Overrun likelihood (Y), worst case overrun duration (D), and the creep or step-overrun characteristic function cannot be casually approached! As with other information gathered throughout IPM project planning, task leaders again must: 1) review past projects, 2) talk to other experienced people, 3) review the baseline and actual Gantt charts from other similar projects, and 4) confer with all task team members in their quest for contribution data (this may take several days, given their other responsibilities). The project manager must be available during this time. A manager's advice and help, especially to people new to the procedure, is essential. Each task leader will report this gathered information to the project leader as soon as possible. The project manager then will accumulate this data as shown in Table 8-1. Only critical path tasks will be used in the analysis. However, some tasks that have very little slack and high Ds can impact the critical path (This explanation of dealing with non-critical path task overruns is left for Chapter 9). The left vertical column of the table follows the sequence of the critical path. (This analysis is not completed until the various causes of contribution are analyzed, which is discussed in Chapter 9.) #### Worksheet The characteristics of Table 8-1 are discussed in the following paragraphs. The # column identifies the task by number in accordance with its place in the Gantt chart task column. Breaks in the numerical sequence reflect the Gantt chart tasks that are not on the critical path. The "Task" column should give the task's name (letters have been substituted for task names). The next column gives the "most of the time" (M) estimate. This is the estimate that appears as a bar on the project Gantt chart. The "Longest Time" (W) column gives the worst case estimate. The "Delta" (D) column indicates the difference between the longest time estimate and the "most of the time" estimate. (On a spreadsheet, an equation can be entered for this column that will calculate the difference of W - M.) The "% of time task overruns likelihood" column (Y) gives the likelihood in a percentage that the task will overrun its original estimate. (This is the likelihood that the task will overrun at all, not the likelihood that the overrun will be the worst case estimate.) ##### Table 8-1. # | Task | "Most of the Time" Estimate (M) (Days) | Longest Time Estimate (W) (Days) | Delta W-M D | % of Time the Task Will Overrun Likelihood Y | Contribution D Times Y | Risk Type | 1 | A | 2 | 3 | 1 | 0.10 | 0.1 | C | 3 | C | 5 | 5.5 | .5 | 0.40 | 0.2 | SF | 4 | D | 1 | 3 | 2 | 0.30 | 0.6 | C | 6 | F | 2 | 2 | 0 | 0 | 0 | | 7 | G | 5 | 6 | 1 | 0.20 | 0.2 | C | 9 | H | 4 | 5 | 1 | 0.20 | 0.2 | SF | 11 | J | 2 | 2.5 | .5 | 0.10 | 0.05 | C | 13 | M | 7 | 7.5 | .5 | 0.30 | 0.15 | SF | 14 | N | 6 | 6 | 0 | 0 | 0 | | 15 | O | 8 | 11 | 3 | 0.20 | 0.6 | C | 16 | P | 10 | 15 | 5 | 0.10 | 0.5 | C | 17 | Q | 20 | 27 | 7 | 0.15 | 1.05 | SF | 19 | S | 3 | 4 | 1 | 0.40 | 0.4 | C | 23 | W | 5 | 7 | 2 | 0.30 | 0.6 | C | 25 | Y | 7 | 10 | 3 | 0.10 | 0.3 | C | 29 | CC | 19 | 28 | 9 | 0.10 | 0.9 | C | 30 | DD | 6 | 8 | 2 | 0.20 | 0.4 | C | 33 | GG | 10 | 14 | 4 | 0.15 | 0.6 | C | 34 | HH | 4 | 5 | 1 | 0.10 | 0.1 | SF | 36 | | 6 | 9 | 3 | 0.20 | 0.6 | C | | Totals | 132 | 178.5 | 46.5 | | 7.55 | | The "Contribution" column is the duration of the overrun in the worst case situation times the likelihood that the task will overrun. This is D times Y, and the equation can be entered in a spreadsheet column. It is the value you seek in doing the risk analysis and it represents the contribution the task makes to the determination of the risk factor for this sequence of tasks. The last notation column on the risk calculation worksheet is for disclosing the "risk type" creep (C) or step function (SF). In certain cases, a task with a large step function D will require special consideration in calculating the risk factor. This worksheet is the essence of IPM risk analysis. The system has been developed through both analysis and trial and error. Likelihood is an estimate. It takes some discussion to help task leaders arrive at a good likelihood estimate. Everyone must realize that it is an estimate, a very useful estimate, but by no means is it mathematically precise. Experience has taught that task leaders can and will make good likelihood estimates if they understand why they are needed. (Likelihood is not probability. Probability is a technical term from statistics that is based on the analysis of a mass of data. This mass of data does not exist for project time buffer analysis.) The delta figure (D) also is the result of estimates. It is the difference between the longest case estimate and the most-of-the-time estimate. Much of the experimentation that leads up to IPM time buffer analysis centered on how to choose D. One experiment was using the average overrun duration for the Y period instead of the longest overrun duration. Although it seemed reasonable, it did not work as well in practice as the D based on longest overrun estimate. (It was also harder to determine). The IPM risk analysis methodology only works if task leaders take it seriously. However, the project manager first must understand it and explain it well at the risk factor discussion meeting. The project manager also must be available to help team members as they seek out the risk analysis data. (The details of determining a task's risk factor contribution are spelled out in the next chapter. Details are separated from fundamentals here because experience has demonstrated that this separation makes the total concept easier to understand.) |