Browser Stats For The BBC Homepage

Lord_Scrumptious writes "An interesting article titled 'The software used to access the BBC homepage' has recently been published on a blog by a BBC employee. It's all about the different browsers and operating systems accessing the BBC's homepage. The analysis is from a week of page requests in September 2005. Not surprisingly, Internet Explorer accounted for 85% of site visits, but Firefox had a very respectable 9.7% share. Even requests from Sony's handheld PSP device were recorded, but interestingly there's no mention of mobile phone devices."

Comments

Finally....

[odaen]

Finally some reliable website records which arn't off some obscure coding page. :)

Re:Finally....

[Bogtha]

No, it's still unreliable. You simply can't correlate traffic to visitors. That's not the way HTTP works. httpd log analysis can tell you many interesting things, but mainly concerning the load on the server. Any attempt to read more into it is based on assumptions that are not only wrong, but wrong by an unknowable amount.

This is true every time somebody posts some bullshit story about how Firefox has a growing portion of the market, and every time somebody posts some bullshit story about how Firefox has a shrinking portion of the market. Even something as simple as AOL tweaking their cache configs can throw off the numbers by a large amount. Sure, it might make you feel good to look at your access logs and see Firefox gaining 1% every month or so, but that doesn't change the fact that that 1% number (or whatever) only has a tentative link to reality.

If you want to know OS stats, browser stats, or anything like that, you need to conduct an actual survey, and not simple observation of HTTP traffic, because if you are doing the latter, you might as well make up your numbers based on your best guess, because it has just as good a chance of being as accurate.

Re:Finally....

[mwvdlee]

And such a survey, assuming it is voluntarily, would be reliable?

I dare bet that FireFox users are MUCH more willing to fill in a survey about webbrowser usage than MSIE users.

I would honestly trust traffic logs a lot more than a survey on this matter. Say 10% of people use FireFox, then log stats might show anything from 5 to 20%, in a survey, it might easily show as 50% simply because FireFox users typically care more about their browser, afterall, they took the time and effort to install a different browser/OS.

Re:Finally....

[searlea]

You make a good point, that cache config can affect the amount of traffic directly hitting your website, and therefore affects your logs.

However, given the headers returned by the BBC site, caches should NOT cache the HTML, as the headers say the content expires immediately:

Expires: Mon, 24 Oct 2005 11:57:59 GMT
Date: Mon, 24 Oct 2005 11:57:59 GMT
Content-Type: text/html
Server: Zeus/4.2
Cache-Control: max-age=0

So, the BBC figures may be more accurate than you think.

Re:Finally....

[Bogtha]

However, given the headers returned by the BBC site, caches should NOT cache the HTML, as the headers say the content expires immediately

Actually, it doesn't say that caches should not cache the resource, it says that caches should revalidate the resource before serving it again, IIRC.

Which BBC site are we talking about anyway? I'm getting completely different headers for www.bbc.co.uk:

HTTP/1.1 200 OK
Date: Mon, 24 Oct 2005 12:10:13 GMT
Server: Apache/2.0.54 (Unix)
Set-Cookie: [snip]
Set-Cookie: [snip]
Accept-Ranges: bytes
Transfer-Encoding: chunked
Content-Type: text/html

That's a cachable resource, and what's more, because there's no Expires header, it's particularly sensitive to public cache tweaking of the kind I just described. Given that you are describing headers from a completely different server, I suspect that you aren't looking at www.bbc.co.uk.

So, the BBC figures may be more accurate than you think.

Well no, even if what you said was true, the public cache was just one example of many. Consider an Internet Explorer user going to the homepage, clicking on a link, hitting the back button, clicking on another link, hitting the back button, and so on. Pretty typical behaviour. Now consider an Opera user doing the same. The BBC homepage will register multiple hits from the Internet Explorer user, but only one hit from the Opera user, because of the differences in the way they handle browsing history (Opera follows RFC 2616, Internet Explorer does not).

Re:Finally....

[Bogtha]

Say 10% of people use FireFox, then log stats might show anything from 5 to 20%, in a survey, it might easily show as 50%

Where did you get these figures from? Correct me if I'm wrong, but it sounds like you just made them up.

I dare bet that FireFox users are MUCH more willing to fill in a survey about webbrowser usage than MSIE users.

And if surveys were typically presenting in such a way, then you might have a point. But they are usually one question amongst dozens, and people usually complete surveys because there's a prize or payment of some kind, not because they want to evangelise whichever browser they use.

Re:Finally....

[searlea]

Which site? news.bbc.co.uk.

As for the number of hits created by IE vs Opera when browsing a site, that makes no difference to stats if you're counting visitors, not hits.

Re:Finally....

[Bert64]

Yeah, IE also rechecks cached pages every time instead of caching them until the site says they should expire.. That causes a lot more hits.. Mozilla and Opera don`t do that.. But i`m not sure if mozilla reloads a page from the server if you press back or not.

Re:Finally....

[Anonymous Coward]

Who the hell would pay attention to raw hits? I'm not sure if they did or not, since the artical is nuked, but I'm sure they tracked IP addresses combined with User-Agent strings within a certain window if they want to be taken seriously.

Re:Finally....

[Bogtha]

Which site? news.bbc.co.uk.

TFA is about the www.bbc.co.uk homepage.

As for the number of hits created by IE vs Opera when browsing a site, that makes no difference to stats if you're counting visitors, not hits.

There's no reliable way of counting visitors short of requiring registration and banning people from viewing the pages until they are logged in, and even then you have to deal with things like BugMeNot.

Re:Finally....

[searlea]

That's a solved problem, to get IE to use it's cache without checking the server, the webserver needs to be configured to set pre-check and post-check extensions to the Cache-Control header.

Re:Finally....

[Anonymous Coward]

Advertisements on the BBC website? I THINK NOT

Re:Finally....

[dan the person]

Hmm yes some reliable numbers from a website that every single firefox install by default hits every 5 minutes to update it's latest news live bookmark.

Re:Finally....

[Anonymous Coward]

" Also, a single page view accounts for many hits on the server, one for each file (images, frames, flash, ads, css, etc).
People who block them (ads, popus, etc) will make less hits."

Sorry, I don't quite follow you here. Ads (banners?) and popups are most usually syndicated content hosted on another server, belonging to another person. They won't make any hits in the host-site's logs, ad-block or no.

Not to diminish your main intended plank: yes, each page with images/media/stylesheets on the same server will make a number of hits, not just one. But then you can look at unique page impressions - how many times this or that file (typically a HTML document) alone was visited by different IP addresses.

@grandparent: It's incredibly crass to say server-logs are not valuable indicators - they're the best you can get if you know how to read them. Looking at the activity on a particular day is not very helpful, but looking at the properties attached to unique IPs on an ongoing basis (I think this is called a diachronic study) is considerably less error-prone than any voluntary Q&A approach. If you become aware of a confound like the AOL cache issue creeping in, you can always re-evaluate that data or do separate study to quantify the impact of that confound.

Trust me, I know a guy who does surveys door-to-door, and he admits a staggering percentage of his questionnaires are rubbish, the answers either 'steered' by him just to get any answer from disinterested respondents, or in some cases actually completed by him after the respondent gave up completely!

Of course that is a sample base of one survey-taker, and in the absence of a control sample of other survey-takers, my study cannot be considered empirically viable. Any survey-takers reading this are encouraged to join the debate at this point.

Re:Finally....

[jez9999]

Anyone have a mirror available? Mirrodot is rather useless as iy only caches the first page. Boy, this site got Slashdotted quickly.

Re:Finally....

[Bert64]

And how is this done?
Also, why should I (as a webserver admin) need to modify my server config to cater to buggy browser behaviour? Surely this is a browser bug that needs fixing rather than a hassle for every webserver admin in the world.

Re:Finally....

[mwvdlee]

Where did you get these figures from? Correct me if I'm wrong, but it sounds like you just made them up.
The first word in the sentence, "say" should have made it clear that it was an example only. Logical, since nobody has any true figures.

And if surveys were typically presenting in such a way, then you might have a point. But they are usually one question amongst dozens, and people usually complete surveys because there's a prize or payment of some kind, not because they want to evangelise whichever browser they use.
The problem is that any type of survey in which the participant has a choice in participating will be, by definition, biased. Paying people to do surveys may help to get a more diverse group (although a paying survey will mostly attract the poorer, considering the price vs. time equation involved) but it will also, by financial necessity, result in a smaller sample.

Re:Finally....

[julesh]

But i`m not sure if mozilla reloads a page from the server if you press back or not.

No it doesn't. Even if you've seen a more recent version of the same page, with the same URL, it will show the page you originally got if it is still in its cache (i.e. most of the time, unless the server requests otherwise).

Re:Finally....

[julesh]

There's no reliable way of counting visitors short of requiring registration and banning people from viewing the pages until they are logged in, and even then you have to deal with things like BugMeNot.

No, but the simple expedient of setting a cookie, then ignoring as a repeat visitor any hit that either has the cookie or is not specifying a cookie but is using the same user agent and coming from the same IP address within some time limit (e.g. 5 minutes) works in over 99% of cases, which is good enough for most purposes.

Re:Finally....

[jhalstead]

Do any sites actually filter out the spam traffic from their logs before performing an analysis?

I've found through scanning logs on the handful of sites I manage that all of the automated spam/hack traffic reports itself as IE 6. For many sites I suspect this is a major contributor to their IE traffic.

-Jerry

http://www.evconvert.com/

Re:Finally....

[Cerv]

Well the BBC doesn't carry advertising, so in this case that's less of a problem.

Re:Finally....

[arivanov]

Wrong. It is an obscure coding page. It is a page you hit only if you miss a link on a navigation menu after an overdose of ethanol solutions on a Friday night. Now the stats for http://news.bbc.co.uk/ and http://www.bbc.co.uk/weather that will be interesting. Especially the latter as it has cookies (if you set your hometown) so it can filter out down to unique visitors.

Re:Finally....

[rrhal]

Well -

You could put GUIDs into cookies. Offer some reason for people to want your cookie (like saving personal preferences). Record the GUID into the web logs for each page access.

I realize its not 100% but it works reasonably well. It works very well for your biggest users - and you can still mine statistics for your unGUIDed users.

Re:Finally....

[whitehatlurker]

It is a page you hit only if you miss a link on a navigation menu after an overdose of ethanol solutions on a Friday night.

Wow, and it got 32 million requests in a week (one Friday/Saturday?). Do the BBC web viewers have a bit of problem with ethanol? ;-)

Would this problem be reflected in the proportions of web browsers recorded? (Oh hell, mark me as a troll - is this why there are so many IE users and almost no Opera users? ;^)

Re:Finally....

[BandwidthHog]

Opera lets users set the user agent string to spoof various browsers, and at least for the past few years, if not forever, it’s defaulted to MSIE. So the percentage of Opera users shown is merely the subset of Opera users who hit the BBC website with their UA string set to Opera instead of the default.

Re:Finally....

