Abstract Data models are the backbone of digital information exchange, since they contain the structure of data to be exchanged. Just as information requirements vary, so do data models – in level of detail, level of abstraction, and domain coverage. Due to these differences, communication between some data models is easy and between others is difficult. Regarding the Building Information Model (BIM) initiative, the IFC standard data model varies in detail, abstraction level, and large domain coverage. In contrast, the Austrian ÖGG-guideline has a consistent detail and abstraction level focusing on a single domain, subsurface engineering. In order to participate in a loss- and distortion-free information exchange, a reliable translation via a multitude of data models is a must-have. In this paper, we present formal criteria for distinguishing between semantics-carrying data models, such as IFC and ÖGG, and translating data models that provide reliable communication bridges between them, such as CAEX and SIMULTAN. We will show that translating data models are an indispensable part of the data model infrastructure even within a single domain. For evaluation purpose, we demonstrate our approach on a subsurface engineering use case.