Researchers and practitioners have recently paid great attention to research and development (R&D) performance measurement, although it is acknowledged to be a very challenging task because of R&D intrinsic uncertainty and complexity levels. In this paper, the problem of designing a performance measurement system (PMS) for R&D activities is addressed; specifically, we investigate if and how the design of the PMS is influenced by the type of activity it is applied to, namely Basic and Applied Research or new product development (NPD). We first develop a theoretical framework that comprises the main constitutive elements of a PMS for R&D. Then the framework is used for supporting a multiple case study analysis involving eight Italian technology-intensive firms. The research results show that the criteria for designing the constitutive elements of the PMS are radically different in Basic and Applied Research and NPD. The reasons behind the observed dissimilarities in the design criteria are widely discussed in the paper, as well as their implications for R&D managers.