Performance metrics, such as Equal Error Rate or Detection Cost Function, have been widely used to evaluate and compare biometric systems. However, they seem insufficient when dealing with real-world applications. First, these systems tend to include an increasing number of subsystems, e.g. aimed at spoofing detection or information management. As a result, the aggregation of new capabilities (and their interactions) makes the evaluation of the overall performance more complex. Second, performance metrics only offer a partial view of the system quality in which non-functional properties, such as user experience, efficiency or reliability, are generally ignored. In this paper, we introduce RoQME, an Integrated Technical Project funded by the EU H2020 RobMoSys project. RoQME aims at providing software engineers with methods and tools to deal with system-level non-functional properties, enabling the specification of global Quality-of-Service (QoS) metrics. Although the project is in the context of robotics software, the paper presents potential applications of RoQME to enrich the way in which performance is evaluated in biometric systems, focusing specifically on Automatic Speaker Verification (ASV) systems as a first step.