The current state of the discipline suggests that the desired fidelity of verification models is defined based on qualitative statements, that is, using heuristics, at the onset of an engineering endeavor. While the use of heuristics in systems engineering is well known, many heuristics have not been validated, and therefore there may be errors in judgement that could lead to systems being engineered that are ultimately not fit‐for‐purpose. This bears the question, based on what conditions should we formally define verification models? In this article, we provide insights from our research to discover the sufficient conditions for defining verification models, taking a science‐based approach that leverages the richness of systems‐theoretic constructs.