Mitigating information overload in e-commerce interactions with conversational agents

Abstract

Information overload influences users’ satisfaction and performance when completing a complex task. In e-commerce interactions, this has the effect that customers’ decision making becomes confused, less accurate and less effective. For websites, numerous countermeasures to mitigate information overload have been presented, whereas not many attempts have been made to reduce cognitive load when conversational agents are used instead. Conversational agents are expected to increase the perceived overload due to the voice interface characteristics. In this pilot study, the cognitive load of subjects was measured during an online shopping task which required different custom shopping skills for Amazon Alexa. It was tested if the countermeasure filtered repetition can reduce subjects’ perceived overload when using the voice assistant and which load differences can be found in comparison to a shopping website. To measure the mental load, the skin conductance level was recorded.