Major projects have to be planned, and the planning process starts by breaking the project down into a number of tasks. Resources may be assigned to each task, and an estimate made of its duration.
Crucially, dependencies between the tasks need to be identified. These are represented by links between the tasks, creating a project network which can be analyzed to estimate start and finish dates for each task. This process is the simplest form of project scheduling, and is known as critical path analysis. (The method is often called the critical path method, or CPM.)
Project schedules produced by CPM often prove to be too optimistic. There are many reasons for this, but probably the main two are:
- As many studies have shown, people tend to be overly optimistic in their estimates of task durations.
- Even if the estimates of individual task durations are not overly optimistic, they are subject to uncertainty and sometimes these uncertain durations combine in a way which causes an additional bias called merge bias.
Project scheduling can be improved by recognizing these problems, and the best way to do this is by replacing CPM with risk analysis. This means replacing each single-point estimate of duration with a probability distribution. This in turn requires the user to enter at least two values for each task, an optimistic and a pessimistic estimate of the duration.
The thought process this entails – for example, thinking of what might go wrong — goes a long way towards correcting the first problem mentioned above, namely the natural tendency for people to underestimate the time it will take to accomplish a task.
The subsequent analysis of this data – using a risk analysis product like Full Monte – corrects the second problem. Merge bias results when multiple tasks converge in project network, that is to say when one task cannot start until several other parallel tasks are completed. Since this task has to wait for the latest of the predecessors, and since this may not always be the same one, the expected start of the successor is later than the latest expected finish of the predecessors. This is the cause of merge bias. It is properly modeled by risk analysis but not by CPM (nor by a modified version of CPM called PERT).
Risk analysis therefore enhances the project scheduling process by addressing the two main causes of overly optimistic schedules. The schedules it creates therefore have a better chance of being executed successfully. For more information, call us today at 281-971-9825.