[BasilBrush]

Consider an Internet Explorer user going to the homepage, clicking on a link, hitting the back button, clicking on another link, hitting the back button, and so on. Pretty typical behaviour. Now consider an Opera user doing the same. The BBC homepage will register multiple hits from the Internet Explorer user, but only one hit from the Opera user, because of the differences in the way they handle browsing history (Opera follows RFC 2616, Internet Explorer does not).

Not true, at least with the www.bbc.co.uk page on which these stats are based. I've just investigated with a HTTP request sniffer, and IE and Firefox do exactly the same as Opera in the scenario you describe. None of them request the page again when you hit the back button.

Re:Finally....

[BasilBrush]

Not relevant for this study. It only counts the request for www.bbc.co.uk home page, not every request to the server.

Re:Finally....

[BasilBrush]

"live bookmarks" are RSS. www.bbc.co.uk does not have RSS. And it wouldn't count even if it did, as this study is simply a count of accesses to www.bbc.co.uk default page.

Re:Finally....

[BasilBrush]

Pardon? You're saying www.bbc.co.uk home page is an obscure coding page? You didn't RTFA did you.

Re:Finally....

[whitehatlurker]

Opera lets users set the user agent string to spoof various browsers,

Not perfectly ;-)
User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; en) Opera 8.50

Okay, it's the current version that produces the above, but as far back as I recall, Opera always appended the "Opera" string at the end. The better log analysers aren't fooled by this.

If you are talking about the modifications to the ua.ini file for types 4 and 5, I don't think there's a lot of people doing that.

Re:Finally....

[BandwidthHog]

Yeah, I should have read through the rest of the thread before I threw in that response. “My bad,” as the kids say.

Re:Finally....

[LPetrazickis]

Opera, as a rule, tells all server cache demands to get stuffed. This means you can instantaneously go back as far as you want without reloading (which is wonderful on, say, my bank's website because I can just wave myself back to the main page and drill down to a different account and not worry about the SSL Page Expired bullshit I'd get in Firefox; ditto for my university's course selection site). It also means that it will be horrendously undercounted on any server that forces IE to reload over and over again.

Mobile devices

[griffinn]

There are specific editions for mobile devices. It's no wonder that they don't access the the front page directly.

Many people go to BBC, CNN and other major sites through their mobile service provider's front pages. These would naturally point to the dedicated mobile editions too.

Re:Mobile devices

[corbs]

Also those web savvy enough to be using firefox would go directly to the section of the bbc webby they need (like news.bbc.co.uk). I find nothing particularly useful about the bbc homepage.

Re:Mobile devices

[Macka]

Ditto. Thats exactly what I do. He may mention something about this on the other pages in his blog, but I can't tell cos of the /. effect.

Re:Mobile devices

[Macka]

Yep. That WAP site is the home page on my mobile, and I probably use their traffic info page more than the others. I use it almost daily when walking to the car to check for reports of problems on the M4 before going to/from work.

Re:Mobile devices

[FuzzyBad-Mofo]

I use Firefox's Live Bookmarks to visit pages on the BBC site at least once a day. But I never use their homepage, so I guess those won't be counted.

Re:mobile devices

[pipingguy]

  The front page is way too big/rich for a limited handset.

If it doesn't feature advertising, it must be communist.

errr

[scenestar]

Linux (various distributions) 0.41%

Windows Vista 0.15%

  MSFT's unreleased os has nearly the same market share as linux?

We've got a long way to go.

Re:errr

[Saven Marek]

> MSFT's unreleased os has nearly the same market share as linux?
> We've got a long way to go.

I wouldn't say so. Most linux people use a browser string to look like windows so sites wont reject them.

Linux market share is about double Mac so its probably about 8 or 9 % so some of that Windows stat is actually Linux.

Not a fair comparison

Re:errr

[Gnutte]

Just put up some news on slashdot that's linked to BBC and it will soone change... By the way, what does the stats for slashdot.org look like?

Re:errr

[odaen]

So what you are saying is that 1 in 11 people I walk up to on the street will be using Linux?

I think not.

Re:errr

[Hey Pope Felcher . .]

. . . can you please point out some sites that routinely reject Linux users?

It was once useful to make sites think that you were visiting using a different browser other than IE, but, for the vast majority of web sites, those days are long gone. I have never, on the other hand, had to pretend to be using another OS to visit a site, never.

I would be greatly intrigued if you could give some examples that require you to be identified as using Windows.

Re:errr

[Anonymous Coward]

probably about 8 or 9 % so some of that Windows stat is actually Linux.

I take it this is referring to the Vista percentages?

Otherwise you would be inferring that NO Windows users (or at least a statistically small amount) use Firefox as their browser.

You're either a troll, or an element of the fanatical, lunatic, hopelessly optimitistic fringe.

Re:errr

[Anonymous Coward]

"Most linux people use a browser string to look like windows... Linux market share is about double Mac so its probably about 8 or 9 % so some of that Windows stat is actually Linux."

HAHAHAHAHAHHAHAHAHAHAHA!!

Oh, shit. BWAAAAAAHAHAHAAAAAH!!!!

Re:errr

[Anonymous Coward]

It is not as common to check for a specific browser as it used to be. Mostly because the checks were done by clueless web programmers that "optimized" for some web browser and refused to let others see his "work of art" with anything less than would allow it to be displayed to his idea of perfection. Of course, people went around this by changing the strings and web programming matured to the point where stupid browser checks were removed. I think the string change just stayed in a lot of browsers because of the trouble it used to give. Very few sites check for IE and those that do are probably closely related to Microsoft in some way.

I only have a vague memory of visiting some site that turned me away a few years ago. More recently, I was refused access to the Wal-Mart downloadable music catalog because of my browser, which was complete crap since my browser didn't have anything to do with whether I could see it. I've heard of people who couldn't do online banking because of the browser checks, but that's only hearsay. The Wal-Mart was is definitely a current example, though.

Re:errr

[Tet]

Most linux people use a browser string to look like windows so sites wont reject them.

Errr... no. Most Linux users will use the default setting for their browser, which for most people will not identify them at using Windows or IE. Yes, a very small number of people will do this, but to claim that it's "most" is just laughable.

Re:errr

[Anonymous Coward]

you are kidding right? you think 1 in 10 people are linux users for web browsing lol. ignoring other browsers not even 1 in 10 are even firefox users and of that only a percentage will be on linux. not even the most optimistic fringe would be putting linux near 10% of home users.

Re:errr

[Anonymous Coward]

Yes, but at work I'm forced to use IE and Windows - as are so many people.

If it wasn't for that, I'm sure that the stats would be even better.

Re:errr

[Anonymous Coward]

Well, most linux distributions ship with firefox as the default browser.
That has to account for something.

Re:errr

[m4dm4n]

A quote that is particularly apt for this whole story: "Lies, damn lies and statistics"

Re:errr

[YowzaTheYuzzum]

MSFT's unreleased os has nearly the same market share as linux?

By that logic, Windows 98 has nearly the same market share as Windows 2000.

Windows 2000 16.5%
Windows 98 6.6%

Re:errr

[AlaskanUnderachiever]

*cough*. . FEMA. . *cough*. . .

Re:errr

[Anonymous Coward]

Just put up some news on slashdot that's linked to BBC and it will soone change... By the way, what does the stats for slashdot.org look like?

There are at least 10-15 links to BBC on Slashdot every month. Oh, and Slashdot stats? :) Let's just say Taco is ashamed of running a Linux oriented site with a majority of users running Windows (and yes, they were once available).

Re:errr

[Tim C]

Do you really think that the sort of traffic the BBC gets will be affected that much by a slashdotting? Or do you not notice the half a dozen or so links a month from here to there?

Make no mistake, slashdot is big traffic-wise, but the BBC is much, much bigger (especially if you consider the whole bbc.co.uk domain, and not just news.bbc.co.uk)

Re:errr

[Anonymous Coward]

So what you are saying is that 1 in 11 people I walk up to on the street will be using Linux?

Around here that would be true. But I live in an area with a lot of tech companies that is also near a university with a big CS department, so there are a lot of geeks walking down the street.

Re:errr

[tritonic]

I'm afraid this doesn't follow. Saying 8% of the site hits are from linux machines does not imply that 8% of people run linux.

Re:errr

[Anonymous Coward]

I was amazed that Linux has such a small marketshare.
On the other hand, the BBC site is one that is respectable enough to be browsing during work breaks, so there people would use whatever the company provides, which in practically all cases is a Windows pc.

- aRTee

Re:errr

[aurelian]

A lot of Slashdot readers (myself included) access the site from work, e.g. at lunch or when slacking off, and may have to use Windows there - maybe even IE aswell. However I'd expect a high share for Firefox.

Re:errr

[Anonymous Coward]

Web Messenger rejects anything that's not IE or Mozilla 1.6

Re:errr

[Fred_A]

I concur, I have a user agent string switcher extension on a few machines, but it's been ages since I've used it to change from the default.

The sites that break with the standard Firefox/Linux string are getting increasingly rare.

And whenever I changed the UA string, I always made sure to change it back afterwards in order to show up in the stats. I'm sure most others did the same.

Re:errr

[spitzak]

I agree, this used to be true but no longer.

Two reasons: first sites started working, in that at least they removed the check and just fed their HTML, whether or not it worked on non-IE. Second is that the newer browsers support *temporarily* changing the string in a user-friendly way, old browsers would be permanently switched to IE as soon as the user fixed it to display one page.

Actually I suspect a large percentage of those very old IE versions they list are actually alternative browsers permanently switched to identify themselves as IE, inluding a lot of old Netscape versions.

Re:errr

[BandwidthHog]

No, but perhaps you could assume that if someone is walking down the street whilst using Linux, they are 11 times more likely to run into you?

Re:errr

[Anonymous Coward]

You're assuming people who use Linux go out in public, when, in reality, it is a scientifically proven fact that most Linux users are nocturnal and smell of Cheetos.

Re:errr

[bergeron76]

What's this we've got a long way to go stuff?

You appear to be a Microsoft shill. Why would you mention the operating system ("Vista") as some point of significance in a browser usage story - particularly in a 'we are losing' context?

If we all set up some bots...

[nmoog]

...with a shiny firefox user agent string - we could easy get that figure up to 30%!

Re:If we all set up some bots...

[Hazzl]

No need to attack the site with a bot net if a good old-fashioned slashdotting will have the same effect ;-) The site is already not responding anymore...

Re:If we all set up some bots...

[Anonymous Coward]

No need to attack the site with a bot net if a good old-fashioned slashdotting will have the same effect ;-)

The power of a Slashdoting is insignificant compared to the BBC.

Re:If we all set up some bots...

[bloodredsun]

Utter rubbish!

Have a look at alexa and you'll see that the bbc site deals with 20 to 30 BILLION hits a day. Slashdots 1 billion is not going to make much difference to their servers.

Re:If we all set up some bots...

[Zwaxy]

I think you are seriously misreading that graph. 20 to 30 BILLION hits per day? 4 hits per day for every person on the planet?

