ProjectFlightDeck

Concept: The To-complete Schedule Performance Index (TSPI) is a prescription for future schedule efficiency. It prescribes the level of schedule performance efficiency required to complete the project within a specified duration.

Practice: Metrics such as the IEACt and IECDt are predictive: given past performance, they forecast future duration. By contrast, the TSPI is prescriptive: given past performance, it dictates the future performance level required to achieve the target duration.

The TSPI calculation is more complicated than the ones previously described in this blog. The TSPI is the ratio between two differences: first, the difference between the Planned Duration and the Earned Schedule; second, the difference between the Target Duration and the Actual Time. The Target Duration can be either the Planned Duration or the Independent Estimate at Completion for time (IEACt).

The U.S. National Defense Industrial Association (NDIA) has a chart that illustrates the factors in TSPI calculations. To achieve the IEACt, schedule performance must equal the ratio between the remaining Planned Duration (PDWR in Diagram 1) and the Remaining Estimated Duration (RD in Diagram 1). In other words, it is the ratio between the two sections of the chart highlighted in yellow. Plugging the NDIA terms into Walt Lipke’s TSPI equation, we have: (PD - ES) / (ED - AD), or: PDWR / RD.

Diagram 1

The NDIA emphasizes use of the IEACt as the Target Duration in TSPI calculations. A later post will explore the NDIA approach in detail. For now, let’s look at the use of Planned Duration as the Target Duration in TSPI calculations.

At ProjectFlightDeck, TSPI calculations use Planned Duration, rather than the IEACt, as the key target duration. To achieve the PD, schedule performance must equal the ratio between Planned Duration less the Earned Schedule (termed the Un-earned Planned Duration) and the Planned Duration less the Actual Duration (termed the Remaining Planned Duration).

Modifying the NDIA diagram to illustrate this version of TSPI, we have the ratio between UPD in Diagram 2 and the Remaining Planned Duration (RPD in Diagram 2. Plugging these terms into Walt Lipke’s standard equation, we have: (PD - ES) / (PD - AD), or: UPD / RPD.

Diagram 2

Later posts will explain how we use the TSPI in our projects.

Concept: The Independent Estimate at Completion for time (IEACt) is the forecasted duration of a project. It is the ratio between Planned Duration (PD) and the Schedule Performance Index for time (SPIt). A related metric is the estimated end date, i.e., the Independent Estimated Completion Date for time (IECDt). It is the project Start Date plus the number of periods in the IEACt.

Diagram 1

The following table shows how the thresholds are structured.

Table 1

Practice: The IEACt and IECDt have many advantages, but they also have some limitations.

Pro:

• Quick and economical to produce. The IEACt equals the Planned Duration divided by the SPIt. The IECDt is simply the IEACt added to the Baseline Start Date.  Given that you are already calculating the Earned Schedule, you have all the factors you need for IEACt and IECDt.
• Improved Accuracy. IEACt / IECDt are derived from objective, time-based data, rather than inferred from subjective, cost-based observations. Studies have repeatedly demonstrated that IEACt / IECDt provide duration forecasts superior to other EVM metrics. (See References below.)
• Widely applicable. Regardless of project size or the delivery approach being used (e.g., Agile or Plan Driven), the thresholds provide the same sort of information. This is especially useful for portfolios of mixed project sizes and delivery methodologies, as it provides a common basis for evaluating schedule performance.
• High-profile metric. Most projects are time-sensitive: the end date matters. Estimates of when the project will finish attract attention well beyond the PM and project team. Management tends to view it as an essential component of project status reporting.

Con:

• Nominal, not probabilistic. IEACt and IECDt assume that the pace of schedule execution in the remainder of the project will be the same as in the part already completed. Generally, projects do not have a uniform rate—it varies over time. The IEACt and IECDt do not reflect that variation. By contrast, ES statistical analysis considers historical deviations and produces probabilistic estimates of IEACt and IECDt. That metric will be addressed in a later post.
• Early Contingency and Reserve. Projects are more commonly late than early. So, risk analysis tends to focus on uncertainties that are likely to cause delays.  The idea of early Contingency and Reserve may therefore be unfamiliar. Still, it’s clear that uncertainties can lead to early delivery and that project plans should account for such cases.
    

References

Cumrine, K.T., Earned Schedule: Utility in Major U.S. Air Force Acquisition Programs, EVM World, 29-31 May 2013.

Crumrine, Capt Kevin (USAF) & Lt. Col. J. D. Ritschel, PhD. “A Comparison of Earned Value Management and Earned Schedule as Schedule Predictors on DoD ACAT I Programs, The Measurable News, 2013 Issue 2.

Henderson, K. “Earned Schedule: A breakthrough extension to Earned Value theory? A retrospective analysis of real project data,” The Measurable News, Summer 2003: 13-23.

Rujirayanyong, Thammasak. “A Comparison of Three Completion Date Predicting Methods for Construction Projects,” Journal of Research in Engineering and Technology, Volume 6 2009 (Kasetsart University, Thailand).

Tzaveas, Theodoros, S. Katsavounis, G. Kalfakakou. “Analysis of Project Performance of a Real Case Study and Assessment of Earned Value and Earned Schedule Techniques for the Prediction of Project Completion Date,” Proceedings of IPMA Conference, May 2010 (Crete, Greece).

Tzaveas, Theodoros. “Earned Value Method (EVM) and Earned Schedule Methodologies Comparison, Greece,” Institution of Civil Engineers, November 2010.

Vanhoucke, M., Vandevoorde, S. “A Simulation and Evaluation of Earned Value Metrics to Forecast the Project Duration,” Journal of the Operational Research Society, vol 58: 1361-1374 (October 2007).