Organizations require personalized solutions to effectively address users’ needs, and stay competitive in the market. In this context, configurable systems offer numerous configuration options to meet user-specific functional and non-functional requirements. However, although configurability makes these systems flexible and versatile, a simple change can result in serious problems in different software variants, such as performance bottlenecks and security issues. Thus, automated approaches based on machine learning have been developed to facilitate configuration management. Our work aims to expand upon previous findings in this field by assessing their applicability to other scenarios. By introducing more efficient practices, we can contribute to cost reduction, higher software quality, and quicker time-to-market. This is particularly relevant in a global context where software plays a crucial role.