Industry 4.0 and Lean – Possibilities, Challenges and Risk for Continuous Improvement An explorative study of success factors for

Industry 4.0 implementation

Joel Larsson & Johan Wollin

Supervisor Johanna Börrefors Karlskrona, Sweden June 2020

DEPARTMENT OF INDUSTRIAL ECONOMICS www.bth.se/mba

MBA Thesis

This thesis is submitted to the Department of Industrial Economics at Blekinge Institute of Technology in partial fulfilment of the requirements for the Degree of Master of Science in Industrial Economics and Management. The thesis is awarded 15 ECTS credits.

The author(s) declare(s) that they have completed the thesis work independently. All external sources are cited and listed under the References section. The thesis work has not been submitted in the same or similar form to any other institution(s) as part of another examination or degree.

Author information:

Joel Larsson joel.larsson@gknaerospace.com

Johan Wollin johan.wollin@volvo.com

Department of Industrial Economics Blekinge Institute of Technology SE-371 79 Karlskrona, Sweden

Website: www.bth.se Telephone: +46 455 38 50 00 Fax: +46 455 38 50 57

i

Abstract

Lean, with its origin in the Japanese automotive production and Toyota, is broadly seen as the most adopted manufacturing philosophy since several decades. One of the core values of Lean is Continuous Improvements (CI). CI is about the many small, simple and cheap improvements, which everyone is involved in, every day. As digitalization is making its way into the manufacturing environment, a hype around what is called Industry 4.0 (I4.0), also known as the fourth industrial revolution, has been created. In short, I4.0 refers to different technology-driven changes in an organization’s manufacturing systems. However, the true implications of those changes remain dimmed by the single-sided discussion of I4.0’s portrayed conceptual benefits. Moreover, despite the importance of CI for corporate success and the overall relevance of the I4.0 topic, no studies have been found to address their potential interaction. Thus, the purpose of this thesis has been to explore how the conditions for CI will be affected post I4.0 implementation. By focusing on the potential negative impacts on the overall rate of improvement, the purpose has also encompassed the identification of specific success factors to mitigate these negatives.

Due to the explorative nature of the research, a two-iteration Delphi survey containing open-ended questions has been chosen as a means of data collection; targeting experienced Lean and I4.0 personnel within the manufacturing industry. The first iteration survey encouraged participants to identify both positive and negative aspects of I4.0 impact on CI, while the second iteration survey encouraged participants to identify success factors. To make the concept of I4.0 more tangible, the technologies have been condensed into three I4.0 value drivers: Connectivity, Intelligence and Flexible automation. The data processing revealed that 64% of the answers provided for all value drivers were positive. This indicated an overall positive belief in the impacts of I4.0 on the conditions for CI both through enhanced problem-sensitivity, built by Lean values and principles, and through increased problem-solving capabilities. While the results reflected the current I4.0 hype, they also highlighted the difficulty in critically assessing the potential impact from technologies that are not yet widely implemented. Nevertheless, for each value driver the participants have also identified Challenges (18% of answers) and Risks (18% of answers) that can adversely affect CI. Based on the Challenges and Risks, a total of 74 success factors have been compiled and divided into four categories: Purpose, Involvement of people, Competence and Implementation strategy.

This research has contributed to the discussion about the Possibilities, Challenges and Risks of the I4.0 value drivers’ impact on the conditions for CI in the manufacturing environment. Furthermore, with emphasis on the identified Challenges and Risks, the authors have tried to cut through the noise of the ongoing I4.0 hype. As such, this research has introduced an alternative perspective that sets it apart from the overwhelmingly uncritical discussions surrounding I4.0. While the research’s theoretical contribution has been built by the insight into the I4.0’s potential impact on the conditions for CI, its practical contribution has been derived from the identified success factors; factors that can work as guiding principles for I4.0 adopters. Keywords: Industry 4.0, I4.0, Lean, Continuous Improvement, CI

ii

Acknowledgements

Sincere gratitude to our supervisor Johanna Börrefors at BTH for constructive and positive feedback during the duration of this project. We would also like to thank Anders Wrenne at BTH for fruitful discussions at the start of the project. Due to the nature of the study and the design of Delphi studies, the expert panel is to remain anonymous before, during and after the study. We would however like to extend a big thank you to the experts at GKN Aerospace, Volvo CE, I4.0 Lighthouse companies as well as Academic representatives for taking the time to participate in this study – without your input this research would have not been realized! We would also like to thank Johan Svenningstorp at Volvo Group, Lina Stålberg, Anna Sannö and Marcus Bengtsson at Volvo CE, Professor Umit S. Bititci at Heriot Watt University, Professor Torbjörn Netland at ETH Zürich, Victor Eriksson at Chalmers University of Technology and Anna Löfgren at Cerner Corporation for their insightful input during the preparation and conduction of this study. Gothenburg, 7th of June 2020

