased stats can be used to convince some web developers that it is not necessary to write for W3C standards; likewise if the stats were skewed the other way then it would be more beneficial for helping to convince web developers to code for everyone. I suppose Martin posted the info to encourage the Mozilla faithful about the increasing marketshare, but 1% marketshare is not going to impress anyone else and thus the current report has very little real value to the Mozilla community.
#38 Re: Reasons to suspect inaccuracy
Thursday October 3rd, 2002 12:59 PM
"Yes there are many reasons to believe they are not accurate. One reason is that there are millions of IE biased sites and it would be rather odd if none of them used OneStat.com, but there are very few Mozilla biased sites and to my knowledge none of them use OneStat.com to keep track of stats."
So in other words, their sample is a very good approximation of the real world.
"If the stats are inaccurate then they are useless as any kind of metric. If the stats are biased then they will never properly reflect growth."
Yeah, these are my arguments against trying to bias it - pro Mozilla or pro IE. Any kind of inaccurate representation of the real world is bad.
"So in other words, their sample is a very good approximation of the real world."
In the real world, many websites are biased against Mozilla (intentionally or unintentiuonally) and so if you base your stats on their usage then your stats are biased.
Regardless of whether their sample of websites accurately reflect the type of websites on the web is irrelevant. They are claiming that their stats reflect the distribution and usage of web BROWSERS, not websites. When you base your statistsics on websites that deny proper access to people who do not use IE then you are ignoring the fact that these people exist.
In the real world there are more games for PCs than for Macs; for sake of example, lets say that there 85% of games are for PC only, 14% are available for Mac and PC, and 1% are for Mac only. So you take a sample of games that reflect this real world distribution and monitor how many PC users play these games and how many Mac users play these games. With this information can you accurately determine how many Mac users play games compared to how many PC users play games? No, of course not. Likewise you can not determine how many people are using Mozilla by sampling usage of IE only websites.
"Yeah, these are my arguments against trying to bias it - pro Mozilla or pro IE. Any kind of inaccurate representation of the real world is bad."
My point is that it is already biased pro IE. Adding in some Mozilla bias will help equal out the already existing IE bias and make the results more accurate. If there ends up being too much Mozilla bias then oops. I would prefer it were not biased at all, but I doubt that is a reasonable expectation and if the results are going to be biased then I'd prefer that it be a Mozilla bias rather than an IE bias.
Like some people have already said, browser statistics should be taken with a grain of salt, especially as the exact methods of generating the statistics (the sources of data, how the user-agent string is parsed to categorize the browsers etc...) are not known.
But ignoring these uncertainities for a moment, it's interesting to notice that in this newsletter, the total percentage of IE versions is 94.9%, in pressbox 7 ( <http://www.onestat.com/html/aboutus_pressbox7.html> ) 95.3% and in pressbox 4 ( <http://www.onestat.com/html/aboutus_pressbox4.html> ) 96.6%. Ignoring the uncertainities, this would mean a drop of 1.7%. Maybe Mozilla and Opera actually have managed to do something?
#9 where are the "ie's global share falling" stories?
Monday September 30th, 2002 9:42 AM
this is a definite noticable trend over the last 6 months, its cool to see the effect of mozilla and opera affecting the numbers.... now when are the major media outlets going to take notice of this, probably once the numbers grow a bit more...
#10 Where are the "ie's global share falling" stories
Monday September 30th, 2002 10:00 AM
It doesn't make a difference whether 97% or 80% of page hits are coming from IE. What matters is how many page hits are coming from Netscape 4 and Gecko-based browsers. If a significant number of hits are coming from Netscape 4, that holds back adoption of newer web standards. If only 5% of hits come from Gecko-based browsers, major web sites will be forced to make their sites comply with the latest standards.
If the "browser war" is defined as the browser with the vast majority of page hits, IE has won, and the war is over. If instead the "browser war" means that there is enough variety in the browser market that web designers and browser makers both need to follow the standards, the war is far from over, and in the end everyone can win.
I wonder who invented the term "browser war". I can't think of any other business where competition is referred to as "war".
The media has over the past years been very eager to announce a winner of this war, ignoring that competition is usually considered good and that the end of the war is actually bad.
Does anybody know about the history of this term?
#21 Re: Re: Re: result....:(
Monday September 30th, 2002 6:11 PM
I can only say that I've seen "The xxx wars" in business circles a few times before the Web boom. In the early 1970s, there was a one-upmanship battle between the big three mens's magazines, Playboy, Penthouse, and Hustler, as to who would show pubic hair or the most pubic hair in their layouts first. The press and media repeatedly called this "The pubic wars". And somewhere I have seen a mention of a historic war in biblical times I think that was called The something Wars where I can't remember what 'something' was, but it was close to the word 'pubic'. Hence the derivation of the Pubic Wars term.
