Techniques for cost estimating in early program phases

Abstract

In the early concept phases of programs, some approximate cost figures are often required to scope system program costs for development, production or life cycle, and to provide preliminary identification of cost drivers. The cost estimation process in these early, formative program stages is, however, hampered by insufficient knowledge of mission, configuration and environmental parameters. It is therefore necessary to use a heuristic approach, i.e., to fall back on cost estimates using modifiers of analogous systems where historical data exist. This paper describes the process of eliciting cost estimates and modifying cost assumptions which enable cost prediction of not well defined systems. It describes what kinds of cost models are acceptable and usable and how the data may be efficiently obtained. The choice of the ‘right’ experts (generalists, people with perception), the proper interviewing techniques (written versus oral, delphi, argumentative, etc.), the important role of the interviewer or moderator (reinforcing behaviour versus objective criticism), and the transformation of question responses into cost model inputs are discussed. The paper reveals the thought process and includes several examples of actually performed, early concept phase cost estimates and contains a list of several do's and dont's the author experienced in arriving at these early cost predictions.