The occurrence of a particular state is a predictor of successful travel consultation

Abstract

Dyadic interactions, such as consultations between clerks (i.e., travel agent sales staff) and customers, involve multiple behaviors that reflect people’s internal states, such as interests and preferences. Thus, the time-series patterns of multiple behaviors between two people reveal the characteristics of the interaction. In this study, we conducted an experiment on travel consultations between clerks and customers, which was as close as possible to a real situation, and analyzed the time-series shifts of multiple behaviors in a 30-minute dyadic interaction using a hidden Markov model. We found differences in interactions between successful and unsuccessful consultations. In particular, in the successful interaction, a customer entered an interactional state associated with interest in the content of the consultation in the first 10 minutes, and such a state was likely to persist during the consultation. These findings suggest that early interactional engagement may serve as a potential behavioral marker associated with consultation success and offer a preliminary process-level perspective on interaction dynamics in naturalistic dyadic consultations.