A Data-driven Design for Deriving Usability Metrics

The complexity of Enterprise Information Systems can be overwhelming to users, yet they are an often overlooked domain for usability research. To better understand the ways in which users interact with these systems, we have designed an infrastructure for input logging that is built upon a data model relating system components, user inputs, and tasks. This infrastructure is aware of user representations, task representations, and the history of user interactions. The interface components themselves log user inputs, so that timing data and action events are automatically aligned and are linked to specific tasks. The knowledge gained about user interactions at varying levels of granularity, ranging from keystroke analysis to higher-level task performance, is a valuable resource for both assessing and enhancing system usability.