In this paper, we consider the value of using Hierarchical Task Analysis (HTA) as a grammar to map actions in context. The grammar approach aims to develop a richer picture of user actions in a task accommodating aspects of the socio-organizational context. This paper reflects on user tasks as emerged from interviews with real users in supply chain industries where Forecasting Support Systems (FSS) are used to predict future product sales. In this paper, we use HTA to describe what the expert literature has identified as the stages in producing sales forecasts. In order to evaluate this against reported scenarios and observations of actual use, we apply an approach that uses a parallel to the grammar of everyday language to represent user activities. In addition to specifying the particular socio-organizational aspects of FSS use, we show how the grammar approach contributes to effective analysis of unstructured, qualitative data. The methodology we adopt also eases our attempt to understand user tasks and validate the constructed HTA model in a given context. The methodological implications and future research directions for task analysis models are discussed.