Was that webpage pleasant to use? Predicting usability quantitatively from interactions

Abstract

Webpage usability is crucial for customer satisfaction and loyalty. Yet, evaluations of webpages are usually tedious or do not provide sufficient information. Thus, we aim at providing a novel layout-independent framework for automatically predicting a quantitative measure of usability from user interactions. A study has shown that it is necessary to take into account differences in user intention and structural features already for very similar webpages. We propose preprocessing steps in terms of structure-based clustering and determining user intention, which will make it possible to provide meaningful usability models that support satisfaction and loyalty. © Springer International Publishing 2013.