I don't think so.

Hint: the graph is labelled "per million" - that means you should divide by a million, not multiply by it. A reach of 20,000 per million means that 20,000 per million or 2% of Internet users visit each day.

Re:If we all set up some bots...

[RoboPimp_3000]

Although he misread the graph (you divide by a million to get a percentage of internet users, kinda like a Nielson ratings share) - his point is correct. BBC is a much larger website and slashdotting won't affect it much.

Here's a comparison of bbc.co.uk and slashdot.org:

http://www.alexa.com/data/details/traffic_details? &range=6m&size=medium&y=r&url=bbc.co.uk#top

Representative of Overall Market Share

[Mad Man]

As of September 2005, Internet Explorer has an 85% market share, while Firefox has a 9.5% market share.

The BBC's numbers are simply representative of this, as any large web site would be.

Re:Representative of Overall Market Share

[tgv]

Is slashdot large enough? I'm sure their statistics differ substantially. A slightly lower share of Explorer, I would hazard to guess.

Re:Representative of Overall Market Share

[pomo monster]

The summary you link to is crap. Where's Safari? Surely greater than 0.15% (Netscape's supposed share)? It seems they've lumped it in with either Firefox or IE; either way, like I said, it's crap.

Re:Representative of Overall Market Share

[diegocgteleline.es]

slashdot is a largue website and it's not representative

Re:Representative of Overall Market Share

[after]

Opera only has 0.23%, and it went down from 0.59%. That's interesting, since Opera is commercially founded, and it went free just some time ago.

It should be the opposite.

Re:Representative of Overall Market Share

[Rocketship Underpant]

"The BBC's numbers are simply representative of this, as any large web site would be."

Well, they'll at least be representative of the overall British marketshare. (And yes, I know not just Brits visit the BBC, but still...)

I imagine for a large German or Finnish media site, you'd see a much higher portion of Firefox users. Many countries in continental Europe seem to have much higher Firefox adoption rates.

Re:Representative of Overall Market Share

[wooger]

It's a British Site, hardly anyone here uses Macs ( I personally know no mc users). The famously claimed 5% marketshare is only for the USA.

Re:Representative of Overall Market Share

[Bloater]

> Firefox has a 9.5% market share.

That's a load of crap, I don't know *anywhere* to buy firefox, Firefox has a 0% market share. It has a 9.5% consumer webbrowsing share though - if that is important.

Re:Representative of Overall Market Share

[Ilgaz]

Opera still identifies as MS IE to webservers. While it was commercial, people using it were changing their id string to Opera manually as they are fan of browser (paid users) and advanced desktop users.

I assume after it went free, flood of newbie user got the browser and they didn't change ID String yet. I can't blame anyone, you can still live problems with identifying as Opera to some sites. While I was using Windows, I was identifying as Mozilla browser.

Its kind of stupid to identify browser features with their names in fact. What they (big sites) need is to check capabilities of the browser, not its name.

Apple got a very good article on that:
http://developer.apple.com/internet/javascript/obj ectdetection.html

Re:Representative of Overall Market Share

[KevCo]

That's a load of crap, I don't know *anywhere* to buy firefox

http://store.mozilla.org/product.asp?code=MZ80006& catid=1.

Hope this helps.

Re:Representative of Overall Market Share

[ZachPruckowski]

Well, that depends on how a web browser brings in its dough. Since all of the major ones are free (or, I downloaded the 8 or 9 big ones everyone mentions for Mac without charge), it seems to me that they must make money through some sort system dependent on page views. If someone could elaborate, I'd appreciate it.

If that is the case, then Firefox is in fact making 9.5% of the "sales" - whatever brings in the money, and it's usage base is the same number. So to sum it up, market share = usage share in this case (if I understand the browser business model). Don't think that something that is free doesn't make money - look at Google.

Re:Representative of Overall Market Share

[ZachPruckowski]

Slashdot is very large, but the second statistical requirement is that it draw from the same population. In this case, it draws from a sub-population that is disproportionately computer literate and/or involved in computing. The difference is that a /. user is more likely to go to the trouble of using a non-default browser. This means that /. data can't be used for this reliably. Large, common sites like CNN or BBC are good, because they are visited by everyone equally (sort of), and thus will reflect the overall numbers roughly.

Re:Representative of Overall Market Share

[BasilBrush]

You obviously didn't read the main article. The Mac has 4.4% of visits to www.bbc.co.uk. Clearly that is going to be biased towards British users. I know plenty of British Mac users - you should get out more.

Re:Representative of Overall Market Share

[tgv]

Ok, since you're getting technical: my point is that I don't believe the BBC draws a representative sample of the population either. Too many English, to begin with...

Re:Representative of Overall Market Share

[joeljkp]

Opera stopped spoofing itself as IE by default a while ago (before it went free, even).

http://operawatch.blogspot.com/2005/07/opera-to-st op-spoofing-user-agent-as.html

mobile devices

[nother_nix_hacker]

The BBC provide specific pages for mobile devices. The front page is way too big/rich for a limited handset.

Opera

[Anonymous Coward]

My install of Opera is set to identify itself as IE to websites as I am sure many others set theirs the same way on install. So in that light, are those figures trustworthy?

Re:Opera

[Sockatume]

There's certainly room for error. If we had figures for how many Firefox and Opera users have their browsers masquerading as IE, we could put together a cludge factor to correct it.

Re:Opera

[Anonymous Coward]

The Opera browser is perfectly recognizable from its user agent string even when it "masquerades itself" as a Internet Explorer. It is hard to imagine a person whose job is to produce these statistics to not know about it.

Re:Opera

[YA_Python_dev]

My install of Opera is set to identify itself as IE... are those figures trustworthy?

Yes, they are.
Old versions of Opera that identify themselves as IE by default use a user agent string like this:

Mozilla/4.0 (compatible; MSIE 6.0; X11; Linux i686; en) Opera 8.02

So the "Opera" string is here and easily identifiable.

New versions should simply use the proper Opera UA string by default.

If you use Opera I suggest to check that it sends the "correct" Opera UA string: the sky will (mostly) not fall down.

Re:Opera

[Anonymous Coward]

So it's still possible for sites to block access to any browser but IE as they have done in the past?

Re:Opera

[GrahamIX]

"...are those figures trustworthy?"

Um, you're the one mis-reporting your User Agent! They can only report on the data they are given. [runs away to find flame proof trousers...]

Re:Opera

[KiloByte]

Yeah, but Internet Explorer still tries to pretend it's Mozilla...

Re:Opera

[peterpi]

To the nearest few percent they are trustworthy, even with your Opera install skewing the figures.

We need to remember that people who do unusual things with unusual browsers are an incredibly small fraction of all internet users. The message of the article is that there's very rougly a 8/1/1 split between IE, firefox and 'other'. That message is not affected in the slightest by Opera, lynx or any other niche browser.

Re:Opera

[Mugros]

For Firefox this error has to be really small. FF never changed its user agent string by default. I had to use the user agent switcher on FF0.8 a long time ago to get access to my banking site. But then they fixed the site and since then i never had to use the user agent switcher again and deinstalled it.

Opera users just needs some self confidence. I have Opera and it is set to identify as Opera, but i am only using it for testing. And IIRC, even if it is set as IE the user agent string is different than that that the real IE. Only some stupid user agent string analyzers don't recognize it.

Re:Opera

[mdwh2]

We need to remember that people who do unusual things with unusual browsers are an incredibly small fraction of all internet users.

Actually what the grandparent post described - identifying as IE - happens by default (it's people who identify as Opera who are the unusual small fraction). The question is whether these stats pick up the "Opera" bit at the end, or whether it's just lumped in with IE.

The message of the article is that there's very rougly a 8/1/1 split between IE, firefox and 'other'. That message is not affected in the slightest by Opera, lynx or any other niche browser.

Except we don't know whether it's a niche browser or not, if we don't know what it's being counted as. That's circular logic: "The accuracy of the stats for Opera don't matter as it's a niche browser, therefore we can conclude that Opera is a niche browser".

Re:Opera

[Anonymous Coward]

At the top end these figures are reliable, yes. 80-odd percent of users don't change their browser settings.

But it wouldn't surprise me if a significant minority of Firefox users didn't fiddle with user agent strings - although of course by default newer versions do reset on reload by default.

Re:Opera

[Anonymous Coward]

We need to remember that people who do unusual things with unusual browsers are an incredibly small fraction of all internet users.

You're begging the question. You label such browsers as niche because your statistics tell you that they are niche. Your statistics tell you that they are niche because you aren't willing to entertain the possibility that a substantial number of users do such things. You aren't willing to entertain that possibility because you've labelled them niche. Circular logic.

Re:Opera

[Midnight Thunder]

This shows you that, while you may change your browser string to access certain sites, in the long run it actually works against you when trying to prove how often Opera accesses the site.

We need to see Opera identify itself properly and give a big middle finger to broken web sites, at least after giving the webmaster a chance to fix their issue. Remember there are some stupid webmasters, but there are many that will fix the issue if the get told about it. If you have to go through a help-desk to sort th issue, then there are chances that bozo 101 is going to give you some catch all answer and file it under "some idiot believes that there is some other browser than IE".

Re:Opera

[cant_get_a_good_nick]

Technically, it still tries to identify itself as Netscape Navigator. Mozilla was the internal nickname at Netscape, short for Mosaic Killa. The original Mosaic team had split into two groups, Andreeson with Netscape and some other guys with Spyglass Mosaic. In the beginning, there was a rivalry for who's browser would dominate.

The company Spyglass is essentially dead, though the code lives in as the base of early IE. You can say that Microsoft killed Mosiac (well spyglass) by incorporating it into IE and giving it away for free. I seem to remember hearing that microsoft offered them a cut of sales of IE, then gave it away free, cutting Spyglass out of a revenue source. I wonder if any lawyer got creative, and since "IE is a fundamental part of the OS" tried to get a cut of Windows OS revenues.

Re:Opera

[whitehatlurker]

If you use Opera I suggest to check that it sends the "correct" Opera UA string: the sky will (mostly) not fall down.

Mostly. There are still websites out there which need to see MS IE coming in, even if there is nothing else wrong with them and they will work okay in a reasonable browser. (Proxomitron comes in handy for lying more convincingly for those sites.)

I agree that Opera users should use the correct UA. (And complain when sites refuse them access.)

To the original question, most web hit counters are knowledgeable enough to figure out "Hey this is really Opera!" The BBC reviewer in particular seems competent, knowing about both Opera and Proxomitron (which is interesting). I would feel confident that he got the count right.

Variability by site

[danfreak]

Interesting. I wonder how much variation there is of browser use by other sites... I imagine BBC is higher than most in the Mozilla-bred catagory, as the BBC News site has posted lots of articles about Firefox over the years. I wonder how different it would be for msn.com, foxnews.com etc.

On a related note, I hosted some pictures on my website last week that were posted into a fark.com forum, 47.6% of fark readers seem to use Firefox (from some 14,000 hits in two days) - I bet slashdot beats this though!

