Don’t Gamble With Uncertainty

A Monte Carlo analysis helps project teams go beyond risk management. By Antonio Sturiale, PMP; Lidia Chicca; and Sergio Gerosa, PMP

One of the first things project managers learn is that we must understand and manage risks if we want to drive our projects to success. But focusing on risks alone often isn’t enough—uncertainties also must be gauged.

At Thales Alenia Space, Europe’s largest satellite manufacturer, our project professionals use the Monte Carlo analysis to measure uncertainties.  The analysis provides the answer to the fundamental question: What is the probability that our project will be completed on time and within budget?  This statistical approach helps increase our ability to control the project’s outcome and deliver a successful project.

While a risk is an event that would create a significant impact to budget, schedule, quality or scope, uncertainty can disrupt the entire project baseline. During project execution, for instance, uncertainties can create several deviations from the critical path. Each of them might have limited impact on the activities themselves. But when combined, they can alter a project significantly.

By following the five steps of the Monte Carlo analysis, you can estimate these variances and obtain a likely time length and cost for your project.

1. Define the model. The process starts by creating a separate project schedule for risk analysis. This detailed schedule must include links among all the tasks but avoid the use of contingencies. For simple projects, all the activities may be in a single chain with a simple “finish-to-start” dependence. This is the case when the overall duration of the project is just the sum of all the possible durations of the single activities. For complex projects, you will have an activities network, rather than a sequence, with different dependencies.

2. Identify the variables. The variables will be different random events—such as a rise in currency exchange rates or inexperienced software engineers who are writing or testing code—that could influence the activities.

3. Generate probability distributions. For each task in the plan, the expert in charge of the task itself should identify likely ranges of cost and duration. For instance, a less experienced engineer may write the code 3 percent slower than the average time.

4. Launch simulations. This is the step where results start to take shape. Each event is used as an input in the mathematical model representing the project. The results will be the sum of all the possible occurrences in terms of overall duration or cost of the project. Because the method is very robust, you can benefit from it by running simulations of just 30-50 events.

5. Verify results and produce final reports. Simulation results can provide the likely duration, cost and even performance of the project. The results give us an indication of the level of confidence associated with the planned project finish date and cost or indicate which plan activities to more strictly monitor.

However, results of the simulations must be analyzed by the project team to validate them.
 And these probabilities should be considered along with other risk management processes to provide a more complete picture of the future of the project. Not combining them could generate unwarranted confidence in the forecast and doom a project.

How might this approach to planning for uncertainties add value throughout the project life cycle? During the bidding phase, it can help determine a bid/no bid decision, minimize risk exposure and determine contingencies. During planning, it helps to show project completion dates and relevant confidence rate. And during execution, it can reveal if you are executing to plan and whether risks or uncertainty hot spots are on the horizon.

One piece of advice for those conducting Monte Carlo analyses: The approach aims to treat uncertainties with numbers, but be careful to avoid being overwhelmed by the data. Keep the focus on how the results can inform your decisions, not the statistics themselves. A Monte Carlo analysis is not the solution; it’s a robust methodology that will support you and your team in your day-to-day decisions.

Antonio Sturiale, PMP, is portfolio manager for inter-entity projects, estimates and investments.

Lidia Chicca is operational and forecast planning and production plan manager and risk analyst.

Sergio Gerosa, PMP, is operational and forecast planning director.

All work at Thales Alenia Space, Rome, Italy. 

Source: PM Network 10/2017 - TO THE CLOUD

Stevbros delivers project management training worldwide, our courses have proven their worldwide acceptance and reputation by being the choice of project management professionals in 168 countries.

Share in