The notion put forth by Michael Hatfield about the use of risk management after the program in on baseline is not only misinformed it is simply bad project management. Let's start with a view of a schedule from the probabilistic point of view. The view of how real schedules behave.
Each activity in the schedule - be it a task, a work package, or a planning package - has a probability distribution associated with it. The reason for this probability is every activity value - a cost, a duration, a technical performance measure - is a random variable. No single point value can be credible.
Failure to understand this is the source of disappointment. The picture below is the starting point.
But let's use a simple examples before we start. If you commute to work like I do, then the time it takes to get from home to the office is shown on the GPS map in my car. This is a calculated value taken from the route and the values on the DVD under the passenger seat for the allowable . In the driveway, the distance shows 52 miles with a 1 hour and 10 minutes drive through downtown Denver. The suggested time from the computer is never the case though. The 1 hour 10 minute drive might be 2 hours on a heavy snow day. Like the drive, all values in project management are random variables.
The picture above shows that. Using Michael's conjecture...
But once the baselines are final, persisting in risk management strikes me as institutional worrying expressed in mind-numbing statistical jargon. To what end? Unless the response to a contingency event (in-scope, uncosted) was to significantly change from how the project team would have reacted normally, what difference does it make if it was anticipated?
misses a critical understanding - and a common one for those unfamiliar with how probability and statistics plays into our everyday lives. All numbers are random numbers no matter what side of the baselining process they live. Just because they are on baseline, doesn't mean their statistical nature changes to deterministic.
This process is call Bayesian Statistics - the past influences the future.
Here's how it works
- A set of activities on the left of the picture above have been completed. If they are completed on-time, on-budget, and on-specification you're a better project management than I.
- If on the other hand they are late, over budget, and have unplanned rework, then those completed activities have impact on the work activities to their right.
- This impact is usually easy to see. The work is pushed to the right. But there are other impacts as well. This included impacts on the underlying probabilistic nature of the FUTURE work.
These types of process are stochastic in nature. That is the probabilistic aspects of the activity network change as a function of time.
Now if you take Michael at his word, the network is essentially static. This of course is not the case on any real project. Nor can it ever be the case on any real project. All numbers are statistically driven by specific probability distributions that underlying the processes representing the work.
Conclusion
All activities on all projects are statistical in nature. Failing to understand this will lead to disappointment. In the Bayesian world if you don't know the underlying statistics your chances are 50/50 it will turn out the way you expect. No good odds on projects.
