Modelling-based teaching activities have been designed and analysed from distinct theoretical perspectives. In this paper, we use one of them--the model of modelling diagram (MMD)--as an analytical tool in a regular classroom context. This paper examines the challenges that arise when the MMD is used as an analytical tool to characterise the modelling process experienced by students working in small groups aiming at creating and testing a model of a sedimentary basin from the information provided. The study was conducted in a regular Biology and Geology classroom (16-17 years old students). Data was collected through video recording of the classes, along with written reports and the material models made by each group. The results show the complexity of adapting MMD at two levels: the group modelling and the actual requirements for the activity. Our main challenges were to gather the modelling process of each individual and the group, as well as to identify, from students' speech, which stage of modelling they were performing at a given time. When facing such challenges, we propose some changes in the MMD so that it can be properly used to analyse students performing modelling activities in groups.