Work

How should conversational recommendation agents organize candidate products? Matching recommendation set granularity with time pressure

Year: 2026

Type: article

Abstract: With the widespread application of conversational recommendation agents (CRA) on e-commerce platforms, an increasing number of consumers use a CRA to support product choices. Existing research has mainly focused on the accuracy of CRA recommendations, with limited attention paid to how candidate products should be organized and whether this organization should be adapted to consumers' current decision contexts. Drawing on the product set granularity perspective and the heuristic–systematic model, this study examines the matching effect between recommendation set granularity (coarse vs. fine) and time pressure (high vs. low) on consumers' intention to adopt CRA recommendations and investigates the underlying psychological mechanisms. The results of three online scenario experiments reveal that consumers show higher adoption intention toward CRA recommendations featuring coarse recommendation set granularity under high time pressure and toward those featuring fine recommendation set granularity under low time pressure. The two matching effects operate through perceived fluency and perceived attractiveness, respectively. Furthermore, consumer knowledge affects the relative effectiveness of the two matching strategies: the match between coarse recommendation set granularity and high time pressure is more effective for consumers with low product knowledge, whereas the match between fine recommendation set granularity and low time pressure is more effective for consumers with high product knowledge. These findings extend CRA research from recommendation content to the organization of candidate products and provide guidance for retail platforms to adjust recommendation set granularity according to consumers’ time pressure and product knowledge.

Source: Journal of Retailing and Consumer Services

DOI: 10.1016/j.jretconser.2026.105165

Authors: Qi Wang, Wen Tong, Tian Yu Wei, Xiao Qian Fan, Wen Zhang

Institutions: Hubei University of Technology, Zhongnan University of Economics and Law

Language: English

Topic: Recommender Systems and Techniques

Subfield: Information Systems

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

Estimated APC paid: $4,510

Funders: National Social Science Fund Youth Project, National Social Science Fund of China