normal_curve“What-if?” These are the two words most heard in the project management industry. Everyone on a project has his own “what-ifs”. Each member of a project team has his own perspective. But what everyone is referring to here is targeted at the same thought process. This is an aspect of contingency planning.

Good project managers, just like good managers of any kind, are constantly thinking of what might happen that isn’t currently planned for. For the most disruptive of conditions, they will likely have already mapped out an alternative strategy for dealing with the issue. This is fundamental to being a proactive manager. Lesser managers often find themselves managing by reaction or by emergency. This is a result of a contingency occurring for which there is no immediate solution.

Risk is often thought of in terms of insurance but it fits very well in a project management context. For those who are involved in schedule and cost management there are many tools now available on the market that can help.

There have been risk analysis project management tools available for your PC for over 10 years. Oracle’s Primavera has Pertmaster, a system which is designed to provide risk analysis and reporting for project schedules. @Risk and Risk+ are two products designed to provide similar analysis to Microsoft Project users. Deltek has WelcomRisk designed to add Monte Carlo risk analysis to its Open Plan product. This

So what does this software do? Does it tell me ho risky the project is? Does it somehow manage the risk for me?

First of all, risk analysis software does not do risk assessment. What it does is take the risk you assess to each task and provide analysis of the combined data to give you a view into areas of the project which might not otherwise grab your attention. What is this analysis? One of the most common algorithms is called Monte Carlo.

Okay, what’s Monte Carlo? Isn’t that where everyone goes to gamble? In fact, that’s exactly what it refers to. Here’s what happens in the Monte Carlo method: The user first inputs the minimum, maximum and best-guess of each task’s duration. In addition, the user inputs a type of curve. Examples of such curves might be a Bell curve or a Uniform curve.

The analysis is, in fact, just a critical path calculation. The difference is that it is done many, many times. And each time a task is considered, instead of just taking the original duration, the algorithm “rolls the dice” and uses a random duration. The randomness is controlled by the input. The algorithm uses the range between the optimistic and pessimistic durations and the curve to determine which duration to use. A Bell curve, for example, will tend more to the middle of the range where a Uniform curve will have an even chance anywhere along the range. The analysis continues through the project then returns and does it again. Often an analysis will perform 100 or even 1,000 passes through the project to deliver adequate results.

The algorithm returns information on each task as well as additional information on key tasks. So what’s the point? Well, consider a hypothetical project. In our project we have five tasks in sequence. Each task has four days of float or ‘slack’ time. Float is the amount of time an activity could be delayed without delaying the project itself. A task with no float is considered ‘critical’.

Okay, so we have our five tasks in a row. Imagine that the first task is three days late. Now the next task must have only one day of float. If it goes a day late, then the next three tasks in the sequence immediately become critical. A normal critical path methodology analysis, such as can be found in most project management systems, will never show you this information. However, a Monte Carlo simulation can find it easily. By rolling the dice, some of the time the first tasks will come up with a longer duration.

In combination with various reports, project managers can no be presented with two kinds of information. First, a listing of tasks which have a high degree of probability of becoming critical at some time in the project. Secondly, for key activities, a confidence curve can be generated showing a histogram.

There is another aspect to these reports which may be the most useful of all. This involves the trend of these analyses over time. Let’s consider this in basic terms.

On the last day of the project there is a 100 percent chance that the project will finish that day because, well, because it just finished. On the first day of a project, a risk analysis might tell that that the project is due to finish no earlier than Jan 1 and no later than July 1 of next year, a six-month range.

Each time the project is updated, the “risk” or the range should get narrower. Well what if it doesn’t? This can be a major indicator to a client or to the project manager that there is something severely amiss. It almost guarantees that there is a change in the project scope somewhere in your future.

Going through the extra effort of risk analysis isn’t often required on every project.  But, when you’re wondering what the impact is on a project with an unusually high number of risky assumptions, Risk Analysis software can go a long way to showing you some empirical results.