This should be a familiar topic to most anybody who reads even a few articles a week.  IE6 quirks continue to be a major headache in designing cross-browser compatible web applications.  It takes up a disproportionate amount of time among my QA team relative to the percentage of lingering users.  And I bet the same is true for your organization.  The unfortunate truth beyond the QA sink hole is that IE6 will continue to limit our  innovation process until it dies, or we decide not support it.  The former is not under my control, so I took a quick look at the latter by asking my trusted SiteCatalyst a few questions.  Here's what I found:

What portion of visitors year-to-date used IE6?

22.4% to be exact, not counting the 0.3% using AOL IE6, which SiteCatalyst tracked separately.  That is definitely significant.

Ok, are those the people buying our products?

They sure are.  Clumping the two aforementioned IE6 editions together comprised 20.5% of revenue so far this year.

Well IE8 became publicly available in March so my visits and revenue numbers must be front-loaded in Q1, right?

Nope!  The interesting part of this whole problem (in our world) is that it's  just not going away.  There are lots of well-documented theories on why IE6 lingers on, including this blog post by Mark Trammell at Digg.com and the work being done at IE6NoMore.com.

Let's take a look at the trend using actual data using a SiteCatalyst report.  First, some info about the report:

Scope: Visitors using IE8, IE7 and IE6
Time Frame:  Jan - Aug 2009
Smoothing: Linear trending

So the trend of IE6 visitors this year is holding steady around 15%, although it's tipping downwards slowly.  At the same time, the trends for IE7 and IE8 are inversely proportional, just about 1:1.  This must be evidence to support the primary theory for why IE6 won't die: that the majority of IE6 visitors are corporate users who cannot control which browser they use (see Mark Trammell's blog for some numbers.  It's about 69% if you lump together the top 2 survey results, which I think you can do safely.)

Last question: Do I really need to continue supporting IE6?

Unforunately, SiteCatalyst couldn't answer this question so I'll have to resign myself to waiting patiently for a bit longer.  We've done as other companies have done, which is discontinue supporting IE6 for logged-in users.  But it seems we can't avoid cutting support for anonymous users at our public eCommerce sites for a while yet.