Joel Larsson and Johan Wollin

iii

Table of contents

1. Introduction _____________________________________________ 1

1.1. Introduction to Continuous improvement and Industry 4.0 ________________________ 1 1.2. Problem discussion_____________________________________________________ 2 1.3. Problem formulation, purpose and research questions ___________________________ 4 1.4. Theoretical relevance ___________________________________________________ 4 1.5. Practical relevance _____________________________________________________ 5 1.6. Delimitations _________________________________________________________ 5 1.7. Structure of the thesis __________________________________________________ 6

2. Background and literature review _____________________________ 7

2.1. Background of Lean ____________________________________________________ 7

2.1.1. The origins of Toyota Production System (TPS) ______________________________ 7

2.1.2. The Toyota Way _______________________________________________________ 8

2.1.3. Toyota Production System ______________________________________________ 9

2.1.4. Waste elimination and Muri, Mura and Muda (3M) __________________________ 12

2.2. CI implementation barriers ______________________________________________ 12 2.3. Background of Industry 4.0 ______________________________________________ 14 2.4. Industry 4.0 technology value drivers _______________________________________ 15

2.4.1. Connectivity _________________________________________________________ 16

2.4.2. Intelligence _________________________________________________________ 17

2.4.3. Flexible automation ___________________________________________________ 18

2.5. Industry 4.0 implementation barriers _______________________________________ 19 2.6. Interrelation between Lean and Industry 4.0 _________________________________ 20

2.6.1. Lean as an enabler towards I4.0 _________________________________________ 21

2.6.2. I4.0 as an enabler towards Lean _________________________________________ 22

2.7. Summary of background and literature review ________________________________ 24

3. Methodology ___________________________________________ 26

3.1. Research design ______________________________________________________ 26

3.1.1. Delphi survey method _________________________________________________ 26

3.2. Data collection ______________________________________________________ 27

3.2.1. Sample selection _____________________________________________________ 28

3.3. Operationalization ____________________________________________________ 29

iv

3.4. Data processing ______________________________________________________ 31 3.5. Validity ____________________________________________________________ 31 3.6. Reliability ___________________________________________________________ 33 3.7. Research ethics ______________________________________________________ 33

4. Results _______________________________________________ 35

4.1. First iteration survey __________________________________________________ 35

4.1.1. Connectivity _________________________________________________________ 36

4.1.2. Intelligence _________________________________________________________ 40

4.1.3. Flexible automation ___________________________________________________ 43

4.2. Second iteration survey ________________________________________________ 46

4.2.1. Success factors ______________________________________________________ 46

4.2.2. Purpose ____________________________________________________________ 47

4.2.3. Involvement of people ________________________________________________ 47

4.2.4. Competence ________________________________________________________ 48

4.2.5. Implementation strategy _______________________________________________ 49

5. Analysis and Discussion____________________________________ 51

5.1. Industry 4.0 Possibilities, Challenges and Risks ________________________________ 51

5.1.1. Connectivity Possibilities _______________________________________________ 51

5.1.2. Connectivity Challenges and Risks _______________________________________ 53

5.1.3. Intelligence Possibilities _______________________________________________ 54

5.1.4. Intelligence Challenges and Risks ________________________________________ 56

5.1.5. Flexible automation Possibilities _________________________________________ 57

5.1.6. Flexible automation Challenges and Risks _________________________________ 58

5.2. Discussion summary first iteration _________________________________________ 59 5.3. Success factors ______________________________________________________ 61

5.3.1. Purpose ____________________________________________________________ 61

5.3.2. Involvement of people ________________________________________________ 62

5.3.3. Competence ________________________________________________________ 62

5.3.4. Implementation strategy _______________________________________________ 63

5.4. Discussion summary second iteration ______________________________________ 65 5.5. Summary of discussion _________________________________________________ 68

v

6. Conclusions ____________________________________________ 70

6.1. Summary ___________________________________________________________ 70 6.2. Answers to research questions ___________________________________________ 71 6.3. Theoretical contribution ________________________________________________ 72 6.4. Practical contribution __________________________________________________ 73 6.5. Research limitations and improvement potential _______________________________ 73 6.6. Future research ______________________________________________________ 74

