Remember that TCO story?
You see, THIS is the kind of crap you need to add to Windows TCO cost calculations. The never ending absorption of bandwidth, the incessant mothering of systems so they stay more or less up, the endless streatm of security problems, the ceaseless interruptions be[Windows would like to reboot, Yes/no]cause updates need it (apps, OS, Java, anti-virus) - it goes on and on. A Windows based platform appears to spend more time coming up with excuses to interrupt people than to do any work, and this is called "enhancing" productivity?
Add to that the compulsory change of user interface with every release with the promise (but never delivery) of better productivity and it becomes really, really hard to defend not switching.
Retraining? What exactly did you have to do switching Vista? Office 2007? How much time did your tech staff spend looking what new devils they had to fight now? Have you found the "insert field" function in Word yet (hint: it's not in the ribbon)? Only in select cases has the upgrade been justified (Excel acquired some decent tools - if you can find them, and when you realise that you may need to switch them off again).
New equipment? Well, no, not for "that Linux thing" - those people don't code with the assumption that crappy, inefficient code is masked by throwing new hardware at it. And they have heard of async coding and real multitasking so .. the .. ma.. ch .. in .. e.. doesn't sl..ow down because you opened another app.
Security? Segregation is part of it's heritage, not imported later. No, it's not perfect either but you have a much longer run up time before it gains prominence as a platform to hack (it's also harder). What do you think you could do with, say, a year of uninterrupted staff productivity?
So there. With a honest TCO calc the picture may look bleak for the continued use of MS products. Aren't you glad nobody does them?