“How to Avoid Earned Value Pitfalls”
The concept of leveraging earned value in measuring project performance is over 50 years old. Initially titled Cost/Schedule Control Systems Criteria (C/SCSC) or “CS squared”, the Department of Defense began to use this concept to better manage and report the expenditure of taxpayer money in government programs. This framework initially provided 35 criteria which government contractors were required to utilize in the management of programs awarded to them by the Department of Defense (DoD). The criteria introduced a third element of performed work measurement to the previous two-dimensional cost and schedule reporting. This is what project managers now refer to as “earned value” and it allowed the DoD the ability to control contract spending. It also gave the DoD a better understanding of the progress they were paying for in the development of missile systems or other military and scientific initiatives.
Today, the concept is used by hundreds of government contractors performing a wide variety of work ranging from military ship building to legacy hazardous waste clean-ups at superfund sites. The criteria have evolved over the years and remain available publicly to anyone interested. In the simplest terms, Earned Value Management (EVM) leverages three primary data sets, on a monthly basis, from a breakdown of work assigned to specific cost centers in an organization: Budgeted Cost of Work Scheduled (BCWS), Actual Cost of Work Performed (ACWP), and the Budgeted Cost of the Work Performed (BCWP). This data informs the contract management team and the government about what the forecast completion date and projected costs will be for a given project. Improper adherence to the criteria can negatively affect the data collected and, thus, strict adherence to these tenets is necessary.
A primary example of improper adherence is faulty creation of the Work Breakdown Structure (WBS.). The WBS is created from a collection of unique work deliverables to which the contractor assigns codes and, essentially, identifies a “bucket” of work on which to collect cost and performance information. When the WBS hierarchy is created, a proper review and approval process of it should be executed to ensure it correctly captures what work is being performed and what deliverables are committed. A common mistake in building the WBS is the inclusion of functional activities, which are, with some exceptions, not to be tracked in a WBS document. If insufficiently defined, WBS elements might overlap causing costs to be charged to the wrong item or items. Thus, a clear WBS with functional deliverables is crucial to not only tracking performance and costs correctly but for providing a correct and concise EVM.
Performance Indexes: CPI & SPI
The proper analysis of data in any EVM system is the prime reason for the criteria and strict adherence to it. The schedule and cost indices are important pieces of this data analysis, but they can also be relied on in other stages of a project. For instance, the Cost Performance Index (CPI) is an efficiency ratio, derived from the BCWP and ACWP and provides the project team with a clear picture of how much budgeted performance or earned value it is receiving for every dollar spent. A CPI of 1.00 means that a project is earning a dollar for every dollar being spent. Likewise, a BCWP of 987 and ACWP of 1974 (or CPI=987/1974=.50) illustrates that a project is spending two dollars for every dollar worth of progress. CPI is measured cumulatively and by period, but, as the project advances and the collection of costs and performance build up, the cumulative CPI is less sensitive to recent events and therefore is less reliable as a problem indicator. The current period CPI at a later stage becomes more important and continues to provide a crucial bellwether of issues as the project approaches completion. Understanding this change in focus is important in analysis of project data and in the reporting of variances that exceed threshold.
The Schedule Performance Index (SPI) is also an efficiency ratio. This metric measures how much work was performed versus how much was scheduled. Similar to CPI, a result of 1.00 is considered “par”. Using the same example as used for CPI, SPI=987/1974=.50 indicates that a project is operating at a 50% efficiency, only earning $0.50 to every dollar scheduled. The primary issue with using SPI is that it reflects an average of all scheduled tasks statuses for whatever WBS items you are tracking. Critical path is ignored in these calculations and therefore needs to be added to any analysis when writing variance reports to a customer.
Completion Indexes: TCPI & EAC
The “To Complete Performance Index (TCPI)” is primarily used to answer the question, “What efficiency will be required to complete the remaining work in order to meet Budget At Completion (BAC)?” It can also be leveraged to address variance reporting after exceeding a customer-defined threshold. As with CPI, the TCPI is less meaningful if the WBS being reported is already greater than 80% complete. Its usage during the early stages of the project, especially after 20% of a project is complete, can be extremely useful in gauging whether the project’s cost forecasts are accurate. A wise customer will point out to their contractor that if performing in the past at a CPI efficiency of .83, there is little chance of meeting a TCPI = 1.15 for the remainder of the contract.