Bibliography_______________________________________________ 75

A. Appendix – Delphi survey 1st iteration _________________________ 82

B. Appendix – Piloting email __________________________________ 84

C. Appendix – Survey email template 1st iteration ___________________ 85

D. Appendix – Email template 2nd iteration ________________________ 86

E. Appendix – Executive summary for 2nd iteration _________________ 87

vi

List of Tables

Table 1: Barrier to organization’s CI capabilities as found in literature. ____________________ 13 Table 2: Definitions of key technologies enabling the “I4.0 Connectivity value driver”. _________ 17 Table 3: Definition of technologies enabling the “I4.0 Intelligence value driver”. ______________ 18 Table 4: Definition of technologies enabling the “I4.0 Flexible automation value driver”. ________ 19 Table 5: I4.0 value driver’s impact on CI as seen from current literature. __________________ 25 Table 6: Benefits and Shortcomings of the Delphi method. _____________________________ 27 Table 7: Data collection steps. _________________________________________________ 28 Table 8: Response rate details for the first iteration. _________________________________ 35 Table 9: Possibility with introducing Connectivity in the manufacturing environment. __________ 37 Table 10: Inertia tendencies derived from Challenges and Risks associated with Connectivity. ___ 39 Table 11: Possibilities with introducing Intelligence in the manufacturing environment. _________ 41 Table 12: Inertia tendencies derived from Challenges and Risks associated with Intelligence. ____ 42 Table 13: Possibilities with introducing Flexible automation in the manufacturing environment. ___ 44 Table 14: Inertia tendencies derived from Challenges and Risks associated with Flexible automation. _______________________________________________________________________ 45 Table 15 Response rate details for the second iteration. ______________________________ 46 Table 16: Success factors related to “Purpose”. _____________________________________ 47 Table 17: Success factors related to “Involvement of people”. __________________________ 48 Table 18: Success factors related to “Competence”. _________________________________ 48 Table 19: Success factors related to “Implementation strategy”. _________________________ 49 Table 20: Summary of themes of Possibilities, Challenges and Risks derived from the value drivers’ impact on the conditions for CI.________________________________________________ 60

vii

List of Figures

Figure 1: Different scenarios with CI only (A) and CI and I4.0 (B). ________________________ 3 Figure 2: The Toyota Way. ____________________________________________________ 8 Figure 3: The Toyota Production System. __________________________________________ 9 Figure 4: The four industrial revolutions (b.telligent, 2020). ____________________________ 14 Figure 5 Industry 4.0 framework (Saturno, et al., 2018). _______________________________ 15 Figure 6: The three value drivers within I4.0 (World Economic Forum, 2019). _______________ 16 Figure 7 Graphical illustration of Possibilities, Challenges and Risk in relation to a Baseline. _____ 30 Figure 8: Operationalization of the collected data. ___________________________________ 31 Figure 9: Number of years of experience of Lean. ___________________________________ 35 Figure 10: Number of years of experience of I4.0. ___________________________________ 35 Figure 11: Number of years of experience of participants per organization. _________________ 36 Figure 12: Summary of positive and negative input on the impact of Connectivity on CI. _______ 36 Figure 13: Summary of positive and negative input on the impact of Intelligence on CI. _________ 40 Figure 14: Summary of positive and negative input on the impact of Flexible automation on CI. __ 43 Figure 15: Summary of responses from the second iteration. ___________________________ 46 Figure 16: Example of Possibilities’ (green) and Risks’ (red) impact on the conditions for CI. ____ 61 Figure 17 Success factors as a means to ensure CI, realizing Possibilities whilst overcoming Challenges & mitigating Risks. _________________________________________________ 67 Figure 18 Summary for Connectivity. ____________________________________________ 68 Figure 19 Summary for Intelligence. _____________________________________________ 69 Figure 20 Summary for Flexible automation. _______________________________________ 69

viii

List of abbreviations

AA Advanced Analytics

AI Artificial Intelligence

AM Additive Manufacturing

AR Augmented Reality

B/C Benefit/Cost

BTH Blekinge Tekniska Högskola

CI Continuous Improvement

CPS Cyber Physical Systems

I4.0 Industry 4.0

IIoT Industrial Internet of Things

IoT Internet of Things

IT Information Technology

LTA Lost Time Accidents

LE Large Enterprises

ML Machine Learning

OT Operational Technologies

PDCA Plan, Do, Check, Act

RFID Radio Frequency Identification Device

SME Small and Medium sized Enterprises

SQDC Safety, Quality, Delivery, Cost

TPS Toyota Production System

VR Virtual Reality

3M Muri, Mura and Muda

1

1. Introduction

This research has been focused on the impact of implementing Industry 4.0 (I4.0) on the conditions for Continuous Improvement (CI) within the manufacturing environment. Since improvements are made by first identifying problems and subsequently solving them, the “conditions for CI” has been referred to as the factors and circumstances that impacts the “problem-sensitivity” and the “problem-solving” capabilities. The latter is built by behavioral aspects, resources, time and competence, while problem-sensitivity is built by Lean values and principles.

