Improving flow is a core Operations Management theme that is set to become even more important following contemporary developments in manufacturing, such as smart products and digital encapsulation that enable new control concepts such as multi‐agent holonic control. But, companies often struggle to realize flow improvements in practice, both with and without new technologies. While the literature agrees on the importance of flow, a structured and independent process that supports managers in identifying the root causes of why flow items wait in inventories instead of being processed is missing. Managers often use a single production management concept, such as lean production or the theory of constraints, when they seek to understand the reasons for a flow problem, which may lead to misdirected and unsuccessful interventions. In response, we use design science to develop a comprehensive approach to diagnosing flow problems that is independent from any production management concept. This diagnosis process results from successive iterations with five companies and supports the selection of appropriate analytical models and flow improvement solutions. It enables an organization to widen the focus of its flow improvement actions beyond the scope of a singular production management concept and complements the application of recent advances in technology, allowing smart products to quickly interpret what is happening in a location without first simulating and analyzing the whole system. Furthermore, the study expands buffer theories by showing that buffers have an internal hierarchy and can be absorbed by other buffers, while enhancing other theories related to coordination, material flow control, and lean improvement.