The measure of “project percent complete” is often subjectively applied and may be used to represent budget spent or elapsed months of the project against total expected project duration. This fluidity of definition can lead to false indicators of project performance and efficacy. By standardizing this measure as cumulative BCWP (BCWPcum) divided by BAC, a project manager can effectively eliminate disparities in measure. Combined with the cumulative CPI, a project manager should be provided with an accurate efficiency to date, as well as the remaining time to correct course.
The Estimate at Completion (EAC) provided by control account managers is critical to projecting where the project costs will be in the future and if Latest Revised Estimates (LRE) are well founded. No single method of calculating EAC is sufficient and choosing formulas for calculation should be based on the stage of the project.
Three formulas appropriate to the early and middle stages of a project are:
1. BAC/CPIcum
2. (ACWPcum + BCWR)/(CPIcum x SPIcum)
3. (ACWPcum+BCWR)/∑CPI
a. ∑CPI = (BCWPt1 + BCWPt2+BCWPt3) / (ACWPt1 + ACWPt2 + ACWPt3)
b. T1 = First month of interest
c. T2 = Second month of interest
d. T3 = Third month of interest
Two additional formulas appropriate for the late stages of a project are:
1. (ACWPcum + BCWR)/(the sum of CPI’s for 6 months)
a. ∑CPI = (BCWPt1 + BCWP next 5 months) / (ACWPt1 + ACWP next 5 months)
b. T1 = First month of interest through T6 = Sixth month of interest similar to above 3-month summation
2. (ACWPcum + BCWR)/(the average of the last 12 months CPIs)
Choosing the correct formula for the correct stage of your project makes a difference when calculating EAC. Managers should not rely on a single method but, instead, leverage standardized and multiple methods to ensure accuracy. Remember, this is using formulas and data generated by the system as designed by the management team. If there are flaws in the system design, your data will not be reliable and therefore your EACs will be unreliable.
Methodology
Finally, the method chosen for evaluating earning performance will have an impact on the data collected since Earned Value Methodologies are the basis for calculating your BCWP. It is both the numerator in efficiency indices and used in calculating cost and schedule variances.
There are several methods for measuring BCWP. The 50/50 or 50/100 method allows 50% of a budgeted value to be “earned” when starting a work package and the remaining 50% to be ” earned” at completion. The strength of this methodology is that it removes any subjectivity from the calculation. This method is ineffective in long-duration work packages (over 3 months, reporting monthly) because it separates potential earned or remaining earned from the month where work was budgeted, thus creating a schedule variance worth 50% of that budget. Similarly, the 0/100 method also removes subjectivity but is only useful in very short duration work packages of 1 month or less.
Next, Milestone Weighting, a very popular method, involves either equally weighted milestones for deliverables within a work package or variant milestones which weights the milestones based on the effort involved. It is most useful for work packages with durations over 3 months.
Lastly, Apportioned Effort (AE,) which measures performance based on the earnings of a related work package. A common example of AE would be Inspection, which could be X percentage of an item being manufactured.
Choosing the correct method requires knowledge of the duration of work and careful consideration of the type of work. With the appropriate methodology, earned performance can be a powerful tool for the success of delivering and managing any project.
Avoiding some of the pitfalls articulated here while building an EVM system and subsequently managing the data generated can be a challenge. However, the resulting outcomes of a well-managed and designed system are invaluable to any Project Management team. It can set the team on a successful path by helping it to anticipate issues and illuminates the appropriate course corrections. Of course, the information presented above is only a small sample of what goes into successfully managing a project using EVM. Schedule analysis and tracking costs present their own difficulties and no two systems are alike due to the work variants on every project. Spend the time necessary to review all the data with a control account manager and project managers to build confidence in managing your project.