In this chapter, section 1.1 briefly introduces the readers to the concept of I4.0, Lean and CI. Section 1.2 covers the problem discussion and section 1.3 outlines the guiding research questions. In addition, section 1.4 and section 1.5 argue for the research’s theoretical and practical relevance and section 1.6, its delimitations. Lastly, the thesis’ structure is outlined in section 1.7.

1.1. Introduction to Continuous improvement and Industry 4.0 “Strive for continuous improvement, instead of perfection” – Kim Collins

Building on the quote by Kim Collins, a former track and field sprinter from Saint Kitts and Nevis, it is safe to assume that the concept of CI has found its way outside of the corporate arena. Needless to say, similar to an athlete, every organization aiming to stay competitive has an essential need for CI (Bicheno & Holweg, 2008). Moreover, as an organization’s long-run survivability is directly related to its competitiveness in relation to other players, CI is, per se, a necessity for all existing organizations subject to competition. One of the best known and most widely used methods to achieve CI is by using Lean which has proven to boost firms’ operational performance significantly (Shah & Ward, 2003). Simplified, Lean is a western attempt to emulate Toyota Production System (TPS) which has been successfully used by Toyota since the end of the Second World War (Fujimoto, 1999). CI is a well-known and established core value within “the Toyota Way” and a pre-requisite for TPS (Toyota Institute, 2001). In general terms, CI enhances an organizations’ products, services and processes though the incremental improvements of value generating activities (Bhuiyan & Baghel, 2005; Lillrank & Noriaki, 1989); which is done by continuously identifying and solving problems. This essentially implies that all aspects need to continuously improve. In an industrial manufacturing environment this would more specifically be things like Safety, Quality, Delivery and Cost (SQDC). In the context of this work however, the focus has been on CI in relation to cost reduction though productivity.

Driven by the current digitalization trend1, the manufacturing industry is at the brink of a new industrial revolution, denoted Industry 4.0 (I4.0) (Rojko, 2017). The term was coined in Germany at the Hannover Messe technology in fair in 2011 as an initiative by the German government aiming to implement a high-tech industrial technology strategy (Benitez, et al., 2019). In short, I4.0 refers to

1 Digitalization is the process of levering the conversion of information from analogous to digital format in order to improve business performance (Burkett, 2017).

2

different technology-driven changes in an organization’s manufacturing systems which have versatile organizational implications (Lasi, et al., 2014). The changes largely follow from the introduction of novel information and communication technologies into the manufacturing value-chain resulting in numerous benefits for the organization (Mohamad, 2018). Nevertheless, the true implications, especially those derived from integrating I4.0 technologies into the value-chain, remains dimmed by the single-sided discussion of the portrayed conceptual benefits of the I4.0 technologies. What is clear however, is that I4.0 will involve a radical shift in how shop floors are operated (Tjahjono, et al., 2017).

According to the World Economic Forum and McKinsey & Company, three technological megatrends are the prominent value drivers for the production transformation during I4.0; i.e., “Connectivity”, “Intelligence” and “Flexible automation”. Adapting technologies belonging to each megatrend can have profound effects on organizations that have escaped the inertia of “pilot purgatory2” such as increased productivity (World Economic Forum, 2019). Generally speaking, I4.0 brings about significant changes in both manufacturing and business (Nosalska, et al., 2019), but most companies are not aware of the lurking Challenges and Risks they might face when implementing I4.0 (Mohamad, 2018).

1.2. Problem discussion

As the industrial manufacturing landscape is transformed with the introduction of I4.0 technologies, the conditions for CI are likely to be affected. However, the impacts on organizations’ CI efforts following I4.0 implementation is yet to be investigated by researchers, despite the relevance and the potential interdependency of the two subjects. Moreover, as stated by several authors, there is a lack of comprehensive frameworks which combines I4.0 solutions with Lean (Kolberg & Zühlke, 2015; Leyh, et al., 2017; Wagner, et al., 2017). From the conclusions of the reviewed articles presented in Chapter 2, there seems to be an overall consensus that I4.0 and Lean are complementary rather than contradictory (section 2.6). Even so, it became apparent that the literature lacks a more realistic and critical approach to the interrelation between the two domains. Furthermore, although a few studies exist within the interdomain between I4.0 and Lean, none of them, to the best of the authors’ knowledge, discusses the possible impacts on CI. Considering that CI is a core value of Lean and that there is a necessity for CI everywhere within organizations subject to competition, this has been deemed a research gap worthy of fulfillment. A few studies were found to investigate barriers towards CI, mainly focusing on behavioral aspects of managers and employees as well as the organizational culture as key factors for successful CI (section 2.2). However, no literature was found that argues for the potential impact on those factors following I4.0 implementation, which further highlighted the need for discussion surrounding the interaction between CI and I4.0. As described above, realizing process improvements to improve productivity requires the ability to continuously improve and to adapt to the dynamics of the environment; a fact that emphasizes the

2 Pilot Purgatory: A term used to describe the difficulty to scale technology projects out of the piloting stage (Boer & Enno, 2018).

