AIS efforts often consisted of "software maintenance" planned and managed as a level of effort. Annual budgets funded teams working on system problems, upgrades and sometimes completely new applications. In this environment the development team generated as much product as possible for the money available that year and moved any remaining work to the next year. Additionally, the customer often had poorly defined or constantly changing requirements. Without specific and controlled product definition there could not be a product, or project, management mentality.
Today many believe earned value represents a natural extension of any structured software development effort, including automated information systems and information technology. Earned value provides a useful tool for any effort that results in products, the normal situation for software.
Current significant changes in both disciplines (earned value and automated information systems) make the melding of the two a natural result. Recognition of earned value as a program management tool moved it from a burden to a benefit. Pressures for improved effectiveness and accountability on the AIS community resulted in applying disciplined software development and resource management processes. This move toward more structure in the software development process encourages metrics including use of earned value management, while the need for resource management makes earned value a critical tool. By its nature the earned value process creates a disciplined approach that ties directly to other program activities such as configuration management and software metrics.
The concept of earned value relies on three principles: integration, planning, and evaluation. Integration refers to always considering the three parameters associated with project performance:
The integration principle requires identifying the work effort, schedule, and resources such that the three parameters can be integrated at the day to day management level. It also requires a control process to assure the three elements remain consistent and tie to the overall program requirements. For our example the integration would be defined as follows:
Develop 5 modules of software:
Plans project completing 2 in first month, 2 in second month, 1 in third month
Resource plan:
1000 hours per month (diamonds)
The planning principle requires accounting for all work to be accomplished, identification and allowance for risk, and determining objective methods of measuring work completion. The planning forces the integration process down to the point at which work accomplishment can be traced. Planning also implies a continuous process to reflect improved knowledge over time due to the typical growth of knowledge over time or the evaluation process. In our case the planning process allocates the resources to specific modules to allow evaluation of performance.
Finally, the evaluation principle requires examining the performance to date, identifying the cause and effect of deviations to the plan, and reflecting this knowledge back to the planning. The evaluation process thus integrates the status of the three elements in an objective manner.
Status after 2 months:
1,600 hours spent.
Modules 1, 2 and 4 complete - Earned Value = 1,400 (triangles)
This simplified example addresses the point that earned value allows differentiating cost and schedule problems through assigning values to the work accomplished. Spending less than the original plan only defines a positive situation when also accomplishing all the planned work. In this case the spending less than planned results from falling behind schedule. This evaluation compares the value of the work accomplished (1,400) against the value of the work planned (2,000) to show under accomplishment. By comparing the value of the work accomplished (1,400) against the cost spent (1,600) we recognize that the effort to accomplish the finished work exceeded plan. Thus, the effort shows as behind in schedule and exceeding planned cost.
Earned value effectively supports software development management since the structured planning and control required by earned value ties closely with the currently recommended software development processes and metrics. The major reluctance in applying earned value often comes from the concern that successful completion of interim steps may not predict successful performance. True! However, unsuccessful completion of current efforts definitely predicts unsuccessful performance. The key lies in the planning itself.
The discipline of a detailed plan and control of changes moves decisions about the software ahead of the development process. This identifies risks and forces evaluation of schedule and cost impact of changes. Thus, earned value cannot assure success in a software effort. It will, however, provide insight to the probability of success, force management of change, and allow objective evaluation of status.
For more information on earned value and software development contact:
| Undersecretary of Defense for Acquisition & Technology Earned Value Management Web Page http://www.acq.osd.mil/pm/ |
NNH Enterprise - Earned Value Page http://www.nnh.com/ |
| Air Force Acquisition - Earned Value Page http://www.acq.osd.mil/pm/ |
Earned Value Web Site hosted by Cost Management Systems, Inc. http://www.earnedvalue.com/ |
