Camille Charles, Louvain Research Institute in Management and Organizations
Caroline Ducarroz, Louvain Research Institute in Management and Organizations
Corentin Vande Kerckhove, Louvain Research Institute in Management and Organizations
Summary
Recommendation agents (RAs) help consumers navigate extensive online product assortments by providing personalized recommendations aligned with their preferences. Yet, an increasing number of online merchants now use biased RAs - systems that integrate not only consumers’ preferences but also the merchant’s strategic objectives. Among the forms of bias that can be incorporated into RA’s recommendations, society-centered bias - such as nudging consumers toward healthier food options - has become particularly salient as retailers position themselves as health-oriented actors. Despite its growing prevalence, little is known about whether consumers recognize such practices or can detect these subtle forms of bias. This research empirically investigates consumers’ ability to detect bias in RA’s recommendations. We test a theoretical model examining how bias intensity in recommendations affects bias detection performance and how this relationship is moderated by consumers’ experience with the RA. We also investigate the psychological mechanisms involved through the mediating roles of perceived personalization and perceived usefulness, and analyze how perceived bias influences consumers’ attitude toward the online merchant. Drawing on agency theory and signal detection theory, five hypotheses are tested across two experimental studies using a fictitious e-commerce platform with algorithm-based recommendations. By identifying when and how consumers detect society-centered bias, this work advances understanding of consumer vulnerability to online merchants’ business practices. It contributes to research on personalization, algorithmic fairness, and digital marketing by clarifying the mechanisms of bias detection and their consequences for consumers’ attitude toward the online merchant, and by identifying the threshold at which health-oriented nudges become noticeable. From a managerial perspective, it highlights the need for greater transparency and enhanced digital literacy to safeguard consumer autonomy in data-driven marketplaces.
Keywords: Recommendation agent, Personalization, Bias detection, Manipulative practices, Society-centered bias
Ophélie Duquesne, Louvain Research Institute in Management and Organizations
Caroline Ducarroz, Louvain Research Institute in Management and Organizations
Simon Hazée, Louvain Research Institute in Management and Organizations
Summary
In an omnichannel context where channel-based price differentiation is widely adopted by retailers, transparent explanations for the motives underlying price discrepancies across online and offline channels play a critical role in shaping consumer responses. Drawing on signaling theory, this multi-study research investigates how and when retailers’ communication practices foster consumer acceptance of higher online than offline prices. Study 1 assesses the relative effectiveness of different communication approaches - emphasizing retailer’s cost structure, consumer added value, and societal considerations - and identifies the communication practice that most effectively enhances consumers’ perceptions of price fairness. Study 2 further examines whether the magnitude of the price difference (low vs. high) between channels moderates the impact of these communication strategies on perceived price fairness. Altogether, these findings provide theoretically grounded and managerially relevant insights to help retailers enhance perceptions of price fairness when adopting channel-based price differentiation.
Keywords: Price fairness, Communication practices, Channel-based price differentiation