3

importance of a manageable organizational inertia3 (Kinnear & Roodt, 1998; Louw & Martins, 2004). Following radical changes however, such as when novel technologies are diffused into the manufacturing environment, there is perhaps a risk that the conditions for CI are affected through the changes in inertia tendencies. This can be caused by, e.g., negatively impacting behavioral aspects or Lean values and principles. Figure 1 attempts to illustrate the potential detrimental effects of inertia tendencies on the conditions for CI following a radical improvement as a result of implementing new technologies into the manufacturing environment. Figure 1 (A) illustrates a scenario with a rate of improvement using only small and risk-free improvements. Figure 1 (B), illustrates a scenario following a large direct improvement which, post realization, results in a lower improvement rate due to realization of various risks (as described below). As observed, scenario (B) has an overall reduced improvement level in comparison to scenario (A) over the same timeframe due to the overall increase in inertia tendencies (here illustrated as the improvement-time derivative). Note that Figure 1 is a rough simplification of reality were the rate of improvement is assumed to be linear; however, in reality, this is likely not the case.

Figure 1: Different scenarios with CI only (A) and CI and I4.0 (B).

The transition to I4.0 and the “smart factory” is likely to increase the importance of employees’ competence and qualification to perform more complex work tasks in an environment that is more connected, intelligent and automated (Glass, et al., 2018; Bonekamp & Sure, 2015). As industry encounter radical changes in the form of disruptive technology diffusions into the manufacturing environment following I4.0 adoption, there is a distinct risk that the time lag between the necessary increase in employee competences and the technology implementation results in a greater organizational inertia. If only temporarily, such inertia risks making necessary incremental process improvement more difficult to accomplish. Apart from the time-lag between continuous learning and technology diffusion, the literature in the I4.0 domain particularly outlines that; the lack of standards and best practices; lack of competences and “creating acceptance for change” are difficult challenges when adopting I4.0 (Schneider, 2018; Bonekamp & Sure, 2015; Glass, et al., 2018). However, the

3 Organizational inertia: Analogous to the definition of inertia found in physics, the phenomena of organisational inertia is described as the organisational resistance towards making transitions, i.e., an organisation’s overall inability to effectively react to change (Kinnear & Roodt, 1998).

4

literature in the I4.0-Lean domain does not include these factors and their potential impact on CI. Other areas that are overlooked include: cyber-security issues, decentralization of decision making; inaccuracy in data interpretation; and outsourcing of competence. Moreover, the introduction of digital waste in the form of over-processing of data and overburden of digital capabilities are often overlooked in the existing frameworks (Romero, et al., 2018). The identified absence of the above-mentioned topics in the discussion surrounding the present Lean-I4.0 integration frameworks potentially indicated an overall underestimation of the I4.0 inertia tendencies. Therefore, there has been a need to investigate the potential impact on the rate of improvement post I4.0 implementation.

1.3. Problem formulation, purpose and research questions

Following the above discussions, the purpose of the study has been to assess how the conditions for CI will be affected by the change in the manufacturing environment brought on by the implementation of I4.0. With a focus on the potential negative aspects, the authors have aimed to explore whether an increased “Connected”, “Intelligent” and “Flexibly automated” manufacturing environment can give rise to inertia tendencies and hence, negatively affect the conditions for CI. The study has been explorative which implies that the underlying aim has been to clarify and deepen the insight of the studied phenomena rather than to derive precise and conclusive results. Thus, the study has been guided by the following research questions:

R1: How does the introduction of I4.0 affect the conditions for continuous improvement with regards to the rate of improvement?

R2: What are the potential factors which can decrease the rate of improvement following an increased Connected, Intelligent and Flexibly automated manufacturing environment?

R3: What are the success factors preventing a decrease in the rate of improvement as a consequence of I4.0?

1.4. Theoretical relevance

Based on the above discussion, the primary focus has been to yield a theoretical contribution to the characterization of the I4.0 impact (both positive and negative) on the conditions for CI in the manufacturing environment. So far, the research concerning the I4.0 topic is overwhelmingly technology centered; thus, regarding management studies, I4.0 remains an under-studied topic (Piccarozzi, et al., 2018; Schneider, 2018). Moreover, I4.0 is the first of the revolutions to be called out a priori of its realization (Jabbour, et al., 2018), and the lack of consensus makes I4.0 an enigmatic term that encompasses several meanings and interpretations. Consequently, it becomes difficult to pinpoint and quantify I4.0 challenges that are not explicitly technology specific (Schneider, 2018). This partly explains why there are a scarcity of studies providing conceptual contributions of the I4.0 impacts on CI, while there are a surplus of articles investigating technology-related challenges. Although, an increasingly interest of the interconnection between Lean and I4.0 from both from industry and academia has given rise to a few studies on the topic (Mayr, et al., 2018), the

