This started as a response to Rob Gonda's "Will Microsoft ever get it? post, but I think there's enough in here to actually build out it's own post:
I tend to break computer users down along the "Neophile" vs "Neophobe" axis. About a third of the computer users out there will use any new technology as long as it's new. These are the people who loved Vista, because it did new things that no OS before had done (or at least, not one that they had used). Most of them will upgrade to windows 7 just because it's new, and they want to try it. Usually within this crowd there's a certain tolerance to bugginess, because the "newness" outweighs the bugginess. In the most extreme case, these folks are Linux users, who run bleeding edge Gentoo releases and build packages of new software whenever they're released. These are the folks who are subscribed to the Freshmeat and Sourceforge new software feeds, and install new software daily. I'm one of these folks mostly.
Then there are "Neophobes", who hate new things simply because they're new. These are the people who say "I've got my toolset, and it works, and it'll take me weeks worth of time to learn something new, so why the hell would I want to upgrade to Vista?" These are the people who were running Windows 98 well into the 2003 timeframe. These people were angry enough that they were "forced" to upgrade to Windows XP, and this whole Vista thing just pushed them over the edge. A few picked up Macs, but then they'll be back on the Windows platform in a few years when Mac OS 11 comes out and they get pissed about the fact that there's now some weird desktop metaphor change that they aren't used to and they'll start reminiscing about good old windows, and how easy everything was there. There's a fair number of these folks in IT too. Ironically, this crowd also hosts a lot of Linux users who are more than happy with their green phosphor command prompt thank you very much. There's also a good amount of bug tolerance in this group because they would rather work with a software package that's buggy, but they know how to work around the bugs, than learn something new.
Then there's the middle ground. I call them the Pragmatists. They don't care whether it's new, or old, they just want something that works. There's almost no bug tolerance in this group. These are the folks who coined the phrase, "Don't use a Microsoft Product until the first Service Pack". This group is actually a fairly small portion of the general IT population, since they tend to get beaten up a lot by the other two factions. However, this group makes up the vast majority of the general, non-IT population. Most of these people won't bother even trying an application unless there's an obvious need at hand for it. Likewise, they won't bother with upgrading their OS until some outside circumstance forces it (They buy a new computer, with the new OS on it, they are forced to by their corporate IT policy, etc..) For most of these people, there's no reason to switch to Macintosh, or Linux, because even the huge changes in Vista don't significantly negatively impact them. Their Office stuff still works, though the toolbars took a bit of getting used to, and their email and web browsing work just fine, and that UAC stuff only affected them like 5 times when they were first setting up the computer, and then it hasn't popped up since then.
The problem with Vista is that Microsoft got too caught up in the Neophile point of view. They made so many radical changes to the OS that it completely alienated all of the Neophobes to the point where even the people who despise change, and want to keep things as static as possible are chosing to change to Macintosh rather than learn the newness of Vista. And of course, whenever you make such massive changes to any large system, many bugs were introduced. This alienated a good chunk of the Pragmatists as well. What Windows 7 needs to do if it wants to succeed is move back towards the middle, to the Pragmatist point of view. New features are great, but don't let them overshadow reliability, and for god's sake don't drop massive user paradigm changes into your most commonly used flagship products first without getting people used to them in other products. The ribbon bar in Office may be a good idea, in the long run. It may make office easier to use, but you're terrifying about a third of your audience. It's just not a good plan.