According to current research, one of the most promising applications for web usage mining (WUM) is in identifying homogenous user subgroups (Liu, 2008). This paper presents a prototypical workflow and tools for analyzing user sessions to extract business intelligence hidden in web log data. By considering a leading Swedish destination gateway, we demonstrate how query term analysis in combination with session clustering can be utilized to effectively explore the information needs of website users. The system thus overcomes many of the limitations of typical web site analysis tools that only offer general statistics and ignore the opportunities offered by unsupervised learning techniques.