5

literature still lacks comprehensive investigations into the interaction between Lean and I4.0 (Pagliosa, et al., 2019). Hence, several authors have called for investigations into a comprehensive framework for Lean and I4.0 (Kolberg & Zühlke, 2015; Leyh, et al., 2017; Wagner, et al., 2017) as well as for investigations between I4.0 and CI (Buer, et al., 2018; Mayr, et al., 2018; Sanders, et al., 2016). Thus, by conceptually exploring the I4.0 impact on inertia tendencies and its constituting effect on CI though Lean values and principles; the authors have aimed to partly fill that research gap.

1.5. Practical relevance

The secondary focus has been to yield a practical contribution to the manufacturing industry by outlining and assessing the potential Challenges and Risks stemming from the I4.0 implementation. This has served to balance the somewhat single-side discussion by adding to the non-existing pool of studies that problematize around the potential negative consequences I4.0 adopters may encounter. Unfortunately, as the perceived benefits associated with I4.0 are highlighted and outlined, negative aspects in the face of I4.0 implementation are rarely discussed, even less problematized (Karadayi-Usta, 2019; Schneider, 2018). Thus, practical contribution has been made by contributing to a more realistic discussions prior to and post I4.0 implementation by outlining not only Possibilities, but also Challenges and Risks. As companies are likely to be blinded by the overly positive rhetoric originating from consultancy firms and other stakeholders that would benefit from selling I4.0 related products and services; such a contribution ought to be warmly welcomed by all current non-adopter companies. The importance of CI for the manufacturing industries cannot be understated, which is why a realistic exploration of the I4.0 impact need to be investigated. In addition to highlighting the potential Challenges and Risks, the focus has also been to identify success factors that can help manufacturers to successfully implement I4.0 without negatively impacting CI.

To put the cost benefit from CI in context, a company in the same size as Volvo CE might have a total operating cost of around 6000 MSEK. With an annual productivity target of 5%, the annual cost reduction targeted is 300 MSEK. If the implementation of I4.0 were to reduce the cost reduction achieved by CI by one percentage point, the impact would be 60 MSEK in lost cost reduction. The benefit of the implementation of I4.0 would then need to contribute with savings well in excess of 60 MSEK, not taking into account the cost of its implementation, to even break-even. In addition, the future CI rate of improvement might be negatively impacted causing further year over year losses of cost reduction. There is a risk that this lost potential is not included in the I4.0 business case since the future barriers (Challenges and Risks) to CI are not visible. One of many reasons might be that there is a lack of understanding between different groups of people in the plant working with CI and working with I4.0 implementation. A gap can hence be identified in trying to understand how this barrier can be avoided, or at a minimum, that awareness exists in the I4.0 implementation considerations resulting in a “smart implementation” of the smart factory.

1.6. Delimitations

The concepts of I4.0 has been limited to encompass three value drivers: Connectivity, Intelligence and Flexible automation. Each value driver has been defined through a set of key technologies to scope their definitions and to ease the impact assessment as described in section 2.4. CI is about the

6

many small, simple and cheap improvements in e.g., SQDC, which everyone is involved in, every day. However, for this project, the focus has been on CI relating to productivity; since that is a major focus for many companies and also relatively easy to measure and to track. Moreover, the impact on CI is assessed through the impact on “problem-sensitivity” and “problem-solving capabilities”. Problem-solving capabilities is built by behavioral aspects, resources, time and competence, while problem-sensitivity is built by Lean values and principles as discussed in section 2.1.3. The concept of “inertia tendencies” has not been further studied or outlined but has been merely used to commonly describe factors that can potentially negatively impact the conditions for CI, resulting in an unchanged or lower rate of improvement following an increasingly digitalized and technology-focused manufacturing environment. In this thesis, these factors are centered around Challenges and Risks as described in section 3.3. The justification for limiting the scope of the study to the manufacturing environment lies in the applicability of the I4.0 technologies. Thus, since I4.0 mainly focuses on the establishment of intelligent product and manufacturing processes (Brettel, et al., 2014), the manufacturing side of the value-chain is likely to experience the overall greatest environmental and organizational transformation. The study has been conducted by consulting academic experts as well as experts within various manufacturing companies within different sectors, ranging from the aerospace sector to the construction equipment sector. The companies are on various transformation scales on their I4.0 journey, ranging from “non-adopters” to “adopters” relative to each other. Lastly, as the topic of I4.0 is currently understudied and its implementation consequences are both unexplored and unclear, the adopted research focus has been exploratory. As a result, care should be exercised when trying to generalize the findings as no conclusive results can be drawn.

