Various novel and data-driven business concepts have emerged during the fourth industrial revolution. Smart manufacturing, for example, utilizes data from manufacturing equipment, human operators, and organizational IT systems to enable dynamic adaptions in production systems. Nowadays, these data are often distributed among multiple partners in collaborative value creation networks. Hence, to identify and collect relevant data for given business cases has become an important, but complex issue. To support the process of establishing comprehensive data provision in industrial practice, a reference model for knowledge-driven data provision processes was developed. It describes a systematic approach to drive operationalization of data provision from knowledge requirements to identify, extract and provide raw data until the application of such data sets. To evaluate the applicability of the reference model, a case study was conducted in which it was used to guide the implementation of an IoT Solution in four Swedish manufacturing companies.