Integrated Cost-Schedule Risk Analysis [Hardcover]

About 18 months ago, I reviewed David Hulett's book 'Practical Schedule Risk Analysis'. The book dealt with statistical analysis of project schedules, and why common deterministic or PERT approaches generally give optimistic predictions. At the time, I thought there was space for a companion volume to deal with the analysis of cost estimates.

In 'Integrated Cost-Schedule Risk Analysis', Hulett has provided much more than a mere repetition of the ideas from his previous book, re-directed at cost; he has provided an approach that considers the interactions and coupling between cost uncertainty and schedule uncertainty in a complex project environment.

This book does not assume that the reader is familiar with the previous work - much of the material in the early chapters on the theory and use of statistics is repeated, covering the manipulation of independent distributions, management of correlated distributions, and the superiority of the Monte Carlo method in combining uncertainty in the time and cost estimates. A few areas where the previous book dives deep into the detail are omitted from this book, and the reader is pointed to references in the previous work. These are generally for completeness, rather than being essential to an understanding of the concepts.

For those familiar with the previous work, this book gets into its stride in Chapter 7: Preparing for Integrated Cost and Schedule Risk Analysis. The importance of robust scheduling is discussed, and pointers for the areas of uncertainty to be considered are provided. The usual criteria for scheduling and planning apply; viz. the schedule (and WBS) has to include all work to be completed on the project, tasks to be linked such that all have predecessors and successors, and resources are to be loaded. In fact, the entry point for analysis is virtually the point at which planning and estimating are considered complete by most organisations.

As expected, uncertainty of events is addressed, but the book also covers analysis of uncertainties in areas such as labour productivity and staff charge rates, and the effects these have on both time and cost of the project. Correlation of uncertainties and their effects is discussed, showing how the analysis considers how single risks can have effects at many points across a project.

Where the previous book introduced the 'S-curve', the cumulative probability of delivery by a given date, this book extends the idea to cumulative probability of delivery to a given cost. However, Chapter 10: Advanced Results, extends this idea by introducing time-cost scatter diagrams, where each run of the Monte Carlo simulation is plotted on a graph of cost against duration at the end of the project. This produces a 'smear' of points around a line, the gradient of which indicates the approximate rate of on the project, and therefore the cost impact of any time delays. This presentation also supports the calculation of the 'Joint Confidence Level', a contour line following those values for time and cost against which one can be, say, 80% certain of delivering.

In summary, this book extends the ideas of the previous work, and provides a thorough approach to estimating and forecasting the duration and cost on a project. The examples demonstrate the sound statistical reasons for a significant fraction of projects overrunning in time and cost; and through explanation of the statistics and reference to commercially available tools, provides the means for an insight into the dynamics of project costs and schedules.

This book is a worthy successor to Practical Schedule Risk Analysis. Anyone with responsibility for project delivery needs to understand these effects, to be able to manage the expectations of their sponsors to understand that project schedules and costs are probabilistic in nature, and cannot be represented meaningfully by a single value.

ABOUT THE REVIEWER: On behalf of Arras People, this review is guest written by John Greenwood (PMP, CPhys, MInstP). John has in excess of fifteen years' experience in project management gained in the engineering and IT industries, and has been an active member of the PMI UK Chapter. He has worked as a Systems Engineer, and is familiar with the application of Monte Carlo methods from his work on analysis and performance prediction of radar systems. As a Physicist, he is of course delighted to see a grand unified theory that explains project cost and time, and wonders how quality can be included.