Corporate decision analysis : an engineering approach

Abstract

We explore corporate decisions and their solutions under uncertainty using engineering methods. Corporate decisions tend to be complex; they are interdisciplinary and defy programmable solutions. To address these challenges, we take an engineering approach. Our proposition is that as in an engineering system, corporate problems and their potential solutions deal with the behavior of systems. Since systems can be studied with experiments, we use Design of Experiments (DOE) to understand the behavior of systems within which decisions are made and to estimate the consequences of candidate decisions as scenarios. The experiments are a systematically constructed class of gedanken experiments comparable to "what if' studies, but organized to span the entire space of controllable and uncontrollable options. In any experiment, the quality of data is important. Grounded on the work of scholars, we develop a debiasing process for eliciting data. And consistent with our engineering approach, we consider the composite consisting of the organization, their knowledge, data bases, formal and informal procedures as a measurement system. We then use Gage theory from Measurement Systems Analysis (MSA) to analyze the quality of the measuring composite.

(cont.) To test this engineering approach to decision analysis, we perform four experiments. The first two are a set of simulations using a company surrogate. Using a progression of experiments, we simulate two major corporate decisions. Simulation data show that there is support for the validity of our decision analysis method. We then perform two in situ experiments: with a manufacturing company and with a technology services company. Findings from these company experiments also support the validity and efficacy of our decision analysis method. The company executives were very satisfied with our findings. Finally, we evaluate our method using method-evaluation criteria. The evaluation suggests that our DOE-based decision analysis method is valid. Unexpectedly every experiment resulted in near-decomposable systems at the scale we formulated our problems. Scaling of corporate decision problems at the appropriate level of abstraction and the resultant properties of their dynamic behavior are identified as areas of future work. This research breaks new ground in corporate decision-analysis as engineering and it furthers DOE and MSA research to a new domain and a new class of problems.