The analysis of usability aspects of multi-user systems, such as cooperative work systems and pervasive systems, pose particular problems because the group behaviour of their users may have considerable impact on usability. Model-based analysis of such features is difficult because of the sheer number of entities that need to be modelled, leading easily to state-space explosion problems when automatic techniques such as model checkers are used. In this paper we explore the use of a recently proposed scalable model-based technique based on solving sets of Ordinary Differential Equations (ODEs). Starting from a formal model specified using the Performance Evaluation Process Algebra (PEPA), we show how different groupware usage patterns may be modelled and analysed using this approach. To illustrate how the approach can contribute to get more insight in the suitability of different design options and their impact on group behaviour we compare two different file access policies in the context of the industrial groupware application thinkteam.