1. Start building a product
2. Screw up
3. Start Over
4. Release an uninteresting product 3 years late
5. Fire your marketing manager
I'll admit that one of the big problems with Vista is that there are no interesting or compelling reasons to upgrade, what is unclear is whether that is a product issue, or a marketing issue. Vista has been in enthusiasts hands for a year now though, so I must imagine that the issue is primarily the product itself since the spelunkers haven't found anything very compelling yet.