Re:Variability by site

[ottffssent]

The IE / FF split is actually pretty representative of the world as a whole. Other recent data points show about 85:10 split for IE / FF.

Of /. users who have visited my site from my sig, the split is almost exactly the opposite, 8:75 IE / FF. That's a somewhat higher percentage of FF users than non-slashdot traffic, but IE is in the minority month after month.

Re:Variability by site

[manojar]

I wonder if that percentage would correspond to real-life too! Considering the demography of the people visiting BBC for news. Within UK, we can say it is true. But, outside UK, we have to consider that people who visit a site like BBC would be more inclined to use FF. Conversely, people more inclined to use IE (and support it against FF) could not be visiting BBC in the firstplace. Just my opinion.

Re:Variability by site

[a.d.trick]

I assume the stats for msn.com have IE at about 99%. I mean who would intentionally go to that site if their browser did not automatically redirect them all the time.

Re:Variability by site

[peterpi]

I imagine BBC is higher than most in the Mozilla-bred catagory, as the BBC News site has posted lots of articles about Firefox over the years.

I doubt it makes much difference. The BBC news site is read by a lot of Normal People who either couldn't care less about what browser they're using, or have no power to change it because it's a work computer.

I'm really surprised that firefox has such a high share. Of course there have been similar stats released by sites like i-am-a-1337-linux-doodz.com and windoxxors-is-teh-suxxors.com, but to get them from a mainstream site like the BBC must be very encouraging for the developers

Re:Variability by site

[westlake]

But, outside UK, we have to consider that people who visit a site like BBC would be more inclined to use FF

You don't need to be running FF to know the value of the BBC.

Re:Variability by site

[Jupix]

It varies a lot and website contents are not the only thing that makes it vary - there are countries where browser education is still nonexistant. For example, the American population uses IE a lot more than the Finns. This clearly shows not only in the media but in website quality as well.

To give an example of how contents affect the portion of each browser, let's look at the development forum of a CSS skin I made for phpBB. The statistics are from 12000 page loads. You'd expect these visitors to be sophisticated in website management and browsing The Net and thus know the difference between IE and Firefox.

Firefox: 36 %
MSIE: 31 %
Opera: 2 %
Konqueror: 1 %

OTOH, let's look at a joke collection site (in Finnish) that I run. You'd expect the visitors to be older, and interested more in the contents than how they view it. Loads: 52000. Here's the stats:

MSIE: 69 %
Firefox: 19 %
Opera: 1 %
Konqueror: 0 %

That pretty much sums it. Firefox still has a long way to go but it will get there eventually unless the internet undergoes some kind of a fundamental change.

Re:Variability by site

[fireboy1919]

Ah...doubt it. My small personal site gets most of its visitors from slashdot in times when I link to it.

Take a look at my stats.

Then again, maybe I don't get enough visitors for any kind of accuracy. I keep getting somebody from "cups.cs.cmu.edu," and I've got no idea who that is. They're visiting enough to be statistically significant, though.

Re:Variability by site

[Ponyegg]

we have to consider that people who visit a site like BBC would be more inclined to use FF. Conversely, people more inclined to use IE (and support it against FF) could not be visiting BBC in the firstplace. How do you draw that conclusion? I know you said that this was simply your opinion but it seems utterly baseless. And that you seem to think that somehow there are 'sides' in which browser you choose to 'support' use is simply pathetic and childish.

Re:Variability by site

[Anonymous Coward]

Why, that's me of course!

Firefox 10 percent!!

[Anonymous Coward]

From BBC..., Can we celebrate the 10 percent share of Firefox then?

Sampling?

[Anonymous Coward]

From the article:

The figures may, however, mask a slightly higher use of Linux. Since the user agents generated are more likely to be unique, they are more likely to have fallen into the statistical long tail.

What is this supposed to mean? If he has the logs he has the entire sample space. He should be able to get the exact distribution. Somebody needs to look through their statistics book for a refresher rather than throw around big words.

Re:Sampling?

[Spad]

He had over 33,000 different user agents to sort through - why don't you email him and offer to trawl through the 22,000 UAs making up the 5% of traffic that he didn't generate stats for.

Re:Sampling?

[Anonymous Coward]

The entire sample space doesn't give you the exact distribution. Maybe you need to read up some more too.

Re:Sampling?

[Midnight Thunder]

He had over 33,000 different user agents to sort through - why don't you email him and offer to trawl through the 22,000 UAs making up the 5% of traffic that he didn't generate stats for.

Actually it sounds like it would be a useful project if someone actually put together a registery of know UAs and then added some associated info on the UA. Maybe something like:
    - client type (browser, bot, etc)
    - platform & version
    - browser name & version

Ideas are easy, putting them into action is something else.

Fatally Flawed

[Anonymous Coward]

I visit the BBC web site multiple times a day, but I haven't been to the "main" page in months. I expect most regular Firefox visitors will have bookmarks or just type a URL that goes past the main page.

The author does point this out:
And I must stress again, these figures don't represent the breakdown of visitors to the BBC site as a whole, they are based on requests to the homepage alone, over the course of one week in September. Nevertheless I think they provide an interesting snapshot of web activity.

but it should have been avoided

Re:Fatally Flawed

[Deaths Hand]

Indeed, I also visit the BBC website many times each day, but might go to the main page about once every three months. Most people (many are Firefox users) in my office have the BBC news main page http://news.bbc.co.uk/ set as their homepage, but this wouldn't show up in these statistics.

Re:Fatally Flawed

[Malc]

And why would Firefox users be any different in this respect to users of other browsers? If Firefox users are bypassing the main page in this way, why wouldn't IE users?

Re:Fatally Flawed

[ZachPruckowski]

There is also the RSS issue as well. One of my browsers somewhere (I have 9, 10 if you count AOL, which I don't) has an RSS feed to BBC. I don't use it much, I read CNN, but I imagine RSS helps people avoid going through the front page.

Re:Fatally Flawed

[Anonymous Coward]

And why would Firefox users be any different in this respect to users of other browsers? If Firefox users are bypassing the main page in this way, why wouldn't IE users?

Firefox users have all ready taken the time to investigate their online surroundings, they are more likely to keep good bookmarks and know which page they want to visit. I have no statistics for this, it's just what my friends and family do.

eg: I would type news.bbc.co.uk to see the BBC's news, while my mother would use the "internet keyword" bbc and then click on news.

Re:Fatally Flawed

[Malc]

I stopped using bookmarks years ago when I switched from Netscape to IE and started taking advantage of its URL completion. Big lists of URLs and well organised bookmarks are so 1995. Both of my parents use Firefox, but they get to the Beeb's website via a Google search, no matter how hard I try to show them a better way. Anyway, I think you're making rather large assumptions.

Site is already slashdotted

[mahesh_gharat]

Just only 10 comments and site is already slashdotted.

Slashdot stats?`

[zerojoker]

Would be interesting too. Browser stats, OS stats ...

Re:Slashdot stats?`

[mr_tommy]

You miss the point of interest with the BBC; it is the number one website in the UK and thus has a reasonably representative audience. Slashdot, however much we love it, does not. I'm thinking male, 14-30, pretty high tech outlook - implying a skew towards Linux / Firefox / etc etc.

Bottom line - the beeb gives us a good painting; it's not a picture, true, but it is a good picture. Mozilla folk should be pleased with themselves; their strategy has worked rather well.

Re:Slashdot stats?`

[lbmouse]

I'm thinking male, 14-30, pretty high tech outlook
 
So basically the same demographic as a porn site.

Re:Slashdot stats?`

[maxpublic]

Yeah, but Firefox supporters are like Mac fans: nearly religious in their zeal. The idea that Firefox's share is tiny in comparison to IE really rubs them the wrong way, as they're convinced that all stats that don't support their view of how things should be are inherently flawed.

Me, I see the Firefox percentages among average Joe and Jane to be hopeful signs for the Firefox developers, and indirectly for improved security overall. But then I'm not a fanatic; hell, I don't even use Firefox (Opera for me).

Max

/.'d

[sam_paris]

Already.

Ah well, at least I had time to read the first page of the article, shame really as I was looking forward to that second page :(

Mirror

[Jugalator]

Well, I for one couldn't access that blog. Here's a mirror...

How about Slashdot generating a mirror link via a neat little "mirror" icon next to the links?

Re:Mirror

[Jugalator]

Damn, disregard that post... As usual they've split it into pages...

Only nyud.net links may help then, although my experiences with those aren't the best and why I tried to avoid it in the first case.

Re:Mirror

[Anonymous Coward]

Nice attempt to karma-whore... one slight problem. The first page is the only one that is mirrored! idiot.

Slashdotted....

[MacGod]

Hmmmmm.... Slashdotted already.

I have a hunch this guy's web stats are going to show a MASSIVE influx of FireFox users, then a long period of downtime...

Re:Slashdotted....

[6*7]

New statistical evidence coming soon:

Firefox (slows) down(s) webservers

And his second report...

[mustafap]

Will be on the visits to his webpage, when the server recovers :o)

How about stats on slashdot readers

[gumnam]

Does slashdot publish browser/system stats about slashdot readers ? That would make interesting reading.

It depends upon your site content

[vivekg]

First here is mirrordot link, if you cannot open page (slashdoteffect).

My site and blog mostly related to Linux and Open source stuff, and here is my exprince so far:
OS
Most of the corporate users, uses Windows XP/2000 desktop
Individual user uses Linux/BSD/Mac OS desktop


Browser
Firefox rules
IE (6.x/5.x)

So it depend upon your site content, if you wanna see this stats they are here

BBC news, typically read at work

[Colin Smith]

So it's probably about right for UK business desktop stats.

Re:BBC news, typically read at work

[Anonymous Coward]

Anything that helps remove the sting :).

Re:BBC news, typically read at work

[BasilBrush]

No, the BBC news site is news.bbc.co.uk not www.bbc.co.uk.

No MSI build for Firefox - no mass deployment

[ph1l0r]

at companies that run Windows clients. I wouldn't bother to install Firefox more of less by hand on hundreds of desktops myself. The Firefox guys should really get a MSI build ready for easy deployment _and_ update. Firefox is just not 100% enterprise ready like IE is with it's managabilty by group policies. I wonder how many people check bbc.co.uk from their workplace. They might even have Firefox installed on their home computer.

Re:No MSI build for Firefox - no mass deployment

[petermgreen]

i'm pretty sure there are ready made scripts out there to build a msi from the latest firefox release if thats your preffered method of deployment.

here at uni they deploy firefox on demand through zenworks like pretty much every other app they have avilible. I think they build thier own packages for that though.

Re:No MSI build for Firefox - no mass deployment

[Anonymous Coward]

There you go..
http://www.frontmotion.com/Firefox/
MSI installers for Mozilla Firefox! Useful for installing Firefox on a single computer for the home user or deploy across thousands of computers automatically with Microsoft's Active Directory. Use Firefox on your corporate computers to decrease virus incidents and increase overall security. Save time and frustration with our installer that is targeted toward the corporate IT administrator with manageability and upgradeability in mind. This is not just a wrapper around the exe installer nor is it another half baked 'captured' install.

