Examining the effects of malfunctioning personalized services on online users' distrust and behaviors

Merchants adopt web personalization technologies to offer product recommendations in the hope of influencing online users' decision making in a shopping process. Although there is a large body of research on the favorable effects of web personalization on influencing online users' decision making, it is often assumed that web personalization functions well. Only scant research examines the adverse outcomes of web personalization on online users' perceptions and behavior when the personalized services malfunction. This research aims to fill this gap. Specifically, we examine malfunctioning personalized services that produce irrelevant and biased product recommendations in online shopping. Irrelevant recommendations are offerings not matched to online users' preferences, whereas biased recommendations are offerings generated for the merchant's interests. When online users encounter such malfunctioning personalized services, they may distrust the personalization agent, which influences their interactions with the agent. We drew on distrust theories to develop six hypotheses. To test the hypotheses, 245 participants were recruited for a field experiment in which they were tasked to download free music tracks from a personalized music download website. Our findings indicate that both irrelevant and biased recommendations lead to high distrust in a personalization agent's competence and integrity. Competence distrust, but not integrity distrust, in a personalization agent negatively influences online users' interactions with the agent.