Evaluating the usability of smartphone applications is crucial for their success, so developers can learn how to adapt them considering the dynamicity of mobile scenarios. The HCI community recommends considering different requirements when evaluating those applications, such as quantitative data (metrics), subjective evaluation (users’ impressions) and context data (e.g. environment and devices conditions). We observed a lack in the literature of approaches that support those three requirements combined into a single experiment; generally one or a pair of them is used. Besides, performing usability evaluation on real mobile scenarios is hard to achieve and most proposals are based on laboratory-controlled experiments. In this paper, we present our proposal for a hybrid usability evaluation of smartphone applications, which is composed by a model and an infrastructure that implements it. The model describes how to automatically monitor and collect context data and usability metrics, how those data can be processed for analysis support and how users’ impressions can be collected. An infrastructure is provided to implement the model allowing it to be plugged into any smartphone Android-based application. To evaluate our proposal, we performed a field experiment, with 21 users using three mobile applications during a 6-month period, in their day-to-day scenarios.