1.7. Structure of the thesis

In Chapter 2, Lean and I4.0 are explained in more detail along with a literature survey with regards to the connection between Lean and I4.0. In particularly, previous work done to understand the impact of I4.0 on CI are outlined and discussed. Chapter 3 outlines the research methodology and Chapter 4 outlines the results and analysis of the findings. In Chapter 5, the findings are discussed in detail and Chapter 6 provides conclusions as well as recommendations for future research.

7

2. Background and literature review

The following sections introduces the readers to Lean and the core value of CI, section 2.1, as well as CI implementation barriers identified by academia, section 2.2. Moreover, a thorough background of I4.0, section 2.3, and its constituting implementation barriers and value drivers; Connectivity, Intelligence and Flexible automation are given in section 2.4, as well as I4.0 implementation barriers, section 2.5. The literature for the background sections has been found using an unstructured approach, utilizing search engines in combination with BTH’s library database. Lastly, for the purpose of identifying knowledge gaps in the Lean-I4.0 domain and to create an initial basis for an implication matrix containing the I4.0 value drivers’ impact on CI, section 2.7; data has been collected through a comprehensive literature review outlining several articles belonging to the studied interdomain between Lean and I4.0, section 2.6. For this section, the literature has been found by utilizing a semi-structured approach, where the Scopus database have been scanned using Boolean logic, filtered by scientific material and publications. Additional articles have been found by “snowballing” literature reviews related to the subject in order to find relevant articles.

2.1. Background of Lean

Lean is the western world’s version of Toyota Production System (TPS). Lean as a term was first used by Womack and Jones in 1990 in their seminal book ”The machine that changed the world” (Womack, et al., 1990). Womack at al. (1990) showed the applicability of TPS outside Japan and its effectiveness in cultures other than the Japanese. This section is providing a brief introduction to TPS to better understand the origins of Lean and how CI fits into Lean as a Toyota Way core value.

2.1.1. The origins of Toyota Production System (TPS)

The history of TPS starts with the founding father of the Toyota industrial group, Sakichi Toyoda. Sakichi Toyoda was an inventor who, as a child, watched his mother using a manual weaving loom. To support the family, he started to improve the loom, amongst other things, and his most famous invention is the automatic loom and particularly the Type G model. The type G model featured autonomous automation which in Japanese is known as Jidoka. This invention meant that if a warp tread in the weaving loom broke the loom would stop automatically. The invention made it impossible to produce defect fabric. Equally important, it also meant that the shop floor operator did not have to watch over the loom to visually make sure all the treads were intact. So instead of assigning one operator per loom, a single operator could instead supervise several machines. The results were fabric with better quality and significantly increased productivity. Jidoka would later become the first main pillar of Toyota Production System (Fujimoto, 1999). The patents for the Type G automatic loom was sold in 1929 to the Pratt Brothers in the UK for one Million Pounds Sterling. This was the starting capital for Toyota Motor Corporation with the aim to produce cars. It was the son of Sakichi, Kichiiro Toyoda, who was tasked with setting up car manufacturing. Kiichiro had travelled to Europe and USA to study and benchmark automotive manufacturing as part of the preparations. In the initial days of Toyota and in the time post Second World War there was a scarcity of capital, raw material and parts to produce cars. This inspired Kichiiro Toyoda to establish

8

the pre-cursor of Just in Time (JIT). JIT would later become the second main pillar of Toyota Production System. JIT means that one should only produce the part, which is needed, at the time when it is needed and in the quantity that is needed. Jidoka and JIT are the two main principles in the Toyota Production System. These needs to be put it into a context and systematically applied to qualify as a production system. The main drive behind the Toyota Production System was Taiich Ohno. Ohno was initially a manufacturing engineer in the loom works and was transferred to Toyota Motor Company to support the improvement of the factory. Through benchmarking, Toyota understood that Japanese productivity was lagging behind the American automotive industry. It appeared that it required nine Japanese shop floor operators to accomplish the same amount of work as one American operator. Ohno set out to reduce this gap and during more than 40 years worked to create the Toyota Production System with support of the senior management (Ohno, 1988).

2.1.2. The Toyota Way

To explain TPS, there is a need to explain the Toyota Way which are the values on which TPS is based (Toyota Institute, 2001). These values enable the successful use of TPS.

Figure 2: The Toyota Way.

The Toyota Way consists of two components, Continuous Improvement and Respect for People. Continuous improvement consists of Challenge, Continuous Improvements (Kaizen) and Go and See and Respect for People consists of Respect and Teamwork (Toyota Institute, 2001).

