Risk and Probability
A discussion of risk would not be completed without a discussion of probability and severity. When we are looking at risky event, we are in essence establishing or assessing the probability of some undesired event coming to fruition. However, even our desired events, for example the completion of a task on the critical path at a certain time, have an associated probability. Contingent upon our estimating methods that risk probability can be quite high. If we have used some historical data specific to that task then perhaps we already know the possible range of distribution for that specific task. If our duration estimates fall within that range, we can perform some evaluation on the probability of successfully achieving that date. What about when we generate a date from expert opinion? What about when we need the task to take a certain length due to the end delivery date of the project? In those cases, the later especially so, we may just injected some additional measure of risk into the project.
Those three methods of estimating, perhaps, do not have the same risk probability, the specific risk being the late delivery of the task to the schedule. Historical data is measured and provides us with some risk mitigation. We know it can happen the way it has happened in the past. The expert opinion is probably not as sure as a measured set over time. We rely upon the expert’s knowledge and past experience, we trust they are not jaded or wear overly rose colored glasses. The last version, we make the duration up due to our stakeholder’s desired project introduction date, the level of risk is much higher.
Often not considered are the compounding implications of risks. Consider the flip of a coin; you can pick heads or tails. The probability of picking the correct outcome from the flip would be one-in-two or 50%. Let us extend that further by considering we must predict two flips of the coin. The possibility of correctly selecting the orientation of the flips is now down to one-in-four or 25%. This translates to project management as well. If we have two depending objectives or activities we can assess the probability of each coming to realization. Knowing or theorizing that probability, we are then able to calculate the probability of the two cascading events producing something we deem acceptable. For example, consider two depending tasks, upon completion of task A, we will start and complete task B. We have considered the successful completion of these tasks to be 90% probable. The probability of these two tasks ending in the desired state is 81% or the product of the two independent probabilities. Just these two actions brought our near A grade down to a C.
Risk applies to all of the tasks we take on in a project, to include our schedule. Sufficient number of those line items in your schedule with less than 100% predictability means the end schedule can be seriously compromised. The truth is, predicting the schedule as many executives and project managers believe possible – is not really possible. We are not talking about routine activities as “operations” can become, but project activities. There will always be risks and the risks and these will always compound. Things are considerably less certain than we believe or plan. This is just one of the topics covered in our up coming PMI event on risk management.