“Hey Google, which smart speaker should I use?” Research on factors that should influence the decision on which smart speaker to use

Abstract

This research for Valsplat focused on gathering more insight into the usage of voice assistants to be able to give better advice to clients. The main research question that has been answered with this research is: “When will a multimodal interface provide a better user experience than a unimodal interface for a specific use case?”. The literature study, focused on characteristics of voice, audio, visuals and cognitive theories, showed that there are various use case characteristics that should play a decisive role in developing for a uni- or multimodal interface. Based on the literature study a classification has been made, consisting of several use case characteristics that should influence the decision on which type of interface to use and several contextual factors that should also be taken into account. The classification can be used as an informative tool for the decision-maker. An experiment has been conducted with 85 participants that interacted with either a uni- or multimodal interface. The participants conducted two tasks, after which they assessed their satisfaction level using a questionnaire consisting of four questions about task completion, time, effort and subjective level of satisfaction. The experiment only tested the complexity characteristic of the classification. A higher level of satisfaction was expected for the more complex task when interacting multimodally. The result showed no significant difference between the two smart speakers. This means that user satisfaction did not increase when the participants interacted multimodally and the task became more complex. It can be concluded that there is a large grey area when choosing for the most appropriate smart speaker since it is not black-and-white when a uni- or multimodal interface should be used. Experimenting with the specific use case at hand might give the best insights.