Measuring Decision Quality Using Recommendation Agents

In addition to helping consumers make better decisions, the use of electronic recommendation agents provides a way for marketers to gather information and assess the quality of consumer decisions. The use and type of agent employed determines the types of measures of decision quality that can be employed by online providers. We examine the assumptions of and relationships among preference-dependent, preference-independent, and subjective measures of decision quality. Analysis of data from a series of experiments that simulate a broad spectrum of recommendation agent approaches shows that preference-independent measures capture about 63% of the variance of preference-dependent measures and that subjective measures capture 9% of the variance of preference-dependent and preference-independent measures. In addition, a superior and parsimonious measure of decision quality can be obtained by selectively combining these measures. Managerial implications for deploying electronic recommendation agents are discussed