Determining the Value of a Prototype

Abstract

Most major defense acquisitions require a technology that is not yet fully developed, introducing a non-negligible amount of risk to the program and its cost. Developing a prototype or technology demonstrator prior to execution of the program can be useful in mitigating this risk, yet these demonstration programs also have associated costs. This paper develops a method to value this risk mitigation, setting an appropriate maximum cost for the demonstration program. This novel application of Value of Information theory and properties of the Bayesian preposterior distribution requires only a program cost estimate distribution and some estimate of possible results of the demonstration program. The method is broadly applicable to programming with varying amounts of technological uncertainty. We describe the method, then show how actual cost overruns of historical programs with and without prototypes can be used to estimate a value of prototype efforts relative to estimated program cost. We conclude with a discussion of other applications and how to explain the method and results to decision-makers.