Project Estimating Using PERT

When determining task estimates for a project, fixed values are generally the norm applied to the project plan. A project manager will ask a programmer how long it will take to complete task X and the response will be H hours and this will be added to the plan. A project plan is really an estimate wrought with uncertainty. All of the great achievements in physics, math, biology and other physical sciences rely on a range values instead of a fixed value. It is therefore prudent to consider ranges of values when estimating tasks on a project.

This concept is not new, and one of the more accepted techniques for estimating using ranges is PERT - Program (or Project) Evaluation and Review Technique. PERT takes into account pessimistic, most likely and optimistic values for a given task and provides a best estimate based on the following formula using a "weighted" average and standard deviations (which are beyond the scope of this article):

Estimate=(Optimistic+4Most Likely+Pessimistic) / 6

For example, task T is to build a customer data entry screen that commits data to the database. Using the PERT approach, a project manager asks programmer Bill to provide the three values and provides the following:

• Optimistic estimate: 5 days
• Most Likely Estimate: 8 days
• Pessimistic Estimate: 13 days

Applying the PERT formula reveals the following Estimate:

Estimate=(5 days+4(8 days)+13 days)/6

Equals

Estimate=((50))/6=8.33 days

Therefore, according to Bill it is "highly likely' that this task will take a little more than 8 days to complete.

The project manager decides to ask for a second opinion from programmer Jane, who provides the following:

• Optimistic estimate: 3 days
• Most Likely Estimate: 6 days
• Pessimistic Estimate: 10 days

Applying the PERT formula reveals the following Estimate:

Estimate=(3 days+4(6 days)+10 days)/6

Equals

Estimate=((37))/6=6.2 days

Therefore, according to Jane it is "highly likely' that this task will take a little more than 6 days to complete.

The project manager decides to ask for a third opinion from programmer Tim, who provides the following:

• Optimistic estimate: 4 days
• Most Likely Estimate: 7 days
• Pessimistic Estimate: 12 days

Applying the PERT formula reveals the following Estimate:

Estimate=(4 days+4(7 days)+12 days)/6

Equals

Estimate=((44))/6=6.2 days

Therefore, according to Tim it is "highly likely' that this task will take a little more than 7 days to complete.

If the project manager took into account just the most likely estimates, there would not be too much variance, but the difference between a little more than six days and a little more than 8 days is significant enough. And in this example the project manager is dealing with ranges and risks that exist on both the optimistic and pessimistic values, which could certainly impact the project for better or for worse.

Asking just three resources for their estimates provides some comfort, but it would be even better if the project manager were able to ask 10,000 "resources" what they thought and to apply probability theory to the estimates. This is what Monte Carol simulations can provide and this concept will be covered in the next article.

Ben DiMolfetta, PMP, is the Director of Grants Management Systems for WinMill Software.