Moving Beyond PERT
The PERT method of project management became popular among business leaders because it routinely identified the “critical path,” or the shortest amount of time required to trigger dependencies and to complete projects. During the era in which PERT was developed, project managers used the process to coordinate some of the largest construction and development projects up to that time. As computers accelerated communication and make many tasks easier to accomplish, projects grew more complex.
Today, many veteran project managers recognize that any PERT-based scheduling method must account for sudden fluctuations in task status. As a result, the “critical path” for a project can change at any time; and may actually change many times over the duration of an initiative. Therefore, when developing a project management schedule, team leaders must not only factor contingencies directly into tasks, they must also cultivate the flexibility to make sudden shifts when team members fall behind.
Visiting Monte Carlo
Novice project managers often hear the phrase “Monte Carlo simulation,” which might suggest the kind of activity that happens at an office holiday party instead of a planning meeting. After a group of physics researchers started using the phrase to refer to simulations that rely on random behavior, the nickname stuck. So, when reading about techniques that require Monte Carlo methodology, managers can expect to encounter predictions that change each time, based on random values and assessments.
Monte Carlo simulations become important when trying to determine whether team members can actually accomplish specific tasks within their budgeted time frames. Many software developers rely on Monte Carlo simulations to project task “velocity.” According to software developer Joel Spolsky, project managers can estimate the typical speed with which an individual team member can accomplish a task by making a series of calculated guesses based on past performance.
Dividing the original estimate by the actual time to complete a past task, managers following Spolsky’s method calculate a series of historical velocity factors that indicate how close that team member often comes to meeting deadlines. When building a project schedule, managers can use the Monte Carlo method to select a random velocity factor from a team member’s past. Dividing a task’s ideal duration by that random velocity factor offers a reasonable guess at how long it will actually take. And with enough velocity figures factored into a large project, the randomization evens out inconsistent guesses into a fairly accurate project timeline.
This approach benefits project managers by accounting for unforeseen delays within the mathematical model, rather than by forcing managers to break out contingency budgets during the planning phase. Managers under pressure to shorten time frames can feel confident about their plans, since contingency resources often appear as “low hanging fruit” to project sponsors and other stakeholders.
This post is part of the series: Steps to Build a Schedule
Learn the best practices of experienced project managers who use five steps to build effective team schedules.