SQL Server Central columnist Steve Jones posted an interesting interview with Bob Dorr, a senior support escalation engineer with the Microsoft SQL Server support team. You really should take a few minutes to read the article because it's worth your time.
I'm going to quickly recap a couple things I found most interesting. Of course, there's the usual biographical things to learn about Bob, of which there are many, such as Bob working on SQL Server as far back as the OS/2 days. Not a lot of guys can say that!
However, I found the organizational details to be even more interesting. For example, I found their organization structure using "pipelines" escalation managers and escalation engineers to be very interesting. I was also surprised that backups & recovery still makes up about 40% of the volume of their support calls.
Another thing he discussed that I found fascinating is that they hired a statistical engineer. The statistical engineer performed a wide variety of manipulations against the data concerning bugs/errors including assigning each incident a value on a customer pain index. While this might seem like a great idea, even an elementary idea, I'm surprised by the number of high-end ISV's who do not employ such a technique to improve their support systems.
Bob also takes the time to speak at conferences, such as last year's PASS 2006 Community Summit. I encourage you to take the time to attend any of his sessions. He's definitely a font of wisdom.