Boiling the IT Frog

second-guessing everything the IT organization does, and IT cant be successful.

You establish trust by agreeing on common objectives and goals and having consistency in how you communicate with each other. Both sides do what they say, and say what they do.

Why is using return on investment for project selection a ticket to failure? Most people would agree that the right projects are those that are in the best interest of the company and make the most strategic sense. But there are two fallacies in using ROI to choose your projects. First is that the project with the highest ROI is best for the business. Thats not necessarily so. For example, a high-ROI project may make a process more efficient by making it more rigid, when what you really need is for it to be more flexible. The second fallacy is that you can compare the ROI of project proposals. Those who write proposals tend to overstate return, understate expenses, minimize transition costs and dependencies on other projects, and neglect risk. So the project with the highest ROI on paper tends to be the one with the most creative proposal writer.

Why do most projects fail? In my experience, there are six primary reasons. First, youre doing the wrong project  its not what the business really needs. Second, youre missing prerequisites in the proposal. You start and then realize you have to beef up the infrastructure, for example, so youre in trouble right from the beginning. Third, youre 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 projects duration is greater than the job tenure of the sponsoring executive. If the project isnt 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, theres not enough contingency planning. Youve got to anticipate things that can go wrong and take them into account in your plan.

Why is it that adding more resources to an IT project thats running late can make it take even longer? Usually, when we talk about adding resources, were talking about adding people, and adding more people requires more coordination. Youhave to rearrange the tasks being done and break them up in different ways. Reallocation of resources takes resources in itself. Also, you would probably have assigned the best people to begin with, and now youre adding people who may not be the best, so you have to manage subpar performers.

Read more about management in Computerworld's Management Knowledge Center.