Source for browser statistics

any suggestions?

[tmricha]

On the last redesign of our corporate internet site, a coworker and I had a disagreement about what level of captability the site should be designed for. She argued the "corporate standard" was IE5...etc etc...Netscape 4.x. The disagreement came from what the 4.x included. In theory the X should include everything from 0-9, but I argued that 4.7 was a much more heavily used browser and was much more feature rich than the previous main version of 4.2. Because of design requirements for what was wanted on the site, I argued that programming for browser checks and compatibility to 4.2 would be way to costly/time consuming. What was wanted could be done for down to 4.7 in a matter of weeks, but to configure compatibility down to 4.3 could take months. I did some *brief* research trying to find browser usage (specific enough to distinguish between 4.2 and 4.7), but had no luck.

So my question is this:
Does the wonderful webmaster world community have any suggestions or known locations for such statistics on a major (approaching global) scale?

Thanks for all of your help

[photon]

Try http*//www.thecounter.com/stats/

[DrDoc]

Be aware, however, that you can not trust the stats 100%. They are skewed and do not say anything about which browser is more used. The only thing the stats tell you is which browser most people use to view pages where the counter is present. Plus, I don't know if I would trust those stats anyway, since they contain browsers that never existed.

[photon]

As far as I know, most browser tracking statitistics rely on the user agent reported by the browser itself. That, of course, can be made to pass along anything the (savvy) user wants it to be. For example, Opera defaults with its user agent set to MSIE 6, I believe. Therefore, all such stats need to be taken with a grain of salt.

Having said that, I think it's safe to infer the level of usage for "legacy" browsers (the 4.Xs) from user agent based stats. Anyone still using those probably isn't a sophisticated user who would change the user agent (not saying that they don't exist, just that it's unlikely).

Is there a way of determining what the actual browser is without relying on the user agent?

<edit>speeling</edit>

[edited by: photon at 9:46 pm (utc) on Nov. 26, 2003]

[DrDoc]

Short of sitting right next to the surfer, not that I know of. All stats are based on user agent strings, or combination thereof.

Hmm... Maybe I should make my Opera 7.22 identify itself as Netscape 2 more often ;)

[carolynf]

Another factor is your customer base. Are they consumers or industrial? How would you categorize them economically? In other words, if they are industrial, are they mid-large entities with specialists who keep office browsers up-to-date? If that's the case, I believe you can design for pages beyond 4.x.

[DCHWeb]

Here's my input on this subject:

If a user is savvy enough to change the user agent, there is usually a reason for doing such things, like getting different pages than intended for dev purposes, and therefore know enough to understand why your page wouldn't view correctly, and change it. Unsavvy users, however, tend to fall into two categories, "conservative" and "liberal," meaning that they either like to change because their computer tells them too, or they never change because they are afraid of a crash, because their system has worked fine without the update. The "conservatives," though my stats, collected by myself, show that there are fewer of them, are the ones you need to worry about. I make a page that tells them why they need to change, and how "facinatingly easy" it is, and more than half of them do.

I also have had very few users hit my site that are not using a version below the 5.x series browsers.

[hartlandcat]

There's no such thing as Netscape 4.2. The 4.x line of the Netscape browser went 4.0, 4.5, 4.6, 4.7, and then finally several years later... 4.8.

Netscape 4.5, 4.6 and 4.7 are almost identical in the way they render pages. 4.0 has a slightly earlier version of JavaScript, and 4.8 has something fractionately different about CSS. However, don't worry about Netscape 4.8, it was released in about November 2002 as a very last minute bug-fix to Netscape 4.79 -- it wasn't advertised in favour of Netscape 7.0, and you'd have to actually go looking for it to find it.

I find it hard to believe that different 4.x versions of Netscape are going to take so much longer. Make your site degrade gracefully in Netscape 4 (by giving them only a very basic stylesheet), and then don't bother anymore. It is possible to create workarounds to get even the most complex pages to work in Netscape 4... but the result is something that looks horrendous in just about everything else.

Unless you're making a site for people in schools etc., your site should only get about 1% Netscape 4 users. Most "Netscape" users use 7.x, or Mozilla (myself included).

[-remembers that I've just switched to Safari yesturday-]

[Zaphod Beeblebrox]

96,5% of the visitors to my sites use IE...I don't bother about Netscape or other browsers.

[Farix]

I use Mozilla 1.5 as my primary testing browser, but I check with IE. Most of the time, I don't need to fix any IE problems.

[DanA]

I know why I am excluded from some sites : I use Firebird or Mozilla with Windows or Linux.
Only 66% of the visitors of my site use MSIE (6,5.5 or 5) as the rest are welcome and get pages according to their favorite browser.
I use Mozilla to build the pages and I test with MSIE, but not to exclude the others I also test with Off by One and Konqueror.
This often leads to many fixes.

[mcguffin]

This year, I spent the majority of the year consulting onsite at a Fortune 100 company. Their corporate standard browser remains Netscape 4.7.

They were taking brand new P4s loaded with XP, and reimaging them back to Win95 and Netscape 4.7 at operations sites across the U.S.

So, even when you're talking about major corporations, there's a lot of legacy applications still out there.

[edit: fixed typos]

[Farix]

Windows 95? Blah! What the heck is causing them to downgrade that far back? Do they have an essential application that just will no run in W2K (much less Win98) or is this just an example of corporate inflexibility?

[Distel]

So, even when you're talking about major corporations, there's a lot of legacy applications still out there.

Although I have no numbers to support it, and it's just my personal thought, I think that major corporations are probably even less likely to switch to more recent applications and systems. They often are not flexible enough, and not willing to spend so much money in what they see as a very low-priority upgrade. I currently work for a large multinational, and most of our systems predate the dinosaurs.

[mcguffin]

Farix, my impression was that Win95 was chosen as a platform many years ago, and the decision had become a part of corporate culture well beyond the OS' functional lifespan.

[Farix]

In other words, it's corporate inflexibility. I can understand not upgrading software on old machines. Even keeping a few old platforms for legacy software. But downgrading software on new machines is something I have a hard time understanding.

[victor]

tmricha,

You need two things....

First: browser stats for your site. So check your own logs.

But don't be fooled by aggegating all pages. An apparent count of "95% of all visitors are IE" may break down into:

75% of visitors to the home page are IE

90% of visitors to the product list pages are IE

100% of visitors to the payment page are IE

What that would mean would need further research, but one obvious possibility is that you are shedding 25% of potential sales while thinking you are catering for 95% of your market.

Second, you need the stats for your competitors. These are harder to come by, but (one possibility) is via any trade association you have.

Any significant differences in a competitor's stats and yours may again need some further research. If they are reporting 3% cell phone visitors while yours is 0.005%, it may be that they have the future sewn up while you believe (from your stats) that PDA/WAP/etc isn't worth bothering with

[jbinbpt]

corporate inflexibility?

Don't fix what ain't broke.....

I only get to replace machines as they die. We run them to death. And then some.....

I'm still running two dos 3.2 machines for a barcoding application. I can't kill them.

[iceman_ertw]

The source for browser statistics can be obtained at bravnet.

[Distel]

In other words, it's corporate inflexibility. I can understand not upgrading software on old machines. Even keeping a few old platforms for legacy software. But downgrading software on new machines is something I have a hard time understanding.

True, but the problem is (as mentioned by jbinbpt) that new machines are often a rare thing in big companies. It costs a fortune to replace so many machines, and such decisions are mostly taken or approved of by the upper management, which often seems to think that those old machines are still more than capable enough of getting the job done.