Project risk analysis

Download a pdf copy of this help file  here

See also: Cost risk analysis, Schedule risk analysis, Cascading risks

Introduction

Project risk analysis is concerned with the assessment of the risks and uncertainties that threaten a project. Project risk analysis has a broad range of applications, just as the definition of a project is broad.

For the purposes of this guide, we consider a project to be the following:

a project is any set of inter-related tasks designed to achieve a certain goal or set of goals, with a limited set of resources, to a particular quality standard and within a particular budget and time frame.

There are two main elements that confound our ability to determine whether a proposed project, or a project already underway, will achieve its goals within the set restrictions:

• The general uncertainty surrounding the duration and cost of tasks within the planned project.

• The set of risk events that may occur, which inhibit the smooth progress of the project.

Typically, a project risk analysis consists of analyzing schedule and cost risk, though other aspects like the quality of the final product are sometimes included. There will also often be an analysis of the cash flow of the project, especially at the conception and bidding stages.

A cost risk analysis consists of looking at the various costs associated with a project, their uncertainties and any risks or opportunities that may effect these costs. Risks and opportunities are defined as discrete possible events that will increase and decrease the project costs respectively. They are both characterised by estimates of their probability of occurrence and the magnitude of their impact. The distributions of cost are then added up in a risk analysis to determine the uncertainty in the total cost of the project.

A schedule risk analysis looks at the time required to complete the various tasks associated with a project, and the interrelationship between these tasks. Risks and opportunities are identified for each task and an analysis is performed to determine the total duration of the project and, usually, the durations until specific milestones within the project are achieved. A schedule risk analysis is generally more complex to perform than a cost risk analysis because the logical connections between the tasks have to be modelled in order to determine the critical path.

Cost and duration are linked together

A project's cost and duration are, in reality, linked together. Tasks in a project are often quantified by, among other things, the number of person weeks (amount of work) needed to complete them. The duration of the task is then equal to the person weeks/people on the job and the cost equals person weeks * labour rate. Costs and durations are also linked if the model includes a penalty clause for exceeding a deadline.

Cost elements and, particularly, schedule durations are also often correlated. Correlation, or dependency, modelling is described in detail here. It is important to be aware that dependencies often exist in a risk analysis model and failure to include them in an analysis will generally underestimate the risk.

A project risk analysis is often completed after a more rudimentary (deterministic) analysis that uses single-point estimates for each task duration and cost. A comparison of the results of this deterministic analysis with those of the risk analysis, where distributions have been used to model uncertainty components, often surprises people. Somehow, one expects that a deterministic analysis based on values one thinks most likely to occur should produce results that equate to the mode of the risk analysis output distribution.

In fact, it turns out that a risk analysis model will provide a mode and mean that is nearly always greater than the deterministic model result. Sometimes the risk analysis output distribution will not even include the deterministic result!

The main reason for this is that the distributions one assigns to uncertainty components are nearly always right skewed, i.e. they have a longer tail to the right than to the left. This is because there are many more things that can go wrong than go right, and because we are always trying to place emphasis on doing the job as quickly and cheaply as possible. Thus the model distributions nearly always have more probability to the right of the mode than to the left which means that, in the aggregate, for most models one is much more likely to have a scenario that exceeds the deterministic scenario.

A schedule risk analysis will diverge even more from its deterministic equivalent than a cost model because any task whose commencement depends on the finish of two or more other tasks begins at the maximum of the samples from the distributions of finish dates of the other tasks, not the maximum of their modes.

Read on: Cost risk analysis