The cost of modelling existing industrial facilities currently counteracts the benefits these models provide for maintenance and retrofit of these assets. 90% of the modelling cost is spent on labor for converting point cloud data to 3D models, hence reducing the cost is only possible by automating this step. The highly-cluttered scene and large number of industrial objects increase the required modelling time. Therefore, modelling is prohibitively expensive. We tackle a part of this issue by identifying the most frequent object categories and object types that require modelling in industrial plants to guide future work aimed at automating the tedious current practice. The industrial facilities investigated are: (a) offshore platforms, (b) food processing facilities and (c) petrochemical plants. The industrial object types obtained from BIM models are hierarchically ordered based on their frequency of appearance and modelling intent. The results showed that structural elements, the piping system and electrical equipment were the most frequent object categories encountered in all case studies. The most frequent object types in these categories are then determined by implementing a statistical analysis on their frequency of appearance in all case studies. The modelling intent of the most frequent object types in these categories is then explored to determine the most important object types. These are in descending order: electrical conduit, straight pipes, circular hollow sections, elbows, channels, solid bars, I-beams, angles and flanges. Automatically modelling these frequent and critical object types can guide future researchers interested in modelling these assets.