When a Windows program crashes, Windows XP gives you the opportunity to send an error report to Microsoft. The process is called Online Crash Analysis. My advice: Do it. Here’s a perfect example of why it’s good for you and for your fellow PC users.
For years, I’ve encountered a sporadic problem with Word. The conditions that lead to the error are easy to identify, although I’ve never been able to reliably reproduce it. Basically, if I cut large blocks of text from a specific type of document (something I do fairly often when writing a book), I run a serious risk that Word will crash with an error message that points to “stamp 424d964d.” I’ve Googled the error and found that other people have it too. But I’ve never found an answer, and I’ve learned to be extra careful when doing mass cut-and-paste jobs. (The good news is that Word always recovers my documents perfectly, and I don’t lose any data – only some time as I reopen and repair each one.)
Today I encountered this error several times, and on the last time I decided to send in an error report. I’ve done this a hundred times or more before, but this time the response I got was different. After the error report finished sending, I was greeted with this Web page:
[Click image to see a larger view.]
Some engineer, using the great big bucket of error reports sent in from all sorts of people all over the world, found the problem, which turns out to be a bug in Word 2003. A fix is in the works. When the fix is ready, the Online Crash Analysis page will offer to install it for me when I bump into this error and send in a report. Imagine that!
Update: Turns out others have been talking about this lately as well. Read this, and this, and this, for instance (thanks, Nicholas). And this informative post from Chris Pratley (thanks, Zaine).
Ed Bott 
http://blogs.msdn.com/larryosterman/archive/2004/06/01/145627.aspx
Straight from the horses mouth, on how they review the data sent in, etc.
On his blog, Chris Pratley (who worked on the Office team for many years and is now with the OneNote team), noted the following:
The remarkable thing about Watson …is that it is a new way to make products better that tries to measure the real world. It is not an excuse to avoid good code architecture, or to not think about product design. One person pointed out that non-crashes can be more annoying than crashes — that’s very true — although crashes and hangs are pretty nasty. In fact we’re looking into ways to extend Watson beyond crashes and hangs. The Watson guys have a bunch of exciting ideas around that…. the goal is to ship a product with the highest known quality, not necessarily with fewer bugs. This is a counter-intuitive concept, so I’ll explain a little more. The naive way to think about bugs is that if you fix a bug, the product is better as a result. The truth is that it probably is better, but you cannot say for certain right after you check in your change that it absolutely is better. You may have inadvertently introduced a different problem by fixing the issue you were dealing with. That kind of bug is called a “regression” — the product quality has regressed (decreased) as a result of trying to improve it.
As a result, advanced users like myself were impressed with the stability of Word 2003 over the 2002/XP version. So because they are “real world” reports on what most users are experiencing, they gain priority for repair. Report them!
Here here. I was having problems with an analog modem in another machine – from time to time I’d get a BSOD. For ages I’d submit the reports to MS and have no response, but then one day IE popped up and said that this was a driver error and that all I had to do was download a driver update from the modem chipset manufacturer – it even gave the link. I did, and the BSODs went away.