The Impact of Position and Contrast Effects in Recommender Systems on Consumer Behavior: A Field Experiment

Lill, Markus und Spann, Martin (2025): The Impact of Position and Contrast Effects in Recommender Systems on Consumer Behavior: A Field Experiment. In: ACM Transactions on Recommender Systems [Forthcoming]

Abstract

This study investigates how position and contrast effects influence user decision-making in recommender systems (RS). Since these effects can induce behavioral biases that affect consumer choices, understanding their influence on RS effectiveness is crucial. We conducted a randomized field experiment, comparing recommendation sets with only relevant items and sets with one relevant item surrounded by non-relevant contrasting items, while varying the relevant item's position across the recommendation list. Our results show that both the position of a focal item and its distinction from contrasting products significantly increase click-through rates (CTRs). Relevant items in second and third positions achieve higher CTRs than first position, while non-relevant contrasting items increased the relevant item CTR despite reducing overall CTR. This suggests that users are more likely to engage with recommendations that stand out. Our results challenge the assumption that the top position in a recommendation list is always the most beneficial by showing that lower positions can lead to higher CTRs. These findings provide strategies for retailers to highlight specific products while revealing fundamental trade-offs in recommendation system design. The study provides valuable insights for optimizing RS design, highlighting the importance of strategic placement and diversity to improve decision outcomes.

• Item Type: Journal article
• Keywords: Recommender systems; User studies; Online shopping; Consumer behavior; Behavioral biases; Position bias; Contrast effects; Field experiment
• Faculties: Munich School of Management > Institute of Electronic Commerce and Digital Markets
• Subjects: 300 Social sciences > 330 Economics
• ISSN: 2770-6699
• Language: English