Decision Support and Cost Analysis
Domain 6 tests the analytic tools used to price requirements, compare alternatives, and track performance: cost estimating, economic analysis, life-cycle cost, and earned value management (EVM).
Estimating methods
- Parametric: a cost-estimating relationship built from historical data, such as cost per pound; fast, but valid only inside the database's range.
- Analogy: adjust a similar program's actuals for size and complexity.
- Engineering (bottom-up): build from work breakdown structure elements; most defensible, most resource-intensive.
- Extrapolation from actuals: use the program's own history as it matures.
Realism checks include learning curves (an 85 percent curve means each doubling of quantity costs 85 percent of the prior unit's labor), escalation into then-year dollars, and sensitivity analysis on the dominant cost driver.
Life-cycle cost and economics
Life-cycle cost covers research and development, procurement, operation and support (O&S), and disposal; O&S frequently dominates. Alternatives are compared in constant dollars using net present value with the OMB Circular A-94 real discount rate; mixing then-year dollars across alternatives or ignoring the time value of money invalidates the comparison. Cost as an independent variable sets affordability targets and trades requirements against them.
Earned value
Planned value (PV), earned value (EV), and actual cost (AC) yield cost variance (CV = EV − AC), schedule variance (SV = EV − PV), cost performance index (CPI = EV / AC), and schedule performance index (SPI = EV / PV). A CPI below 1.0 means work is costing more than planned; a simple independent estimate at completion is EAC = BAC / CPI.
Worked example
Budget at completion is $10,000,000; EV is $4,000,000; AC is $5,000,000; PV is $4,500,000. CPI = 4.0 / 5.0 = 0.80, so the program is running 20 percent above plan on cost. EAC = 10,000,000 / 0.80 = $12,500,000, a $2,500,000 overrun. SPI = 4.0 / 4.5 = 0.89, so the work is also behind schedule. The decision-maker should identify the variance driver before deciding whether to re-baseline.
Traps that cost marks
Comparing alternatives in mixed dollar bases, using average rather than marginal cost for an added unit, reading SPI as a cost measure, and applying a learning curve beyond its valid range.