Customizable automatic detection of bad usability smells in mobile accessed web applications

Remote usability evaluation enables the possibility of analysing users' behaviour in their daily settings. We present a method and an associated tool able to identify potential usability issues through the analysis of client-side logs of mobile Web interactions. Such log analysis is based on the identification of specific usability smells. We describe an example set of bad usability smells, and how they are detected. The tool also allows evaluators to add new usability smells not included in the original set. We also report on the tool use in analysing the usability of a real, widely used application accessed by forty people through their smartphones whenever and wherever they wanted.