Re:No MSI build for Firefox - no mass deployment

[diegocgteleline.es]

The Firefox guys should really get a MSI build ready for easy deployment _and_ update. Firefox is just not 100% enterprise ready like IE is with it's managabilty by group policies

Stop blaming software makers for that. Microsoft pushed for a _long_ time the _crazy_ idea that "programs must have their own .exe installer", which lead to companies to create (crappy) installer products (!!), which leads to thousands of different installers not being able to update your system as they should, which leads to the DLL hell and other hells. Just because Microsoft hasn't been able to fix their OS and make software developers easier to have decent installers for a long time doesn't mean everybody has to support it right now.

There're thousands of non-MSI installers out there not just Firefox, there're _tons_ of installers that have dialogs that need to be clicked to be installed (they don't have command line options), there're tons of installers that won't even _start_ unless you've administrator rights. There're thousand of programs that can't be "upgraded", you need to uninstall (if you're lucky, you can't automate it either) and then install the new version. It's the Windows plataform who is not "easy-management-compliant" not software, and it's Microsoft who must be blamed for not adopting package-like technologies when windows 98 was out, not software makers.

Re:No MSI build for Firefox - no mass deployment

[naelurec]

I have been using http://www.frontmotion.com/Firefox/ for about 7 months for a network of ~50 systems. I use this with FirefoxADM for some basic browser configuration. There will be the "ADM XPI for Firefox 1.5" which from my brief understanding will integrate Firefox config w/registry and allow for tighter group policy management.

It would definitely be nice to have an official MSI package (as it would attract many more admins) coupled with something like ADM XPI .. *hopefully* that will happen.. I thought I read that 1.5 would have an official MSI but couldn't find a reference.

Re:No MSI build for Firefox - no mass deployment

[Malc]

What the hell are you talking about?

What's the installer got to do with DLL hell? That's a problem of the application packager. To be honest, I haven't seen DLL hell for years.

It makes complete sense to use MSI for Firefox. Leverage both the deployment infrastructure that exists and the installer functionality that doesn't need to be reinvented.

As for your point about some installers requiring admin rights - wtf? Apt on my Debian system won't even run unless I'm root, so how is this any different?

Most of the custom installers are crap compared with MSI. We have one with our products, and as a software developer it's a PITA compared with MSI due to bugs that we keep finding many years, the fact that we have to evolve it with Windows, and the fact it doesn't offer everything that MSI does. How many custom installer will do a proper roll-back if there is a problem during the install process?

Re:No MSI build for Firefox - no mass deployment

[ph1l0r]

I would need a script to build a german msi build. Other would need spanish etc etc etc. These 3rd party scripts are a start. Since Firefox ain't distributed as zip file anymore it even got harder to create a MSI own his own. Especially for localized builds.

Re:No MSI build for Firefox - no mass deployment

[Midnight Thunder]

The MSI needs to be on the http://www.getfirefox.com/ page and not on some obscure site that only well informed techies seem to know about.

Re:No MSI build for Firefox - no mass deployment

[ph1l0r]

Doesn't help me all since i need a localized version of Firefox.

Re:No MSI build for Firefox - no mass deployment

[ph1l0r]

msi is out for ages and communicated as _the_ way to install software on the windows platform. I see no reason as to why Firefox couldn't or shouldn't be distributed as msi package instead of the current .exe only installer builds.

Re:No MSI build for Firefox - no mass deployment

[amchugh]

Funny when I google for MSI for firefox, that obscure site comes up first.

Re:No MSI build for Firefox - no mass deployment

[petermgreen]

iirc the reason ff moved from a zip to an installer was some registry key that was needed to make it work properly with a newer version of suns java.

they do however distribute xpis which are apparently just zips with some scripts inside maybe those would be a good starting point for autobuilding custom packages without re-building firefox itself from source. ftp://ftp.mozilla.org/Public/mozilla.org/firefox/r eleases/1.0.7/update/win32/

Re:No MSI build for Firefox - no mass deployment

[RoLi]

That's not really very helpful, because most will not be satisfied with some "unofficial" hacked together msi from some random page on the internet.

But actually that's good news because Firefox still has a lot of potential to grow. :-)

Coral cache

[holy zarquon's singi]

http://www.currybet.net.nyud.net:8090/articles/use r_agents/index.php, but that seems to be /.'d also.

Opera

[Anonymous Coward]

I wonder if the percentages for Opera are accurate as Opera tends to identify as MSIE by default.

Super Respectable

[Mulletproof]

"but Firefox had a very respectable 9.7% share."

I use firefox and even I can't keep a strait face reading that line. I mean have some self-worth, man. There's nothing respectable about that. Can't we aim just a tad higher here? Especially if we're gonna tag on the word "very"?

Re:Super Respectable

[grimJester]

It does sound a bit condescending. "Good boy, you've done a very respectable job. Have a lollipop."

Re:Super Respectable

[Midnight Thunder]

You have to remember that IE is installed standard with MS-Windows, just as Safari is on the Mac and Konquerer on KDE. A lot of people are quite happy with what they get and even if they aren't, they aren't necessarily adventurous enough to try some alternative.

In English there is the saying: Better the devil you know than the devil you don't.

All I could get of the article (page 1 and 2)

[a.different.perspect]

It started with a casual enquiry from a colleague - "I wonder how many Firefox users visit the BBC homepage?" - and before I knew it I was involved in a lengthy statistical analysis of the browsers and operating systems that request the BBC homepage at http://www.bbc.co.uk/

Our old stats reporting tool at the BBC gives a breakdown of requests from different user agent strings, which is where the browsers and operating systems people use to navigate around the web leave their digital fingerprints. It is about to be phased out in favour of a new solution, but I'm not sure that the new system gives the same granularity of data, so once I'd started, I thought I'd look at the figures in some detail before the old system gives up the ghost.

Now if you've never looked at user agent strings, they are rather dull and geeky, and full of lots of technical gubbins like these examples:

* Mozilla/5.0 (Windows; U; Windows NT 5.2; en-GB; rv:1.7.10) Gecko/20050717 Firefox/1.0.6
* Mozilla/5.0 (Macintosh; U; PPC Mac OS X; en-us) AppleWebKit/85.7 (KHTML, like Gecko) Safari/85.5
* Mozilla/4.0 (compatible; MSIE 6.0; America Online Browser 1.1; Windows NT 5.1; SV1; .NET CLR 1.1.4322)
* Mozilla/4.0 NETIKUS.NET GetHttp v1.0
* Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; Hotbar 4.5.1.0)
* Mozilla/4.0 (compatible; MSIE 6.0; Windows CE)
* Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031007 Firebird/0.7
* Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0; T312461; BT [build 60A])

There are of course some caveats around the figures I'm about to talk about.

User agent strings aren't an exact science. Or rather, they ought to be, but in the real world the come out a right mess. I've done my best to untangle them, but I still ended up with a significant number of user agents that I could not identify properly. And that is before we get started on the corporate networks that use the UA string to broadcast their corporate branding to the world whilst masking their operating system. Or requests claiming to come from both Internet Explorer 6 and Internet Explorer 5.5. Or that claim to be from a particular Linux distribution and Windows 98 at the same time. Or the plain weird like the inadvisably named KummClient from Hungary that proudly proclaims 'Linux rulez' to anyone like me dull enough to be delving through their logfiles.

User agent statistics on something as big as the BBC homepage could almost be the very definition of the long tail. The most popular user agent string - IE6 on Windows XP - clocked up nearly 6 million requests. I only counted user agents that had made more than 50 requests, but between 6 million and 50 requests there were nearly 11,000 different user agents to look at. Examining that number of requests accounted for 95% of the reported traffic, but only around 1/3 of the stats report. I initially suspected that counting the whole of the tail was likely to increase the market share I derived for the quirkier set-ups, but a random sample showed that a large proportion of the tail consisted of the most popular browsers and operating systems, but with different installed toolbars or corporate network messages that distinguished them as a unique string.

And I must stress again, these figures don't represent the breakdown of visitors to the BBC site as a whole, they are based on requests to the homepage alone, over the course of one week in September. Nevertheless I think they provide an interesting snapshot of web activity.

In total I've examined around 32 million requests to the BBC servers - although some of these have been discounted as 'unknowns' and some originate from crawlers and spiders.

The complete dominance of Windows XP and Internet Explor

Re:All I could get of the article (page 1 and 2)

[pubjames]

In total I've examined around 32 million requests to the BBC servers - although some of these have been discounted as 'unknowns'

I would be interested to know what percentage were discounted as 'unknowns'

Re:All I could get of the article (page 1 and 2)

[Anonymous Coward]

What's unfortunate is that Opera isn't represented here, or in the OP.

Some of us have been using Opera's tabbed, secure browser since long before this upstart Firefox came on the scene...and we still think it's better than Firefox.

Go figure.

Re:All I could get of the article (page 1 and 2)

[Dunkelzahn]

Probably all the Opera users using the handy little user agent spoof thingy to make the web server think its IE. Just my guess.

Re:All I could get of the article (page 1 and 2)

[Hal_Porter]

Yes it is

Opera 0.32%

from here

http://www.currybet.net/articles/user_agents/4.php

I prefer it too, seems quicker and more stable than FF. Probably at 0.32% market share it should be a non tempting target too for exploits too. And it's free now, the Google ads in the old versions have been disabled.

Incidentally, even when its told to identify as IE, it still mentions Opera at the end of the user agent string. It's kind of cool really, the old Mozilla would identify as Mozilla. Internet Explorer would identify as

Mozilla X.XX ( compatible; MSIE X.XX; Windows NT X.XX; YY)

where YY is a language code, e.g. en for English

So now Opera identifies as

Mozilla X.XX ( compatible; MSIE X.XX; Windows NT X.XX; YY) Opera X.XX

i.e. Opera pretending to be IE pretending to be Netscape.

If you use netcat nc -l -p 80 and go to 127.0.0.1 in Opera, you see these headers -

GET / HTTP/1.1
User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0; en) Opera 8.50
Host: 127.0.0.1
Accept: text/html, application/xml;q=0.9, application/xhtml+xml, image/png, image/jpeg, image/gif, image/x-xbitmap, */*;
q=0.1
Accept-Language: en
Accept-Charset: windows-1252, utf-8, utf-16, iso-8859-1;q=0.6, *;q=0.1
Accept-Encoding: deflate, gzip, x-gzip, identity, *;q=0
Connection: Keep-Alive

As always, defaults play a role

[The Hobo]

Remember, with every English-US installation of Firefox comes a preloaded RSS feed on the bookmark toolbar that points to the BBC for news (I say this as an avid Firefox user)

Re:As always, defaults play a role

[Anonymous Coward]

he said 'these figures don't represent the breakdown of visitors to the BBC site as a whole, they are based on requests to the homepage alone' so anyone using rss feeds wouldn't have been counted!

