Software Quality Assurance involves all stages of the software life cycle including development, operation and evolution as well. Low level measurements (product and process metrics) are used to predict and control higher level quality attributes. There exists a large body of proposed metrics, but their interpretation and the way of connecting them to actual quality management goals is still a challenge. In this work, we present our approach for modelling, collecting, storing and evaluating such software measurements, which can deal with all types of metrics collected at any stage of the life cycle. The approach is based on the Goal Question Metric paradigm, and its novelty lies in a unified representation of the metrics and the questions that evaluate them. It allows the definition of various complex questions involving different types of metrics, while the supporting framework enables the automatic collection of the metrics and the calculation of the answers to the questions. We demonstrate the applicability of the approach in three industrial case studies: two instances at local software companies with different quality assurance goals, and an application to a large open source system with a question related to testing and complexity, which demonstrates the complex use of different metrics to achieve a higher level quality goal.