The problem, in many cases, is that Planned Value (PV) cannot realistically be computed at all. In a project that is expected to take about 18 months to complete, with an average of 25 engineers on the project throughout, and the number of engineers changing over time, we can only begin to compute the planned value as time progresses.
A simplistic model, that I realized was possible during a discussion with my fellow colleague Richard Wheeler, is what is presented here, as an example. No need to worry about actual engineers involved, etc., by making the following, very realistic, assumptions [A click on the graph will open it in its own tab/window]:
- Planned Value is 0 at the beginning of the project. (7/1/2010)
- Planned Value is 2 at the planned end of the project. (12/31/2011). This is also the Budget at Completion (BAC). (The choice of 2 for PV is only to separate the graphs of EV and SPI in the same presentation).
- An S-curve is fitted between the beginning and the end of the project. I have used a very simple Excel model of an S-curve in this example; your environment might be better served by a different model suited for your project's environment.
- For each of the interim milestones, 4 in the example, there is an associated PV, based on elapsed time. (Milestone 5, M5, is the completion of the project).
- When a certain milestone is reached in actual execution, the EV(t) is considered to be the PV at the milestone.
- When M3 is reached in actual execution, any schedule variance at that time, SV(t), EV(t) - PV(t), is assumed to persist until M4 is reached.
- Beyond M4, I have assumed a linear improvement in EV(t) until M5 is reached.
Of course, one limitation of this simplistic model is that actual costs are not considered; even then, it can be a useful model in some project situations.