Re:As always, defaults play a role

[IngramJames]

Firefox comes a preloaded RSS feed ... that points to the BBC for news

Maybe so, but that's not the homepage, which is from where the stats were taken :-)

Re:As always, defaults play a role

[petermgreen]

surely that would made ff LESS likely to produce hits on the homepage as people would be getting the stories fed to them directly by RSS.

Firefox comes with a "Live Bookmark" to the BBC

[osbjmg]

One thing I noticed when I installed Firefox, is that it comes with just one live bookmark. It is called: "Latest Headlines", and pulls the feed from http://fxfeeds.mozilla.org/rss20.xml/ But, this feed is the same as the main stories feed at BBC. I would figure people would click on these and get some more exposure to the BBC site, more than usual. This has actually made myself more aware of those stories, and made me more likely to visit again.

I run a website with 20% Firefox usage as of now!

[shaka]

I run a website, a webmagazine in Sweden. It started out as a music/lifestyle webmag, and is now more of a collection of blogs, mostly about music and related things (sports, debate, feminism, lifestyle, TV). In other words, the visitors are not at all tech type guys, but it's definately an inner city, trendy type of crowd. I would not hesitate to call them early adopters. Nonetheless, I was amazed when I checked the browser stats for October after reading this article. WE HAVE 20% FIREFOX VISITORS! Please see below the figures for the top 5 browsers (not counting the mysterious "Unknown" browser, which is mostly RSS aggregators and its ilk):

(Sorry for the bad formatting, why can't Slashcode support the pre-tag?)

Month IE Firefox Safari Mozilla Netscape
May 64.2 16.7 5.3 0* 1.3
June 64.7 17.9 5.4 0* 1
July 65.9 15.6 4.6 3.8 2.1
Aug 66.8 16.6 5 2.5 1.6
Sep 62.9 19.2 5.9 3.2 1.7
Oct 59 20 5.8 3.9 2.4

* Before this date, Firefox & other Mozilla were lumped together.

5% not a good penetration for windows, non-ie

[Anonymous Coward]

You have

everyone that went to the bbc: 100%
everyone that went to the bbc that used windows: 95%
everyone that went to the bbc that used non-windows: 4.9%
everyone that went to the bbc using ie: 85%
everyone that went to the bbc using windows but not ie: 5%

A 5% penetration on windows is not very good. It's always been that, going back to netscape 4, and even more before ns4.

Re:5% not a good penetration for windows, non-ie

[Toutatis]

Have you assumed that everyone not using windows is using Firefox? I don't think so.

Re:5% not a good penetration for windows, non-ie

[yomcat]

Thats because it's wrong.

Thecounter.com has stats as well.

[Anonymous Coward]

Thecounter.com gets like 600 million hits a month since they make software for many different web sites, here's a link to their browser stats for sept. 2005:

http://www.thecounter.com/stats/2005/Sept ember/browser.php

1. MSIE 6.x    62223734 (83%)
2. FireFox       5806423 (8%)
3. MSIE 5.x      3170911 (4%)
4. Safari        1311540 (2%)
...

Here's a link to their OS stats for the same month:

http://www.thecounter.com/stats/2005/Sep tember/os.php

1. Windows XP    54945908 (73%)
2. Win 2000       8468339 (11%)
3. Win 98          6806316 (9%)
4. Mac             2317188 (3%)
5. Unknown         1132090 (1%)
6. Win NT           473897 (0%)
7. Linux            322362 (0%)
...

Probably more accurate since they count more hits on a wide variety of different types of web sites.

Re:Thecounter.com has stats as well.

[The Second Horseman]

The really interesting thing is the Mac stats that we're seeing -- 3%, 4%, etc. Makes me wonder if the Mac market share is really growing, or if there's been a bubble of existing customers replacing older machines. Certainly, the iPod isn't driving some huge increase in Mac users - unless they have a population that only uses their Mac to go to the iTunes site.

Obvious solution

[LaughingCoder]

The obvious solution is to make the BBC homepage the default homepage for Firefox!

Re:Obvious solution

[Bernie]

They kind of already do, with a live bookmark hitting the main RSS feed.

Let IE and Firefox battle it out

[GroeFaZ]

I'm all for a diverse browser ecology, it only makes exploit writers' lifes more difficult. I, for one, am a most pleased Opera customer/user and while I hope that Opera the company will stay in business and refine their browser for a long time to come, I also kind of hope that Opera the browser remains in its 5% niche where it attracts no major attention from mentioned exploit writers.

Early adopters

[Peyote Pekka]

Smuggness about browser stats for general sites is fine. Don't let that distract you from the fact that early adopters pick up first what everyone else will be using later on. Those sites now frequented more by early adopters of ICT are using more Firefox than the general sites. That doesn't mean Firefox is niche or MSIE will always dominate, it means that those sites are preview of what we will probably see on the general sites.

Re:Early adopters

[Anonymous Coward]

Just like Linux on desktop right? Or is that revolution still to come?

Point them at various microsoft.com pages

[Colin Smith]

Scare the living shit out of them.

Re:Point them at various microsoft.com pages

[Anonymous Coward]

Somebody mod this up and funny already!

... and this is page 3

[YA_Python_dev]

Windows Operating System Share

Concentrating on just Windows alone we can see that Microsoft have done a very thorough job of converting their user base to the most recent iteration of the software. Windows XP accounts for just under 70.5% of the Windows requests, and Windows 2000 a further 17.4%. That means in total around 88% of users of Microsoft Operating System products are using the two most recent consumer releases.

Windows 98 features in 7% of requests made from a computer running a version of Windows, and after that the figures are very small in terms of market share. In fact the next largest figures is a clump of 'Windows other' including Windows CE, and various unspecific Windows NT user-agents that I couldn't pin down to a precise version.

Operating System Share of Windows Requests to the BBC Homepage
Windows XP - - - 70.5%
Windows 2000 - - 17.4%
Windows 98 - - - 6.99%
Windows (Various versions including CE, 3.1 and ambiguous UAs) - 2.23%
Windows NT - - - 1.90%
Windows NT5.2 -- 0.63%
Windows 95 - - - 0.21%
Windows Vista -- 0.16%

Chart illustrating the version share of visits to the BBC homepage using Windows software

Mac Operating System Share

I was frustrated in my attempts to similarly breakdown the different versions of the Mac OS that people were using to request the BBC homepage. I established that from the requests we saw I could identify Panther as supplying 31%, Tiger supplying 21%, with Jaguar lagging behind at 3%. However there were 41% of requests where I could identify that the computer was a Mac, but not the specific version. That is because Safari helpfully supplies in the user agent string the WebKit build, allowing the precise version of the OS to be identified, but most other browsers do not.

Linux Requests To The BBC Homepage

The number of Linux requests to the BBC homepage was very small, representing only 0.41% - less than 100,000 - of the 32 million requests included in this study. With such a comparatively low number I didn't take the time to delve into which different distributions were driving the requests.

The figures may, however, mask a slightly higher use of Linux. Since the user agents generated are more likely to be unique, they are more likely to have fallen into the statistical long tail. However I should add that my random samples of the tail did not show that it consisted entirely of Linux, in fact as I mentioned earlier, a lot of corporate-branded Windows networks show up in the tail.

Legacy OS Systems

We have some fairly strict standards for supporting legacy technology at the BBC on the client-side - but the long tail of older OS software visiting the BBC homepage is amazing. We still saw over 300 requests for the BBC homepage coming from machines claiming to be running Windows 3.1, and around 200 requests from machines claiming to be persevering with 0S/2 Warp.

Small AuPair website in the UK

[Kroc]

I'm the webdesigner for a small AuPair company in the UK, our demographic is entirely UK Families and young foreign nationals.

For this month, this is the breakdown of browser access

5250 Views this month:
* 77.5% Internet Explorer (inc. Maxthon & AOL) = 4070 Views
    v5 (57 views) v5.5 (27 views) v6 (3703 views)
* 10.9% Mozilla Firefox (inc. Netscape & SeaMonkey) = 574 Views
* 02.3% Apple Safari (inc. Linux Konqueror) = 122 Views
* 00.4% Opera Browser = 22 Views
* 08.8% Other (Unknown, bots and rare browsers) = 462 Views

Even with this incredibly Windows/IE centric demographic (almost all being "regular" people), I'm very pleased to see a 10% Firefox Usage. The site only counts 1 view per IP per 24 Hours and ignores views from my IP and the Company's business address IP.

Preferred at many (most?) companies

[mgkimsal2]

i'm pretty sure there are ready made scripts out there to build a msi from the latest firefox release if thats your preffered method of deployment.

My guess is that this is the preferred method at most companies of > 50 people. I've worked at a number of companies over the past 3 years. This is by far the primary reason given for not deploying Mozilla/Firefox. MS gives tools to easily customize IE and push it out to everyone on the network very quickly. I'm working with a company now that realizes there are problems with IE - specifically DOM and scripting issues - yet we still use IE for the intranet apps despite these problems because customizing and pushing out Firefox and new versions to handle security fixes is deemed too much work relative to the IE answer.

Microsofts true competitor...

[spectrokid]

The market share of Win98 is bigger then Apple + Linux. That is a ten year old OS pretending to be a seven year old OS. And they say Windows is not stable?

Re:Microsofts true competitor...

[Anonymous Coward]

Win98 is stable. Well, not compared to a *nix system, but it sure beats XP.

Re:Microsofts true competitor...

[Kroc]

I'd only love to know where you got that authoritive piece of information. I've not had to reboot my XP machine in months. It's as stable as OSX and some Linux's as long as you look after the basics,...

unless that is your machine has so many virii/spirii that they've started to copulate? Win98 used to crash on me for just pressing the shift key sometimes, and I had an ultra clean install.

Re:Microsofts true competitor...

[Vo0k]

Install it on PentiumII 200MHZ with 64M RAM and 4GB hdd and see how long you can go without a reboot. I'm surfing the net from similar a configs on daily basis, with W98SE, WinNT4.0 and Linux and they are all pretty stable and fast. Meantime, a Celeron 1.5GHZ with 128M RAM and 40GB hdd, running XP Home Edition needs to be rebooted every few hours, as it slows down to a crawl.

Re:Microsofts true competitor...

[Kroc]

That is true, but the same can be said of GNOME which has memory fitting issues. I'm tired of people saying that WindowsXP is crap, no it's not, only very ignorant people say it's crap. You can get better than WindowsXP, but saying Windows98 is better is utter bull any which way.

A few years ago my laptop died and I was pretty stuck so I spent my last £100 on a PII300MHz, 256MB RAM and 3GB Hard disk. I installed Windows98 on this in compact mode and it ran insanely fast. opening word without a splash type fast. However Win98 would do it's normal random crashing for no actual reason. I later installed WindowsXP on the same machine and though it did lose its' snappiness it was usable due to the oddity that WindowsXP runs better on a PentiumII than a PentiumIII (a lot of people attest to this)

At the end of the day, XP crapness aside, Windows98 is not better. Fact.

