154 Project Management
to have truckloads of uncertainty at the outset of a project. You
should publicize this uncertainty, not mask it behind a single fig-
ure—namely, that “best guess” you were asked to provide.
Using Commercially Available Estimating Software
Through the use of the sophisticated estimating software cur-
rently available, some of the attitudes and reactions depicted
above may begin to change. Several products are on the mar-
ket now that use the power of statistics to create probabilistic
estimates. The capability of these software tools is astounding
and well worth mention. They are quite powerful, yet easy to
use. Though they vary somewhat in style, most work the same
way, following the approach described below. (This is an exam-
ple for estimating costs; the procedure for estimating a schedule
is more complex, but similar in approach.)
Step 1. Estimate Minimum, Maximum, and Most Likely
Values. That sounds a little like PERT, doesn’t it? Actually, the
principle is similar, but, as you’ll see, the software tool takes the
evaluation to a much greater depth. Figure 8-6 shows a portion
of a typical input form.
Step 2. Select the Appropriate Risk Profile.The risk profile is
intended to model the most likely distribution of potential costs
for each activity. Profile A, for example, is suggesting that the
cost of the activity could very well fall anywhere within the over-
all range. (See Figure 8-7.) A triangular distribution that peaks
at the most likely cost is the most common choice, as it repre-
sents a very reasonable assumption. Entering data for this step
consists of a simple push of a button.
Step 3. Incorporate Any Known Dependencies.In this step,
you define any dependent relationships that may exist between
activities. This is not a big deal in estimating costs, but would
be for scheduling.
Step 4. Run the Simulation. At this point, the statistical software
takes your input and simulates the execution of your project as
many times as you wish—500 to 1000 is a popular choice.