Concept: The To-complete Schedule Performance Index (TSPI) is a prescription for future schedule efficiency. It prescribes the level of efficiency required to complete the project within a target duration. The target duration may be either the Planned Duration (PD) or the Estimated Duration (ED).
Practice: Previous posts describe two ways to assess TSPI:
- SPIt - TSPI Threshold: The U.S. National Defense Industry Association (NDIA) looks at the difference between the current Schedule Performance Index for time (SPIt) and the TSPI for Estimated Duration. Each difference is compared to threshold values that indicate the confidence level warranted by the duration estimate.
- 1.1 Threshold: Walt Lipke compares the current TSPI to the threshold value of 1.1. If the TSPI exceeds the 1.1 threshold, there is reason to believe the project will exceed the Planned Duration (PD) or the Estimated Duration (ED), whichever is being assessed. Otherwise, the project is on track to achieve the desired outcome.
The two assessments might seem to offer similar information, but a comparison reveals important differences between the two. (Note: IEACt = Independent Estimate at Completion for time.)
Table 1
Comments
Item 1: Lipke’s approach works when the Estimated Duration equals the IEACt, whereas the NDIA’s approach does not. Given the importance of the IEACt as a duration estimate (see the previous post), this is a significant gap in the capability of the two assessments.
Item 2: The NDIA approach assesses whether or not the forecast is overly pessimistic or overly optimistic. Lipke’s approach, on the other hand, does not ascribe pessimism or optimism to the estimated duration (or, for that matter, the planned duration). Instead, Lipke focuses on whether or not the target date can be achieved. The NDIA’s assessment, therefore, adds another dimension to what TSPI can tell us.
Item 3: Although the NDIA applies thresholds only to TSPI for Estimated Duration, ProjectFlightDeck has found that the SPIt - TSPI thresholds also fit TSPI for Planned Duration. Given that the NDIA approach can be applied to both, its scope is comparable to Lipke’s.
Item 4: While both approaches present a binary evaluation, ProjectFlightDeck adds an intermediate status to Lipke’s analysis: TSPI < 1.1 and TSPI >= 1.0 means that the predicted outcome is achievable, but caution is warranted because the TSPI is approaching the 1.1 threshold. That leaves TSPI < 1.0 as the only unconstrained indication of achievability. By further delineating Lipke’s threshold, we extract more information from the TSPI.
Item 5: Both approaches reach a limit in the final period because the associated equations break down. For both approaches, the denominator (i.e., ED – AD for NDIA, EACt – AT for Lipke) goes to zero in the final period. That is not surprising because the TSPI measures the performance level required to meet the target date. It makes sense that, when you arrive at the target date, there is nothing left to measure.
Conclusion
At ProjectFlightDeck, we most commonly use Lipke’s approach. It works for the IEACt, which is a demonstrably superior duration estimate. Also, by adding an intermediate status, we align the TSPI with other Red-Yellow-Green metrics that we use on projects.
Still, we acknowledge that there are situations in which the NDIA thresholds are useful, especially when it is important to highlight the optimism or pessimism of the estimate (or plan). |
