Cost Management as Described in the Guide 3rd Edition
In the 3rd Edition, the Guide laid out a framework for understanding reserve analysis in a way that was fairly straightforward. A graphic making this clear is shown in Figure 1.
Figure 1: Reserve analysis as presented in PMBOK®, 3rd Edition
In this model, we assume that the estimator is using a bottom-up method of estimating and rolling up costs to higher levels of summarization. Estimates include what it is expected to cost to accomplish the work at the work package level (e.g., along with labor hours, quantities of materiel, etc.). Thus, the project cost (denoted "Project" in Figure 1) is based on these realistic (i.e., most likely) estimates.
Once this estimate for the project has been produced, a contingency reserve is added. In the language of the 3rd Edition, this is to cover known unknowns. Known unknowns are cost variations attributed to normal variability in materials costs, material quantities, labor hours, and labor rates. In these cases, we know that we are going to have to pay for labor and materials, but we may not know exactly how many labor hours the work will take, what the labor rate will be, exactly what quantity of materials will be required, or what material unit costs (e.g., commodities costs) will be.
Another way of looking at known unknowns is to look at it from a risk management point of view. Some risks are sufficiently likely to occur (have a high enough probability of occurring) so as to warrant treating them as certain events. They are then accounted for in the work breakdown structure and schedule, to include their cost and time impacts. If treated in this way, they are considered a certainty and are automatically included in the performance measurement baseline.
If these risks are treated as a possibility (i.e., not imbedded in the work breakdown structure and schedule) and the associated risk events occur, the portion of the budget set aside for contingency reserves can be tapped to cover those expenses and/or time extensions. According to the 3rd Edition, contingency reserves are added to the project cost estimate to arrive at a cost baseline. This cost baseline is then used as the basis for establishing the planned value baseline and so becomes the standard against which performance measures (earned value management measures) are referenced.
Contingency reserves can also apply to time and schedule. Impacts to the schedule can occur as a result of unplanned work stoppages (natural disasters, failure to pass critical inspections, rework, etc.) As well as monetary reserves, a contingency reserve of time should also be incorporated into the project plan.
According to the 3rd Edition, contingency reserves are used at the discretion of the project manager.
Managing Contingency Reserves
Two methods are presented in the 3rd Edition, for managing contingency reserves. The first is the use of a zero duration activity "... that is placed across the network path for that group of schedule activities, and is used to hold the cost contingency reserve." When I contacted PMI headquarters for clarification, what I got was a "deer in the headlights" response-meaning they couldn't explain it.
Having pondered it further, I believe the authors were talking about using what is essentially a milestone-type object in the project schedule that is separate from the tasks to be done, but that has a monetary sum attached to it. Thus, the tasks identified in the schedule contain the realistic estimates, and this new, separate element contains the cost contingency reserve. Done in this way, the cost contingency reserve rolls up into the project's performance baseline over time, with the cost contingency reserve becoming available (from a scheduling perspective) at the time that work is scheduled to be accomplished, and for the tasks for which is was inserted.
While this method enables one to incorporate the cost contingency reserve into the project schedule and performance baseline, it does not accommodate a time contingency reserve to absorb the impact of tasks that take longer than expected to complete.
The second method presented by the 3rd Edition is to add a "buffer activity" that "is intentionally placed at the end of the network path for that group of schedule activities." This method allows the buffer activity to have both a duration and the cost contingency reserve attached to it. This method is also associated with the critical chain method.