Data-based identification of knowledge transfer needs in global production networks

Abstract

Manufacturing companies’ value chains are increasingly distributed globally, which presents companies with the challenge of coordinating complex production networks. In general, these production networks grew historically rather than having been continuously planned, leading to heterogeneous production structures with many tangible and intangible flows to be coordinated. Thereby, many authors claim that the knowledge flow is one of the most important flows and the source of competitive advantage. However, today’s managers face major challenges in transferring production knowledge, especially across globally distributed production sites. The first obstacle to a successful knowledge transfer is to identify what kind of knowledge should be transferred between whom and at what time. This process can take months of information collection and evaluation and is often too time-consuming and costly. Thus, this paper presents an approach to automatically identify at what point knowledge should be transferred. In order to achieve this, the company's raw data is being used to identify which employees work on similar production processes and how these processes perform. Therefore, production processes, which can be compared with each other, need to be formed, even though these processes may be performed at different production sites. Still, not every defined cluster of production processes necessarily requires the initiation of knowledge transfer since performing a knowledge transfer always entails considerable effort and some processes might already be aligned with each other. Consequently, in a next step it is analyzed how these comparable production processes differ from each other by taking into account their performances by means of feedback data. As a result, trigger points for knowledge transfer initiation can be determined.