Pre-Test-Market Models: Validation and Managerial Implications

Abstract

The predictive accuracy of a widely used pre-test-market model (ASSESSOR) is analyzed. The standard deviation between pre-test-market and test-market shares is 1.99 share points before adjustments for achieved awareness, distribution, and sampling and 1.12 share points after adjustment. Sixty-three percent of those products tested passed the pre-test screen and 66% of these were subsequently successful in test market. A Bayesian decision analysis model is formulated and a “typical” case shows a positive value of information. Although some conditions are identified under which a test market may be bypassed, in the authors’ opinion both pre-test and test-market procedures should be used in all but exceptional situations.