Most visited site in the UK

[smallguy78]

I've read the bbc news website is the most visited website in the UK, so it's probably the best indication of what UK people use to browse the web. I wonder how many of the IE stats are Opera however.

Re:Most visited site in the UK

[Vo0k]

All the fake "I'm MSIE" Opera ID strings contain "(Opera)" by the end, so any self-respecting stats program should register them correctly.

Re:Most visited site in the UK

[Anonymous Coward]

"I've read the bbc news website is the most visited website in the UK" ... and the article is about the beeb homepage. Damn shame. I'd really like the browser stats for the BBC news. I suspect the results would be rather different and more representative. It'd be nice to see if it's true.

I wonder if there's a way to convince the cash-strapped beeb that regular posting of this sort of information is a worthy news item and public service? It'd sure cut a lot of FUD among developers, and the point of the web is everyone in a developer.

a start at Overall Market Share

[twitter]

Actually, the BBC numbers show that IE has LESS than 85% of the market. The author was unable to identify or otherwise did not count 5% of the strings, non of which is IE. So IE has 85% or 95% of the visits to BBC, which is 80.75%. I'm waiting for it to dip lower than 80% on such a huge site.

It looks like the business world is learning. The BBC is a work safe site, so statistics should be dominated by corporate desktop visiting where the user has no choice of software. That's good news for everyone.

These statistics can be skewed by the botnet. The kinds of people who use botnets would be very sad if the world dumped windows, so you can be sure that DDoS attacks use a standard IE string when they are not busy taking sites down. Microsoft, judged by their record of dirty tricks, might even pay them to do that.

Re:I run a website with 20% Firefox usage as of no

[YA_Python_dev]

(Sorry for the bad formatting, why can't Slashcode support the pre-tag?)

The table, reformatted:

Month| I.E. | Firefox | Safari | Mozilla | Netscape
May -| 64.2 | -- 16.7 | -- 5.3 | - - 0* -| - - 1.3
June | 64.7 | -- 17.9 | -- 5.4 | - - 0* -| - - 1
July | 65.9 | -- 15.6 | -- 4.6 | - - 3.8 | - - 2.1
Aug -| 66.8 | -- 16.6 | -- 5 - | - - 2.5 | - - 1.6
Sep -| 62.9 | -- 19.2 | -- 5.9 | - - 3.2 | - - 1.7
Oct -| 59 - | -- 20 - | -- 5.8 | - - 3.9 | - - 2.4

* Before this date, Firefox & other Mozilla were lumped together.

Re:I run a website with 20% Firefox usage as of no

[leomekenkamp]

IAAKW (I Am A Karma Whore):

Month  IE    Firefox  Safari  Mozilla  Netscape
May    64.2  16.7     5.3     0*       1.3
June   64.7  17.9     5.4     0*       1
July   65.9  15.6     4.6     3.8      2.1
Aug    66.8  16.6     5       2.5      1.6
Sep    62.9  19.2     5.9     3.2      1.7
Oct    59    20       5.8     3.9      2.4

Under [ ] No Karma Bonus and [ ] Post Anonymously there is a dropdown which you can change to 'Code'; your post will be formatted without the need for <br/>, <p></p>, etc.

User Agent Switcher

[GotenXiao]

Thanks to the awesomeness of the Firefox extension model, you can dynamically switch your user agent to whatever the hell you like.

Sploooooooooooo!

Re:User Agent Switcher

[Vo0k]

Yeah, switch to "googlebot" and have free access to all these pay-for-registration sites.

Re:User Agent Switcher

[F�an�ro]

that used to work, but nowadays it seems most sites are using IP or some other criterion too.
Can anyone confirm this? OR am I using the wrong useragent?

Re:User Agent Switcher

[m50d]

Thanks to the awesomeness of the Firefox extension model, you can dynamically switch your user agent to whatever the hell you like.

Or, if you use a real browser like Konqueror you can just do it without having to hunt down an extension to let you.

Re:User Agent Switcher

[YA_Python_dev]

... am I using the wrong useragent?

The googlebot UA string currently is (remove the space in "ht tp"):

Mozilla/5.0 (compatible; Googlebot/2.1; +ht tp://www.google.com/bot.html)

or:

Googlebot/2.1 (+ht tp://www.google.com/bot.html)

Yahoo:

Mozilla/5.0 (compatible; Yahoo! Slurp; ht tp://help.yahoo.com/help/us/ysearch/slurp)

MSN:

msnbot/1.0 (+ht tp://search.msn.com/msnbot.htm)

Re:User Agent Switcher

[Anonymous Coward]

I've seen googlebot two hours ago with: "Googlebot/2.1 (+http://www.googlebot.com/bot.html)", this came from 66.249.71.69, which actually belongs to google. At one time, I've seen two googlebots spider our sites, one with the googlebot.com URL, the other with google.com, and otherwise identical UA-strings. Does anyone around here know what this difference is about?

Re:I run a website with 20% Firefox usage as of no

[shaka]

Ok, didn't know that. Thanks (this goes out to YA_python_dev too)!

Now we can discuss the actual numbers! IE UNDER 60%!!!

Plus, there is Self-Selection

[Anonymous Coward]

Another thing that makes the stats suspect is self-selection.

For example, the BBC's choice of RealMedia, for their sound and video, drives away a lot of the potential Linux and Firefox visitors. They would do better, in terms of balanced statistics, if they included mp3 and mpeg, or ogg formats.

Respectable

[twitter]

There's nothing respectable about that [10% market share for firefox].

First, in a free world, you would expect Firefox to get about 33%. Why? Because there's at least two other good browsers out there, and most run on Windoze. If users were informed and were able to chose a browser that they actually liked, you would expect them to move to one or the other of these based on the quirks of each and personal preference. As Microsoft dies and user choice improves the net will move towards a mix of standards based browsers and none will gain more than 33%.

Second, do realize the enormous effort required to break out of the Microsoft vendor lock in? BBC is a work safe site, so it's statistics are dominated by corporate desktop browsing. Big dumb companies buy from other big dumb companies, like Dell. Microsoft makes sure that companies will pay through the nose to have any non M$ approved software on those computers. So, the 80% market share IE shows here is a major step towards user choice. Each one of those browsers is a determined punch in M$'s face.

POINT OF THESE SURVEYS?

[Anonymous Coward]

WHAT IS THE POINT OF THESE SURVEYS?
 
I use what I use - I don't give a stinking monkey what other people are using. Why would it interest me?

Re:POINT OF THESE SURVEYS?

[Sockatume]

You could just, y'know, not read it. Instead of reading it, then taking the time to post a comment about it.

Default?

[zaguar]

Remember when you first install firefox, there was this shing RSS bookmark with 'Latest Headlines' and pointing to the BBC news pages?

Anyone considered that, maybe, that might have influenced the results? Having a default bookmark as the page of the study? You wouldnt take browser results from MSN.com or whatever IE's default home page is.

Nevermind me though, I just suggested that a pro-Firefox poll might be biased. Karma be dammed!

Re:Default?

[Anonymous Coward]

But those don't touch the BBC homepage. They go straight to the news. They arn't necessarily counted in this, except for people who eventually navigate to the homepage.

Re:Default?

[Anonymous Coward]

Nevermind me though

Good advice. If you would RTFA and think a little bit then you would realise you were talking utter BS.

Re:Default?

[BasilBrush]

An RSS bookmark on news.bbc.co.uk would have absolutely no effect on this count of accesses to www.bbc.co.uk home page, no.

Is not having a fucking clue an essential qualification for being an IE fanboy?

Re:Default?

[zaguar]

Are there any links from news.bbc.co.uk to bbc.co.uk?

Yes.

Nelson Ratings (was Finally....)

[mitcheli]

"If you want to know OS stats, browser stats, or anything like that, you need to conduct an actual survey, and not simple observation of HTTP traffic, because if you are doing the latter, you might as well make up your numbers based on your best guess, because it has just as good a chance of being as accurate."

That's funny... Nelson has been telling what show's are the most popular for years and they've never called me to ask me what I'm watching?

Re:Nelson Ratings (was Finally....)

[fossa]

What, they know the frequency to which you are tuned? I thought Nielsen paid people to put a special box between the cable and the TV... in other words, "took a survey".

Re:Nelson Ratings (was Finally....)

[joeljkp]

I got a Nielson survey once. It was just a 'viewing log' that asked you to write down what you watched and when for a week.

I donno about this...

[oztiks]

I run a web hosting business and here is the breakdown for webstats last month, and yes its linux hosting btw:

Windows 16400 70 %
Linux 5497 23.4 %

MS Internet Explorer No 14012 59.8 %
Firefox No 7579 32.3 %

Though windows is the dominator in this respect but it goes to show the website content does really depend on who visits the site and therefore produces the stats.

Lets face it BBC is a news network business people and general interest users are reading these articals; ofcourse nerds and geeks are but on the scale of things we really only make up that 9% :)

That seems about right

[Cyburbia]

The BBC stats aren't that far off from the stats for my planning-related Web site. Considering how many people visit the site from their workplaces, I'm really surprised to see a somewhat higher percentage of alternative browsers than the Beeb.

MS IE 81.9 %
Firefox 10.4 %
Safari 2.2 %
Netscape 1.6 %
Konqueror 1.1 %
Mozilla 1 %
Unknown 0.8 %
Opera 0.6 %
Camino >0.1%
I-Mode phone >0.1%
Others >0.1%

Re:That seems about right

[dbmasters]

I run a reasonable popular home studio recording community and FireFox is as high as 13% on that site with IE shrinking down to 60% this month...pretty damn excited to see IE shrinking at that rate.

Re:That seems about right

[belmolis]

You can see the disparity caused by different audiences in some sites that I run. My personal site's stats look like this:

Firefox 48.3
IE 24.2
Mozilla 9.2
Unknown 4.9
Safari 4.1
Konqueror 2.6
Opera 2.4
Netscape 1.1
Galeon 0.5
Others 1.4

Another site that I maintain is quite different:

IE 71.7
Unknown 10
Firefox 7.8
Safari 3.6
Konqueror 2.1
Mozilla 2
Netscape 1.1
Links 0.3
Opera 0.3
Others 0.6

The difference is that traffic for my personal site comes primarily from people downloading free software (most of which will run on systems other than GNU/Linux but is known mostly to Linux-oriented people), whereas the traffic for the other site, which deals with the native languages of British Columbia, presumably has no such OS or geek bias.

A third site that I manage, for some friends' fishing lodge, looks like this:

IE 74.8
Firefox 17.6
Mozilla 5.2
Unknown 2.3

If anything, this site will probably be the most representative since the ydli site probably attracts a geekier class of person. Unfortunately, the number of hits is much smaller so that stats are probably not terribly reliable.

