Imagine you work in an industry that employs large numbers of expensive people to apply their expertise to a problem, and all you get is the result. Even though their expertise can be partially codified, that information remains in their (expensive) heads, and it's never written down, much less automated. Imagine how expensive that would be!

An "approximate calculation" is the dictionary definition of an estimate. Nearly everyone expresses interest in getting better at estimating, but what do they really mean, and is what they mean reasonable? I want to discuss estimation by focusing on several key words: target, estimate, approximation, expectation, and uncertainty.

During a recent class of executive MBA students in which we looked at a bunch of successful and unsuccessful technology implementation projects, someone asked why so many large-scale enterprise software projects fail. Since the class was full of student-executives from over 30 companies, we were able to explore the question with the benefit of multiple, real-world perspectives.

In nature and in human affairs, there are two kinds of animals: the hunters and the hunted. In lower animals, the distinction is species-dependent, but among humans, it is a function of attitude. As a project manager, which are you?