Estimating the effects of ramp-up and learning from cost performance data

Purpose The US Department of Defense routinely receives monthly cost performance reports (CPRs) to monitor program status, justify payments and identify production issues. This paper investigates how ramp-up, disruption and learning affect costs in long-cycle, made-to-order defense production. Using CPR data from the US Army’s Black Hawk helicopter program, we estimate cost due to learning and production rate. We show how disruptions and experimentation altered production. Design/methodology/approach We estimate parameters of the Camm and Womer (1987) model of crew allocation from 133 months of CPR data spanning five Black Hawk lots (1977–1982). To separate learning-driven cost reductions from ramp-up and disruption effects, parameters are estimated from a later, stable period of production, then applied to earlier disrupted periods. Simulations compare predicted costs under optimal ramp-up with observed outcomes. Findings Results show a steep underlying learning curve of about 71% when production rate effects are considered. In contrast, observed curves appear flatter when ramp-up is ignored. Simulations suggest that better planning with smoother ramp-up could have reduced labor requirements by 13,000–15,500 worker months and program start could have been delayed by nearly two years. Practical implications Our approach provides defense managers with a new way to use CPR data to validate contractor cost-at-completion estimates early in programs. The model helps estimate the impact of disruptions, project labor requirements and evaluate schedule alternatives. While based on a historical case, the framework can be applied to current acquisition programs, offering insights into cost, schedule and risk trade-offs during ramp-up. Originality/value This research is the first to use CPR data for these purposes.