The contribution of this paper is threefold. Firstly, we measure the incidence of user innovation in a broad sample of firms. Previous work has collected repeated evidence on the frequency of user innovation in a variety of industries and products, but so far its incidence has not been demonstrated in samples of larger business populations. Secondly, we assess if current innovation surveys adequately capture user innovation. Surveys such as the CIS (Community Innovation Survey) take a producer perspective and seem to overlook that in practice many innovation efforts are done by users to satisfy their process needs. Thirdly, we explore to what extent user innovations are transferred to producer firms. In doing so we assess if user innovation is marked by voluntary spillovers which is a strong argument to justify policies for user innovation. Drawing on survey data of 2 416 SMEs in the Netherlands, we find that 21% of all SMEs engage in user innovation, i.e. they develop and/or significantly modify existing techniques, equipment or software to satisfy their own process-related needs. We also find that user innovation is remains largely invisible in the current innovation surveys. Next, in a survey of technology-based small firms in the Netherlands we identified 364 specific user innovations. We found that users tend not to patent or protect their innovations, and that one out of four is transferred to producers. The data suggest a significant feedstock of voluntary knowledge spillovers from users to producer firms. We conclude that future innovation surveys should explicitly capture user innovation, and develop some recommendations to guide this effort. We also plea for more research on policies for user innovation.