While most OM texts cover Project Management (see Chapter 3), there is little discussion as to where the time estimates for activities in PERT and CPM come from or how accurate they are. This question arises in, of all places, an easily overlooked article in The Wall Street Journal (Oct.16-17,2010).
Here we learn that “planners underestimate costs in nearly 9 out 0f 10 projects” and “cost overruns for building projects are typical”. From my decades of teaching MIS and working as a consultant in IT, I can add that completion times in software development projects are also regularly underestimated.
Why is this the case? Research, according to the WSJ, shows that “people allow their best hopes to dominate the planning process”. This case of “irrational optimism” suggests that project managers would do well to keep data from prior projects. “Looking forward makes you more optimistic”, says a Norwegian researcher in the article. “Looking backward makes you more realistic”.
So when you and your students are discussing large projects that suffered major overruns (like Boston’s “Big Dig” and others they will bring up), you may want to remind students that doing PERT charts and running MS Project is all well and good, as long as the inputs are meaningful.
Finally, did you ever wonder about the the empirical basis of PERT’s 3 time estimates and the use of the Beta distribution in project management? If large projects do consistantly underestimate activity times, should we be giving more weight in the PERT formula to the “pessimistic” time estimate, b? Perhaps the formula in Equation (3-6) in our text should be t=(1a+3m+2b)/6 instead of t=(1a+4m+1b)/6?