The Value of IT-Enabled Retailer Learning: Can Personalized Product Recommendations (PPRs) Improve Customer Store Loyalty in Electronic Markets?

Abstract

Personalization is a strategy that has been widely adopted by online retailers to enhance their customers' shopping experience, with the ultimate goal of building a strong and enduring customer relationship. Personalized product recommendations (PPRs) are product recommendations adapted to individual customers' preferences and taste. So far, very few empirical studies have ever investigated the impact of PPRs from a consumer behavior perspective. Whether PPRs generate any value for consumers and ultimately, retailers, is still an open question.

To fill this gap in the literature, in this study, drawing upon the household production function model in the consumer economics literature, I develop a theoretical framework that explains the mechanism through which PPRs influence customer store loyalty in electronic markets. Online shopping can be viewed as a household production process and customer store loyalty is driven by shopping efficiency. Building upon retailer learning, higher quality PPRs can increase consumers' online product brokering efficiency, which in turn increases their repurchase intention. A two-phase lab experiment was conducted among 253 undergraduate students in the business school. The subjects completed a simulated purchase at Amazon.com and the quality of PPRs they received was manipulated. Empirical analyses indicate that higher quality PPRs improve consumers' online product brokering quality, which in turn increases their repurchase intention. Consumers make higher quality purchase decisions and experience more fun during the online product brokering process. A surprising finding is that higher quality PPRs increase consumer online product brokering cost. Consumers spend more time on decision making and have more difficulty reaching a purchase decision. Implications, limitations, and contributions of this study are discussed and areas for future research are suggested.