The paper describes the conceptualisation process in the ReQuest approach, a wide-spectrum methodology for web-based information systems analysis and development. This methodology includes a strong involvement of end users in the requirement elicitation process by building prototype user interface fragments of the future application. The paper focuses on the analysis step of these fragments that yields a draft conceptual schema of the application domain. The analysis includes a tree-based representation of the fragments, the detection of shared subtrees through mining techniques, their normalisation and the derivation of the conceptual schema. A short description of a supporting tool is given.