Recommendation agents (RAs) in interactive online retailing: an investigation on consumers’ shunning recommended products

Purpose This paper elucidated how recommendation agents (RAs) in online retailing elicit a counteractive perception among consumers from a perspective of control, which results in their intent to avoid RA recommendations. For marketers who attempt to reduce consumers’ negative responses in the context of RAs, the paper examined whether an RA empowered by algorithmic transparency would confer an effective solution. Design/methodology/approach One pilot test and two studies of scenario-based experiments were conducted to capture the proposed effect. Findings The findings support that sense of control (SOC) and avoidance intention (AV) are negatively related, where privacy concern (PC) mediates the relationship. Transparency empowerment exerts a detrimental effect on leveraging consumers’ SOC. The assertion that an increased SOC eliminates consumer avoidance is found to be substantially stronger in the absence of transparency. Research limitations/implications The research consolidates the compensatory control theory (CCT) and adds a novel perspective on the relationship between one’s control and PC, which has yet to be adequately addressed. It enriches empowerment literature by revealing that algorithm transparency would not significantly influence consumers’ avoidance of RAs’ recommendations, whereas non-transparent RAs would attenuate such impact. Practical implications This study entails significant implications for technicians and marketers, emphasizing the need to promote feelings of control through RAs while also highlighting the necessity to rethink the interpretability and transparency attributes in RAs. Originality/value This research contributes to the literature on human–technology interaction, documents key theoretical and managerial implications, and sheds light on the counterintuitive effects that span the literature on individual control and consumer empowerment.