ProjectFlightDeck

Concept: ES Statistical Analysis factors historical variation into the calculation of duration estimates. By analyzing the project’s previous schedule performance, ES stats identify the high and low bounds for the estimate at completion for time. From high/low bounds and target dates, we can determine how well or poorly the project will perform. Thresholds guide the assignment of Green-Yellow-Red status.

Practice: At ProjectFlightDeck, we use ES Statistical Analysis to steer the project. In particular, we compare the high and low duration estimates to allowances for uncertainty. The comparisons provide insight into the status of the project.

As with other ES metrics, ES Statistical Analysis has advantages and disadvantages.

Pro:

• Probabilistic. Unlike the IEACt and IECDt, Statistical Analysis takes account of the historical variation in schedule performance. The estimates are probability functions, rather than point values.
• Empirical Estimates. Given the unique characteristics of each project, the best basis for estimation is the project itself. As noted in the Agile framework, the team learns from its experience on each individual project. ES Statistical Analysis formalizes the approach. It uses results actually obtained over time to predict the likely range of future outcomes for the project.
• Guidelines for Application. You do not need to understand the details of how the statistics are generated in order to use them to steer your project. You can apply stats to assess allowances for uncertainty (as suggested by ProjectFlightDeck), or use them to identify trends (as suggested by Walt Lipke). Neither application requires a detailed understanding of the math.
• Available Tools. There are easy-to-use, affordable tools available for producing ES statistics. For a free trial copy of our Excel tool, add a comment to this post or email Robert.VanDeVelde@ProjectFlightDeck.com.

Con:

• Calculation Complexity. ES statistical calculations are more complex than other ES calculations. Understanding the math can be daunting, although not really necessary for using the stats.  
• Innovation Challenge. If you want to develop new applications of ES statistics, you will benefit from understanding the math. Without such an understanding, it will be difficult, but not impossible, to create innovative uses for the stats.
• Re-Baseline Woes. A re-baseline effectively destroys the data pool used in ES Statistical Analysis, rendering the stats ineffective for steering the project. As Walt Lipke puts it, “The impact of a re-plan could (and usually will) overcome the forecast implications” (Lipke (2009), p 156).

Concept: ES Statistical Analysis factors historical variation into the calculation of duration estimates. By analyzing the project’s previous schedule performance, ES stats identify the high and low bounds for the estimate at completion for time. From high/low bounds and target dates, we can determine how well or poorly the project will perform. Thresholds guide the assignment of Green-Yellow-Red status.

Chart 1

Practice: At ProjectFlightDeck, we use ES Statistical Analysis to steer the project by comparing the high and low duration estimates to allowances for uncertainty. The comparisons provide insight into the status of the project.

There are other ways to interpret the results of ES Statistical Analysis. For instance, Walt Lipke proposes the following  interpretation based on the graphical display of ES statistics (Lipke (2009), pp 156-7).

• When the high and low bounds are symmetrical around the nominal value, the nominal estimate will be fairly close to the final outcome.
• When there is an upward trend in the graph, the duration estimates are climbing, and that indicates performance is worsening over time. Experience with such cases has shown that the high bound ends up as slightly higher than the final duration.
• When there is a downward trend in the graph, the duration estimates are falling, and that indicates performance is improving over time. Experience with such cases has shown that the low bound ends up as slightly lower than the final duration.
  

Other Observations

Walt recommends using a 90% confidence level to generate ES statistics. While higher levels, for instance, 95% or even 98%, yield extremely reliable results, they tend to overestimate the bounds. Exaggerated estimates are less valuable for steering the project.

Walt acknowledges that ES statistical methods may seem “overwhelming”. Certainly, the math is daunting. But, it is not necessary to understand the intricacies of the calculations in order to use them. All that’s required is a grasp of how to interpret the results. The posts in this blog are intended to give you the information you need to steer your project with ES statistics.

Finally, Walt notes that there has been a lack of tools for producing ES statistics. ProjectFlightDeck has closed the gap. If you want a free, trial copy of our Excel ES Statistical Analysis tool, add a comment to this post or email Robert.VanDeVelde@ProjectFlightDeck.com.