Unfortunately, I have to deal with a much different picture in my case. The web application is often used by either complete computer illiterates or users in large corporate installations (or both).
My IE numbers (looking at the three biggest installations):
I nearly fell off my chair when one of our new enterprise customers asked me if we could provide them the app... but with Firefox support because that was all they used.
I worked for this large company, who had just bought another smaller company. Wisely, the smaller companies IT department had moved everyone onto Firefox, but since they had to have access to our new system, which was IE, we had to tell them to move back to painsville.
It broke our hearts, the IT people, and the users. The new system will work on all browsers (except IE6), whenever that will be out.
Out of curiosity, did you look into using something like IETab? You can configure it to render specific sites in IE's Trident engine, and all other sites using the usual Gecko engine.
However, I'm not certain how easy it is to push a Firefox extension + configuration to a large number of Windows machines.
Chrome is a disallowed program on our network by "Symantec Endpoint Protection". You can install Firefox but it's discouraged by the IT department. I wouldn't doubt if these setups are common among larger corporations.
That's what it's like here. Fortunately, you can just rename chrome.exe to not-chrome.exe, then everything works fine. The "protection" software works just well enough to get the contract, but not so well that you can't work around it.
I suppose the issue was more that there are barriers which require the effort beyond that which the 98th percentile will exert for the benefit that Chrome gives, and therefore 98% of users will stick with what's given to them.
Woobius traffic specs (quite representative of the enterprise space, so more realistic if you're doing an SaaS that also targets enterprise users/customers), are not too far off, but not 50% yet.
In other words, dropping IE6 support is not yet plausible. But we've already opted to provide only functional support for IE6 - i.e., things need to work in IE6, and not look totally broken, but they don't have to look great.
It all depends on how much work goes into supporting IE6. If the amount of time is substantial, personally I would consider dropping it. But even until the bitter end, that decision depends on your target audience and how much you want that group to be forced to make a choice: upgrade or leave your site.
That's for functionality. I think we're long past the stage where you should care whether or not IE6 gets the same exact _visual_ experience.
For us, it certainly does. Our product's success depends on it working "for everyone". If you invite your structural engineer to the project and they can't use Woobius, you'll revert back to email or some other method, and Woobius will have failed. Therefore, if 5.6% use IE6, given that those 5.6% include some of the largest engineering firms in the UK, we have to support IE6.
I run a business blog that is in the Alexa top 10K for US users. It probably leans slightly more techie than average. Most people read it at work. Here are its stats for the past 30 days:
Wow, Google runs ads on Television? I consume all of my video media from the Internet these days so I guess I just didn't know, but that's surprising to me. Do they run the same type of promotional videos that you find on YouTube about Chrome, or are they totally different adverts?
You make a good point, I consume all my television via streaming services, too. It's possible that the adverts are different for that compared to broadcast TV. A cursory web search for Chrome TV ads does appear to confirm they're broadcasting them though.
At deviantART we're dropping IE6 support shortly... it wound up representing a tiny percentage of our traffic, and we like the idea of being able to use the crazy innovations of IE7.
It's slightly painful, because we've been forcing IE into quirks mode for ages (with a comment before the doctype) so that all IEs behave like IE6, and thus we only had to test one IE.
It dropped from 50.17% in March to 49.05% in April, continuing to steadily decrease since then.
My traffic is about 70% North American. I have many students. Macintosh is a 20% slice of my OS pie and Safari isn't too far behind Firefox in browser share.
September => IE: 48%, Firefox: 24%, Safari: 16%, Chrome: 10%
On top of that, Mac users have less need to even think about what browser they use because IE isn't the default, which weights things disproportionately in Safari's favor. Safari is "good enough."
The biggest problem with IE6 that I see with my company, is that a lot of corporate clients still have IE6 on their machines because of their lazy IT departments. So they don't care if only 0.0000000000001% of the traffic is from IE6 because it happens to be them.
There are also apps that rely on IE6 and simply do not work properly in IE7 nor IE8. These apps would also have to be updated which, with tight purse strings in some companies, is not bound to happen soon.
We had this problem with publishers sticking with OS9 and Quark 4 a few years back: that kept them on Mac IE5. In the end we told that that while the admin system would continue to work in Mac IE5 the front end would not be viewable. Didn’t go down too well, but it simply wasn’t possible to provide the layouts they wanted in a format that worked with that browser.
Yep! Started a new job on Monday with a fresh install of IE 6. I couldn't believe it, I'm now working at one of the companies where ie6 is forced upon all staff, even the web dev team. I complained that the majority of users are using a newer browser, and I was told that IE7 is currently being beta tested. I'm not holding my breath.
Another obvious but often overlooked factor that contributed to the decline of IE is the increased number people converting to Mac (anybody have a figure for this?). Because MS discontinued supporting IE on Mac years ago, there's not way to recover users lost to Safari and other browsers on the Mac front.
In my small subset of the enterprise market, amongst my clients (large financial services firms) IE usage is 100%. Unfortunately I don't see this changing any time soon.
Anyone have good stats for the browser-vs-non-browser (i.e. dedicated FB/twitter client, Instapaper, etc.) use of HTTP to consume HTML. I bet you'd find that c. 2002 IE had 90% of the browser and browsers had 99.9% of HTTP+HTML, but that now browsers have more like 90% of HTTP+HTML. Then there's the desktop browser vs mobile browser issue, and RSS ...
The site I work for has been running at under 40% IE for over a year and a half. We're at about 15% Safari, 40% Firefox, 38% IE, and 7% Chrome. IE6 is under 4%, thanks goodness!
We develop our products for the latest version of popular browsers: IE, Chrome, Firefox, Safari, Opera. If a customer has a problem and is using a older browser, we simply ask them to upgrade.
In order to stay modern and write good code with as few hacks as possible, you have to let older browsers go.
The end of life of Windows 2k in July hopefully will mean this will change soon. Thankfully many companies will upgrade once there is no support whatsoever for their current OS.
a blog I run had 125,266 visits from 10/4 - 11/4. The browser usage percentages are as follows: firefox: 41%, chrome: 22%, internet explorer: 19%. Most of our traffic is referred from tech inclined sources. Personally, I have heard a ton of complaints about the amount of memory firefox uses while running compared to chrome...
My IE numbers (looking at the three biggest installations):
IE percentage: 80, 90, 98 - IE6 percentage: 30, 50, 40
No Chrome. No Safari. Rest is Firefox (2.0, 3.0, 3.5 and 3.6).
I weep when I consider the amount of development time we waste catering for these IE users refusing to (or unable to) update.