Exploring the Correlation between Predictive Usability Measures and User Tests

For software products to succeed they have to satisfy customer expectations, including usability aspects of the software. Early during the development cycle important design decisions are made for projects that are typically on restricted schedules and budgets. There is a need to support this decision making process, by empowering the decision makers with better tools and techniques. In this research we present a prediction approach that allows potential stakeholders to examine and choose among different user interfaces based on our usability prediction model. The usability predication model establishes correlations between predictive usability metrics and the results of usability tests performed by users. Based on our case studies we applied a multiple regression method to find relationships between response usability quality attributes and explanatory usability metrics.