Merchandising/Marketing/Retailing: Management

Do You Trust My Recommendations? Impact of Recommendation Agents' Filtering Method and Presentation Style on Consumers' Trust and Decision-Making

---

Abstract

Recommendation agents (RAs) are artificial intelligence algorithms that capture consumers' preferences and interests to give them personalized recommendations during online shopping. Online retail websites, such as Amazon and BestBuy, use RAs to help consumer decision-making by reducing the alternative filtering complexity and information overload. RAs may form their recommendations by employing two types of filtering methods-- content filtering or collaborative filtering-- and present them to consumers in two styles-- vertical listings or side-by-side comparative charts. However, the impact of these RA filtering methods and presentation styles on consumer decision-making is unknown. This research employed a 2 x 2 between-subjects online experiment (n = 306) to test whether the RA filtering methods and presentation styles elicit different levels of consumer trust in the recommendations, reduce consumers' perceived decision effort, and increase perceived decision quality. Structural equation modeling results revealed that consumers trusted content filtering RAs and comparative presentation style more than collaborative filtering and listing style, respectively (both ps < .05). Both the filtering method and presentation style of RAs had significant indirect effects on PDE and PDQ through consumer trust. These findings illuminate the importance of RA filtering methods and presentation styles on consumer decision-making efficiency.

Keywords: Recommendation agents, trust, decision effort, decision quality, content filtering, collaborative filtering

How to Cite:

Harrison, E. N. & Kwon, W., (2022) “Do You Trust My Recommendations? Impact of Recommendation Agents' Filtering Method and Presentation Style on Consumers' Trust and Decision-Making”, International Textile and Apparel Association Annual Conference Proceedings 79(1). doi: https://doi.org/10.31274/itaa.15907