I was reading an article in ComputerWorld  which really caught my attention. Here is an excerpt that I found so true across various companies I have worked. This is part of an interview between Kathleen Melymuka and the Harwell Thrasher, author of the book "Boiling the IT Frog".

Why do most projects fail? In my experience, there are six primary reasons. First, you’re doing the wrong project — it’s not what the business really needs. Second, you’re missing prerequisites in the proposal. You start and then realize you have to beef up the infrastructure, for example, so you’re in trouble right from the beginning. Third, you’re going for home runs instead of base hits; for example, a global rollout of a major system instead of chipping away at it a little at a time. Fourth, the project’s duration is greater than the job tenure of the sponsoring executive. If the project isn’t completed before he or she leaves, you may be in deep trouble with the successor. Fifth, you gather requirements instead of negotiating them. If you go around asking all the stakeholders what they want and your project goal is the sum of all those things, you will undoubtedly have contradictions — for example, a simple system that must do everything. Sixth, there’s not enough contingency planning. You’ve got to anticipate things that can go wrong and take them into account in your plan.

Boiling the IT Frog