The use of multiple performance criteria which account for both the static and dynamic aspects of marketing systems is proposed as a necessary step toward scientific marketing analysis. Currently, over-all evaluations using multiple performance measures are difficult to make and even more difficult to interpret meaningfully. Fisher's technique of discriminant functions is utilized to estimate the mathematical relationship between the relevant marketing performance measures which best "matches" subjective classifications agreed upon by a number of business executives. An application of the technique involving the evaluation of a consumer product's performance in various market areas is discussed. The problems of bias and multicollinearity, their relationship to the conceptual approaches, and their effect on the use of the performance measure are also discussed using the particular problems encountered in the application to illustrate the various points.
