Statistical Forecasts
Statistical forecasts use past performance to calculate a performance factor (PF) that is applied to the remaining budget.
The standard formula is:
Forecast = PF (BAC - EV)
Performance factors correspond to numbers that are derived from specific calculations.
Performance Factor | Meaning |
---|---|
1 | The project is expected to complete the remaining work exactly according to the budget. |
<1 | The project is expected to complete the remaining work under budget. |
>1 | The project is expected to go over budget to complete the remaining work. |
Cost Performance Index
One method for calculating a performance factor is an index called Cost Performance Index (CPI). Past studies show that project performance rarely improves once the project is 10% complete. To apply past performance to the remaining effort, multiply the remaining budget by a performance factor calculated using the CPI.
Forecast = 1/CPI (BAC - EV)
where:
CPI = Earned Value/Actual Cost
For example, assume that a CPI of 0.85 has been calculated indicating that for every dollar being spent, only 0.85 dollars worth of work is getting done. In this case, dividing the project target cost by the CPI of 0.85 provides a realistic cost estimate that the remaining work performed on the project will continue with the same cost overruns.
Forecast = 1/.85 (BAC - EV)
Performance Factor Calculations
Cobra supports the following methods for calculating a performance factor:
- Performance factor = 1 — This method assumes that the project will perform all remaining work according to budget.
- Performance factor = 1/CPI cumulative to date (where CPI = Earned Value/Actual Costs) — This method assumes that all remaining work will be performed at the same rate of efficiency (cost performance index or CPI) as has been achieved so far.
- Performance factor = 1/CPI last status period — This method assumes that all remaining work will be performed at the same rate of efficiency as has been achieved in the current fiscal period.
- Performance factor = 1/CPI last three status periods — This method assumes that all remaining work will be performed at the same rate of efficiency that has been achieved in the current period plus the two previous periods.
- Performance factor = 1/CPI last six status periods — This method assumes that all remaining work will be performed at the same rate of efficiency that has been achieved in the current fiscal period plus five previous fiscal periods.
- Performance factor = user-defined value — This method allows you to enter a performance factor at the time new forecasts are generated.
- Performance factor = 1/((a * CPI) + (b * SPI)) (where a + b = 1.0) — This method allows you to define a performance factor that reflects the cumulative CPI and SPI and in which the relative weighting of CPI and SPI are user-definable. This method allows you to indicate the relative importance of cost and schedule performance when calculating performance factors.
For example, assume that the cumulative CPI for a work package is 1.5 and the cumulative SPI is 0.6. You want to assign a relative weighting of 75% to the cost performance and 25% to the schedule performance. As a result, Cobra calculates a work-package performance factor of 0.889 as follows:
Performance factor = 1/[(0.75 * 1.5 ) + (0.25 * 0.6)] = 1/[(1.125) + (0.15)] = 1/1.275 = 0.889 By contrast, if you assume the same values for CPI and SPI, but assign a weighting of 25% to cost performance and 75% to schedule performance, Cobra calculates a work-package performance factor of 1.212 as follows:
Performance factor = 1/[(0.25 * 1.5) + (0.75 * 0.6)] = 1/[(0.375) + (0.45)] = 1/0.825 = 1.212 - Performance factor = 1/(CPI * SPI) (where SPI = EV/Budget) — This method allows you to define a performance factor based on both the cumulative cost performance index and the cumulative schedule performance index.
For example, assume that a work package originally budgeted at $10,000 is halfway through its schedule and has a cumulative budget of $5,000. 30% of the work package budget has been earned, resulting in a cumulative EV of $3,000. Cumulative actual costs, however, are $2,000. Thus, the work package has an unfavorable SPI of 0.6 (3000/5000) and a favorable CPI of 1.5 (3000/2000).
As a result, Cobra arrives at a work-package performance factor of 1.111 as follows:
Performance factor = 1/(0.6 * 1.5) = 1/0.9 = 1.111 - Multiple performance factors — This method allows you to have multiple performance factors. Cobra calculates a different performance factor depending on how much of the project has been completed. Cobra determines how much of the project has been completed by comparing the cumulative EV to the BAC.
You can define up to four ranges of completion. For example, if you want to set up a forecast that uses a performance factor of 1 (forecast method 1) for the first third of the work, a user-defined performance factor (forecast method 6) over the second third of the work, and the cumulative CPI (forecast method 2) over the final third, you can use the following definition:
Percent Complete Range Forecast Method 0 to 33 1 34 to 66 6 67 to 100 2