The progression towards data-driven manufactu : A case study of four Swedish factories

Abstract: Today’s technologies, such as Artificial Intelligence, Internet of Things and Digitals twins, have made data very valuable for most industries. The manufacturing industry is not exempted from this fact. In recent years there has been a lot of literature addressing the topic of utilizing data within manufacturing, which is also referred to as data-driven manufacturing. But there is a lack of literature exploring the concept of data-driven manufacturing in practice. Therefore, this paper aims to bridge the gap between theory and practice.  In doing so four Swedish factories have been studied. Two of the factories are a part of manufacturing firms that are considered large manufacturing firms, while the other two factories are a part of manufacturing firms that are categorized as SMEs. Semi-structured interviews were conducted with employees from the different factories. A thematic analysis was then conducted on the interview data. By combining the empirical findings, with the Maturity Model for Data-Driven Manufacturing (M2DDM) as a framework for digital progression, the progression towards data-driven manufacturing for the factories was assessed. Additionally, industry experts and researchers were interviewed in order to get a more nuanced view on the situation, by including a larger variety of sources of data.  The study showed that none of the factories had come especially far in their progression towards fully utilizing data-driven manufacturing. Keeping this in mind, it is worth noting that the factories associated with larger manufacturers had come further than the factories associated with smaller manufacturers. Mainly, due to the smaller manufacturers outsourcing functionalities such as IT and maintenance, but also due to the smaller manufacturers having less economical resources. Generally, the four factories had, at this moment of time, a lack of necessary competencies needed for achieving the end goal of self-optimizing and self-learning factories, such as dedicated data scientists with knowledge in machine learning, and other such areas. Lastly, the study explored the discrepancy between theory and practice. Theory suggests clear competitive advantages with fully utilizing data-driven manufacturing. Meanwhile, the empirical findings show that, in practice, there is no urge to achieve such competitive advantages, rather there is an impression that data-driven manufacturing will be achieved with time, as old machinery gets replaced by new machinery.