With a declining number of nuclear data evaluators in the world and an increasing demand for high-quality data, there is a risk that evaluators will concentrate on producing new nuclear data to the detriment of developing new models and methods for evaluating existing data. In this context, it is essential to identify the basic physics issues that are going to be important for future nuclear data evaluation processes. At the same time, demand for new types of data, which will be needed in emerging nuclear applications, could warrant new evaluation techniques that are presently only used in the context of fundamental research and not in nuclear data production.  These proceedings present the main findings of the "Perspectives on Nuclear Data for the Next Decade" workshop, which explored innovative approaches to nuclear data evaluation with the aim of opening new perspectives, building new research programmes and investigating prospects for international collaboration.