Towards automated detection of mobile usability issues

Abstract

While evaluating the usability of mobile applications in the field has proven to lead to better results than in laboratory settings, in practice it is still not carried out after deployment - typically due to the required resources. In this paper we demonstrate a lightweight automated method for revealing specific usability issues of mobile applications in the field. Based on application usage data, we derive a simple heuristic which detects low discoverability by analyzing view transitions of mobile applications at runtime. We show the applicability and feasibility of our approach in a user study with a real application. Our results are promising and call for further research.