#25 Re: Re: Re: Re: result....:(
Monday September 30th, 2002 11:12 PM
Do you mean "The Punic Wars" (wars between Rome and Carthage, <http://www.wsu.edu:8080/~dee/ROME/PUNICWAR.HTM> )? I would not say biblical times, but pagan times :-)
#8 Web Browsers Used To Access Google
by filip <email@example.com>
Monday September 30th, 2002 9:41 AM
There is a little graphic showing the web browser used to access Google. And the 'other' categorie take a small increase in the last month... Peharps it's the mozilla effect :)
(Your language preference must be set to englih content). Filip
More stats per site can be found on <http://www.nedstatbasic.net/> . Klick on the little graph icon at the right of each site, and then top left "With What" or in whatever language it appears. But Moz counts as Netscape 6 - or 7 perhaps? Strange how it varies - one pron site still has 1% Netscape 3 on 100K views, but 2.3% too for NS6/7/Moz. A US music site has to make do with only 0.3% Moz-based viewers. For F1-racing Finland, 'our' score is 4%. Even Konqueror appears at 0.3% in the land of Linus.
#24 How Mozilla counts on Nedstat
Monday September 30th, 2002 10:12 PM
Mozilla counts as Netscape 6 Netscape 6 counts as Netscape 6 Netscape 7 counts as Netscape 7
sorry, folks, but it seems very likely that any supposed increase in global mozilla usage lacks any form of statistical significance, and likely falls well inside any definable margin of error.
and of course, the real-world significance is equally negligible. consider the fact that in the 2000 election, the percentage of the popular vote gained by the third-party candidates *not* including ralph nader was greater than the percentage of browser use that mozilla can proudly claim.
-a proud supporter of mozilla and harry brown (lib.) in 2000
<http://extremetracking.com/open;sys?login=arc154> <http://extremetracking.com/open;sys?login=aurel31> <http://extremetracking.com/open;sys?login=blaker12> <http://extremetracking.com/open;sys?login=guildftp> <http://extremetracking.com/open;sys?login=knoda> <http://extremetracking.com/open;sys?login=mozdave> <http://extremetracking.com/open;sys?login=mxweblog> <http://extremetracking.com/open;sys?login=nvgdsm> <http://extremetracking.com/open;sys?login=ocautemp> <http://extremetracking.com/open;sys?login=oceanfrg> <http://extremetracking.com/open;sys?login=rarlab> <http://extremetracking.com/open;sys?login=sreactor> <http://extremetracking.com/open;sys?login=taming> <http://extremetracking.com/open;sys?login=tarqwak> <http://extremetracking.com/open;sys?login=viahw>
For example sreactor is gone from 2.19% in 2002.08.12 to 2.36% today Mozilla is counted under "Netscape 6"
#20 Red Hat 8 should help
Monday September 30th, 2002 4:53 PM
Because Moz 1.0 is the default browser in Red Hat 8, I think once RH8 starts getting some user share from previous versions, which defaulted to NS 4.x (ICK), that Mozilla's share will definitely benefit.
I would imagine that many Linux users already use Mozilla, but in any case it doesn't really matter; Linux has a miniscule market share on desktop machines (it's probably doing quite well on servers, but people don't browse from servers sort of by definition). If you look at the Google page somebody else linked, they include OS stats too; Linux is 1%.
Of course, all the niche operating systems add up, and Mozilla is the only serious browser on most of them. So they do help Mozilla's share (and I think there's probably particular room for Gecko on the Mac). But if Mozilla/Gecko is ever to improve its share by a really significant amount, it's the 93% of Windows users that need to be convinced to bother installing an alternate browser.
Tuesday October 1st, 2002 7:27 PM
Here are the results from TheCounter.com (<http://www.thecounter.com…/2002/October/browser.php>)
Note that Mozilla and Netscape 7.0 are both counted as Netscape 5.x
1. MSIE 5.x 5332148 (48%) 2. MSIE 6.x 4657320 (42%) 3. Netscape 4.x 335986 (3%) 4. MSIE 4.x 259060 (2%) 5. Netscape comp. 126894 (1%) 6. Netscape 5.x 96931 (0%) 7. Opera x.x 84898 (0%) 8. Netscape 6.x 57497 (0%) 9. Unknown 26700 (0%)