"Measuring the innovation performance of regions or nations has been repeatedly done in the literature. What is missing in the literature is a discussion of what innovation performance of a region means. How do regions or nations contribute really to the innovation output of firms? And how can this contribution be investigated in an empirically sound way? We argue that while the literature offers many suggestions, their theoretical foundation is often weak and the underlying assumptions are rarely discussed. In this paper, we systematize various mechanisms by which spatial units influence firms' innovation activities. On the basis of this, common innovation performance measures and analyses are discussed and evaluated. It is concluded that there is no general best way of measuring the innovation performance of spatial units. In fact, the most interesting insights can be obtained using a multitude of different approaches at the same time."