Comment removed

[account_deleted]

Comment removed based on user account deletion

older stats

[TheRealDamion]

I used to be about 60% Netscape 35% IE 5% misc like lynx etc, when I started in 1998.
Then it got to about 95% IE, so 85% is quick a marked drop in IE support.

Do remember that the BBC is hardly a generic site for your average Internet user, it attracts a significant quantity of beginners and is dull for anyone technical (there are a higher proportion of technical users on the Internet than you'd meet on a street). So these stats are quite good.

I know the way they are worked out should be quite fair.

Re:older stats

[ZachPruckowski]

Dude, you want older stats? I got them right here. ::Reaches into coat, pulls out a cool-looking book:: This is the sacred wikipedia. It draws on the learning of thousands of bright people, and changes with the times. Use it well.

Ok, enough silliness: http://en.wikipedia.org/wiki/Usage_share_of_web_br owsers/

But the FP is still bigger than it need be.

[burnttoy]

Look at the source. As I've pointed out to the BBC (several times) it's full of junk. about 30% of it is useless white space. Then there's the overly long URLs, comments (why!). In total it could be about 20K smaller. Or to put it another way, the pictures could be bigger and you'd notice little difference in download time.

I even offered to fix it for them (i have the scripts here) but finally got redirected to their "careers" page... nothing like being proactive is there... hmmm....

So, i gave up and wrote my own rss->html front end in PHP instead. It's still a WIP, I have a newer version that aggregates feeds and organises by pub date. I'll stick it up soon and _yes_ I will opensource it.

http://www.burnttoys.co.uk/rss.php?url=http://news rss.bbc.co.uk/rss/newsonline_uk_edition/front_page /rss.xml

as you've probably guessed the bit after url= is user definable.

mode=whore... anyone looking for an HTML/PHP/PERL/XML/SVG/C++/ASM etc coder? ;-D

Re: RSS Syndication

[EddyPearson]

The great thing about the BBC is all their stories are available syndicated into RSS, and of course as a Brit and a Firfox user i have it in the in-build aggregator, so unless BBC count every RSS clickthru as wee, then the figures will DEFINATLY be off, because remember, IE doesnt have an aggregator, until crappy vista (god, it sound like a feminine hygine product!) comes out the vast majority of frontpage hits, will be through MSIE.

Trend matters, not snapshot

[fnurb]

The relevant information is not the raw number, but the trend. If you see Firefox gaining 1% every month of so, then is is reasonable to conclude that Firefox is gaining marketshare--in fact, it is even reasonable to assume that that gain is about 1% per month, since statistical anomalies and distortions caused by "AOL tweaking their cache configs" averages out to noise in a long-term trend.

While you are right that an accurate snapshot is impossible, snapshots only matter to magazine writers facing a deadline. In both the economic and intellectual marketplaces, what matters is the trend.

Stats for my site

[Anonymous Coward]

2% IE, 1.7% Firefox, .01% Opera, 60% MSN spider, 30% Google Spider, 6.29% Yahoo! spider

*sob*

Re:LATE BREAKING NEWS!!!

[julesh]

Is the BBC homepage supposed to reflect some important or signifigant user base?

Yes. It is probably the broadest cross section of mostly British web users you are likely to find on a single site.

The fact that nearly 10% of those users use firefox is particularly relevant, and is a good weapon for those of us who do commercial web design to persuade our clients that the extra work to support alternative browsers properly *is* worth it.

Those figures aren't influenced by ff's rss

[mjmartin_uk]

IMHO Those figures aren't influenced by Firefox's 'Latest Headlines' because that RSS feed points to article pages, not the homepage, these results were from the homepage therefore anyone that says that FF's 'Latest Headlines' unfairly influences the results should RTFA!

FWIW Westminster Abbey's stats are (taken 16-10-05 to 22-10-05):
IE Variants: 65.9%
Netscape variants (mostly Firefox and other Geckos): 24.9%

This is compared to 6months ago (taken Jan 2005 - Feb 2005)
IE vars: 69.4%
Gecko vars: 21.1%

And finally last year (taken May 2004 - June 2004)
IE vars: 80.9%
Gecko vars: 11%

Our results do tend to suggest that Firefox has gained a significant market share from Microsoft's browsers.

Ipod Studio

[Thabenksta]

I'm the webmaster for http://ipastudio.com/ and out of the 10,000 visitors we have a day, 50% of them or more use Mozilla/Firefox.

Re:Ipod Studio

[Anonymous Coward]

50% Firefox? What about Safari? On a site dedicated to iPod gear, I'd expect a few Mac users...

9.7% isn't that un-usual

[Anonymous Coward]

I work for a restaurant search engine company (www.dine.to) in Toronto, Canada, which is aimed about as far from web-hacker as possible. Firefox gets about 9.5% of the visitors. Like the story says, not a massive amount, but still respectable.

Yet another Opera is undercounted theory coming!

[porneL]

IMHO Opera stats can be skewed in a different way.

It is plausible that some IE users have BCC homepage set as browser's start page and create large number of hits.

but other browsers have alternative mechanisms, that allow user to visit homepage even less often than usual. For example Opera on each start reopens previously open tabs, from cache, so rarely anyone uses start page feature. Opera and FF have RSS that leads users directly to articles, etc.

All is relative.

[jotaeleemeese]

10% of a market competing against a monopolist (convicted for abusing that monopoly) with the resources that MS has is mighty respectable.

85% too dumb to differ a webbrowser from a shell

[Anonymous Coward]

And about 100% of webstatistics are made of idiots, because they include IE as a webbrowser. Cleaned up, statistics clearly show that about 65% of all users access BBC Online _with a webbrowser_ are using Firefox.

What confuses the author is explained

[Anonymous Coward]

On Macs the author states a stunning figure:

Internet Explorer 30%

"Safari's lead is no surprise, but I wasn't expecting Internet Explorer to have so big a penetration in the Mac market."

Is it possible that Opera reporting itself as Internet Explorer by default has anything to do with this figure?

Re:Opera and lynx

[Anonymous Coward]

I use lynx for automated gets. It visits about 30 websites, reads posted content, completes search forms and parses results in a nicely formatted wordpad document... in about a minute.

anyway, I changed the useragent to Mozilla so I assume I am logged in with IE and also misrepresented in the data. I do this to blend in.

Competition Commission Now!!

[Trogie]

In 2000, a Competition Commission found that Interbrew's (now Inbev) acquisition of Bass Brewers was posing competition concers for the British beer market. Interbew would have a combined 32% of the British beer market!! ONLY 32% and they put up a competition commission!! Why not investigate M$ monopoly??

Re:Competition Commission Now!!

[Trogie]

http://www.realbeer.com/news/articles/news-001541. php

Wait for October for movile phones

[adelayde]

"but interestingly there's no mention of mobile phone devices"

Wait for October's stats, I used my Nokia 6630 to access the BBC's News site only the other day and very impressive their 'mobile' version of their site is too. The Beeb should get an award for a great attempt at accessibility and browser compatability.

On the street, sure

[iabervon]

Of the people using any operating systems when you walk up to them on the street, 1 in 11 probably will be using Linux. It's a reasonably common embedded OS for smart phones, although it's completely hidden from the users. If you're making a phone with an entirely custom UI, Linux provides a nice programming environment, and you're not losing anything you want, anyway.

legacy OSs

[belmolis]

My personal web site gets on average one or two visits a month from systems identifying themselves as running CP/M. I always wonder whether there really are people still running CP/M and surfing the web or whether these are just joking misidentifications.

Re:LATE BREAKING NEWS!!!

[koonat]

"Yes. It is probably the broadest cross section of mostly British web users you are likely to find on a single site."

That doesn't mean ANYTHING though. PROBABLY ... LIKELY to find... whatever.

These results would be a LOT more relevant if they came from a website that people make use of, actually visit. The website of a major media company is a poor reflection of the public. You show me the statistics from google, yahoo, amazon, ebay, or some other site with an actual diverse userbase.

Linux More Popular Than Windows 95

[djKing]

And it only took 10 years!

Re:LATE BREAKING NEWS!!!

[Anonymous Coward]

You think BBC News isn't a website "people make use of, actually visit"? You need to get out on the internet more.

Yeah, you didn't read that correctly.

[heinousjay]

Slow down, Cowboy! Slashdot doesn't require that you check yourself, but it should have, because you wrecked yourself.

Re:Yeah, you didn't read that correctly.

[heinousjay]

I take that back - I wrecked myself. DOH.

Mod appropriately, please.

Spider analysis missing

[gbartoli]

I used to analyze corporate web metrics and my comment about this article is that a bit more information should have been provided about the methods that he used to filter spiders and other automated retrieval systems, especially if you are going to make your own claims about browser/operating system share.

I have seen as much as 30% of activity on a single page be due to some some sort of automated retrieval.

It looks like he paid quite a bit of attention of the wide variety of user agent strings (kudos!) and I'm guessing that much or all of his filtering was based on his analysis of these strings.

Unfortunately, this is not enough. Many spiders identify themselves as some Windows/Internet Explorer variation without any indicator that they are a spider, business intelligence application, or other non-eyeball-on-the-page type browser.

One lo-fi suggestion for folks who want to check their current spider filter setup is to create a temporary honeypot link on one of their pages. An invisible hyperlink to a non-production page using a 1x1px transparent gif is enough. Then run a metrics report on that non-production page and include the IP addresses (watch for entire subnets of crawlers) in your production filters.

Also, I assume that whenever the author says "requests" he means "page views." It is good to make that specific distinction and focus solely on page views only for any kind of metrics reporting that is about eyeballs rather than server load. It provides for apples-apples comparison between browsers that might be more conservative at caching and re-using images, etc.

cheers

Other browser ID's

[solune]

I'm using Opera, and have the default set to identify itself as I.E. Certainly this messes with the count. Granted, Opera probably doesn't get as much press as others, but I do know Konqueror can be set to identify itself as another browser, Mozilla among them. In other words, I find this statistic useless because it's not *really* able to identify what browser is being used, only what browser the user says is being used.

ugly

[einolu]

*comic book guy voice* ugliest website ever

Firewalls Don't Browse Much

[billstewart]

I was surprised not to see BSD listed at all, and *somebody* had to make the "BSD is Dying" joke. But I suspect there are several things happening that account for its invisibility:

• Maybe it's lying about what OS it really is? Not an uncommon thing for browsers to do, as well as lying about what browser type they are.
• Lots of BSD use, especially OpenBSD, is for appliances like firewalls that aren't running browsers very often.

Re:Firewalls Don't Browse Much

[Anonymous Coward]

I access bbc.co.uk from a FreeBSD box regularly. The UA string reveals BSD.

Not accurate for other reasons

[LPetrazickis]

Opera caches every document and lets you go back as far as you want without reloading. IE reloads every time the back button is hit. Each reload counts as one or more unique hits. This means Opera will always be undercounted in any stats analysis that goes only by hits.

Well, I run a website with over 27% Opera usage!

[LPetrazickis]

My stats. The thing is that I use my wiki to take notes, so I am always there myself, and I use Opera whether at home, at school, or at work.;)