Why IE6 still matters, even if you do happen to be a cross browser testing company!
This post should more properly be titled 'How could we be so arrogant' - read on to discover why.
Two weeks ago, one of our customers emailed and said some of the developers on his team could not launch a live test or an automated screenshot test - they did not even see the buttons! In questioning further, he shared that they were accessing the system with IE6. Now, I have to admit... I did not think there were ANY users of our system that access it via IE6 - these are web designers and developers after all - surely they all are using the latest, greatest verison of Firefox, Chrome, or Safari? My partners shared that conviction. While we did try to make sure the site functioned properly in IE6, we did not worry about the appearance. We had some pretty major flaws on our home page with rendering in IE6, which we were okay with - because NOBODY (ok, no designer / developer) uses IE6. Since none of our customers use IE6, we focused our energies on things that mattered.
Except we were wrong!
Checking Google Analytics feed us some humble pie. Here are a couple charts looking at the 'logged in' area of our site, specifically the 'My Testing Center' page. First, overall browser usage:
As you can see, roughly 50% of our customers are using Firefox, 29% IE, and 10% Safari. This fits with our assumptions about the makeup of our customer based. Heavily weighted by designers, the percentage of Firefox users is vastly greater than a 'normal' consumer site's makeup. Drilling down further on the Internet Explorer users, we can look at the customers accessing the My Testing Center with IE6:
A fifth of our customers that are IE users are using Internet Explorer 6! And we were providing them with a layout that did not render properly, and, in this case, a broken product... NOT GOOD.
It only took a few minutes to correct the issue on the My Testing Center page which was preventing the customers with IE6 from testing with our service. We had done a redesign of some of the blocks on the My Testing Center the previous day - these changes were what caused the problem. And, in light of our little wakeup call, we are in process on redesigning our home page so it renders properly on IE6. We still have a little work to do, but here are some before and after screenshots:
As you can see, the original design's center content area was wrapping, leaving a large unbalanced blue area on the right side of the layout. It also had several other issues. The new layout still needs a png fix done on our logo and some padding work, but overall it is much more presentable.
BTW - this issue lead us to add capability to take automated screenshots of areas behind authentication to the CrossBrowserTesting.com system. This new capability will help us easily screen for issues on our pages that require user sign in. We believe it will also help our customers.
Lesson learned - cross browser compatibility is important, even for a cross browser testing company!
Ken, Mike and Tony
CrossBrowserTesting.com
Pick an OS | Pick a Browser | Test a Website
(we may change this tagline to "We can eat crow with the best of them")
Delicious
Digg
StumbleUpon
Propeller
Reddit
Magnoliacom
Facebook
Google
Yahoo- ken's blog
- Login or register to post comments
Comments
I think your conclusion is flawed.
I expected some insider stuff on the prevalence of IE6 that might change my mind about putting her to rest, but your info here seems to make the opposite case. 1/5th of of 1/3rd isn't a big number - around 6%.
Then consider that we are opening ie6 in order to test IE6 - after all, yours is a cross-browser testing site that's metered. I know I personally use my own IE6 while I'm using your site sometimes since I need to hack around the most in it and it would be too expensive to do that on the clock. That doesn't mean I'm an IE6 customer. I just happen to have it open and might hit your site by accident with it. I could guess that could easily make up 3% of your hits, with another 3% actually being ie6ers. I'd gamble you have more genuine Opera traffic than IE6.
You could do a poll to ask a sample of your ie6 visitors if they are primarily IE6 types.
Consider then also that roughly half my time in cross-browser support work is spend on Ie6. 1/2 my time for 3% of visitors isn't worth it. Think of what 37 signals would say...I might also point out the famous 80/20 rule of business here. 80% of your time is pent catering to 20% of your customers. When a browser reaches that number, I drop it.
When we all started going HTML 4x and dropping IE3, IE3 was around 20% use. When we dropped NS4.7, it was around 17%. We did this to encourage the clingers to ditch these and most of them did once sites stopped working.
Then also consider the responsibility we face as developers and designers to encourage visitors to embrace more compliant browsers.When the industry tries to move forward, don't we have some obligation to help? Coddling IE6 is probably the best way to cave to customers, but it only leaves them with a browser that will serve increasingly broken sites. It may be a better service to them to encourage them to upgrade.
And for your service, I can't help but doubt any of us web developers are genuinely browsing primarily with IE6.
High Tech Creative Services
Techism Seattle
Good points. Thanks Ken
Good points.
Thanks
Ken