The optimization of regression testing (RT) has been widely studied in the literature, and numerous methods exist. However, each context is unique. Therefore, how to tell which method is appropriate for a specific industrial context? Recent work has proposed a taxonomy to aid in answering this question. The approach is to map both the RT problem and existing solutions onto the taxonomy, aiming to determine which solutions are best aligned with the problem. This paper presents a case study that evaluates the approach in a real setting. The context is the development of R&D projects at a major automotive company, in the domain of connected vehicles. We used the taxonomy to characterize the RT problem in terms of measurable effects, and to identify the technically feasible solutions from a set of 52 papers. We report on the beneficial aspects but also the difficulties of the approach, due to unclear taxonomy elements, missing ones and paper classification errors.