Challenge means that current status should always be challenged as to drive improvements. Questions such as: Why are we getting the results we are getting? Why is the current manufacturing process looking the way it is? Why are the job elements being done in this sequence? Should be asked in order to continuously challenge the status quo as to drive improvements. Continuous Improvement (CI) means that one should always try to improve. “Always a better way” is a frequently used slogan within Toyota. When talking about CI the Japanese word Kaizen is used. Kaizen means change (kai) good (zen) and is used interchangeably with CI. When teaching CI, it is emphasized to focus on the many small improvements rather than the few large improvements. The reason is that the many small improvements are easy to find and to come up with, they are also normally quite simple and relatively cheap with a low implementation risk. Large improvements, however, might take a longer time to implement during which time there are instead no improvements. A large improvement is potentially also more complicated, more expensive and hence entails a higher risk. If the large improvement does not work as intended or do not give the results

9

anticipated, then the step back is also larger than if a small improvement does not work. Toyota employees all have four different job tasks: 1) do their daily tasks, 2) improve their daily tasks, 3) improve themselves and 4) improve their team. This only works if there is an expectation from leadership that everyone should constantly be involved in CI (Suzaki, 1987). Go and See. The concept behind Go and See is that management should be present where the value adding process is being undertaken. In production this means that managers are not hiding in their offices but rather are out on the shop floor to fully try to understand the current conditions and to try and provide support in terms of problem-solving and challenging the current best level as to drive CI. Teamwork is the emphasis on working as a team. The background, knowledge and different perspectives from each shop floor operator is used to ensure that the contribution of the team is greater than the sum of its individual members. This applies to all functions and should drive strong cross-functional collaboration. Respect is probably the value of Toyota that is the most misunderstood when organizations try to implement Lean. Respect means, of course, to treat each other respectfully. Respect in Toyota also means giving people challenging jobs and tasks as a sign of respect of their capability. If someone’s performance does not meet the expectations then, as a sign of respect, feedback on this is provided. It would be considered disrespectful to not give feedback on poor performance because it would hide an improvement opportunity for the person. Therefore, respect is far more than being nice and is probably one of the more difficult aspects of Lean to apply and also the part where the connections with Japanese culture and society is the strongest. Soft topics like this is inherently more difficult to implement compared to things like hard Lean tools.

Successful Lean implementation requires building a culture that reflects the Toyota Way values. Without them, the right Lean thinking and culture will not flourish and instead there will be a focus on specific Lean tools which in time will be difficult to sustain (Suzaki, 1987).

2.1.3. Toyota Production System

As seen in Figure 3, the Toyota Production System (TPS) consists of three main principles: Standardized Work, JIT and Jidoka.

Figure 3: The Toyota Production System.

10

2.1.3.1. Problem-sensitivity

Before discussing the TPS principles in detail, the reader needs to understand in which context TPS was developed, i.e., limited availability of material and capital as a consequence of the Second World War and the need to close the productivity gap towards US car production (a gap estimated by Toyota to be nine to one). This means that the underlying purpose of TPS is cost reduction. The mechanism to achieve this cost reduction is “problem-sensitivity”. At Toyota, they want to be sensitive to problems. This sounds counter-intuitive; as one might think that the opposite is wanted (i.e. not being sensitive to problems) so that even if problems occur, the production can continue un-disturbed. However, the thinking at Toyota is that: if you are sensitive to problems, then there will be an issue when a problem occurs. Thus, it is the issue itself which allows us to see the problem. If we can see our problems, then we can do something about them and if we solve them, we will gradually become better; or in other words “continuously improve”. When Toyota is talking about conducting CI, it is not just because they know it is good, but rather that they are operating within a system where problem-solving and CI are the only possibility. Thus, the role of management is to continuously make manufacturing more and more sensitive to problems as to drive more and more CI (Modig & Åhlström, 2015). When applying Toyota way and TPS it is important to understand the core values of Toyota way as well as the principles, methods and tools of TPS and their inter-connection. Visiting Toyota factories and when reading about Lean, it is easy to grasp the application of the different Lean tools. There is then a need to appropriate this tool to the specific organization. It might be counter-productive to apply a Lean tool from mass-production of motor vehicles to a specialist health care situation. Instead it is important to understand the Values, Principles and Methods which Toyota tried to achieve as to make it work efficiently outside Toyota. TPS is not a toolbox which can just be implemented, instead the only purpose is to increase the problem-sensitivity and then the use of the Toyota way core values means we will drive the business forward by continuously improve (Ohno, 2013). It is clear that leadership has an important role in any Lean transformation. Leadership must facilitate “the right way of thinking”, commitment as well as persistency. Failure to establish a Lean culture (based on the correct core values) and failure to appropriate tools and methods are the main reasons to why Lean initiatives fails.

2.1.3.2. Standardized Work

Standardized Work should be centered around human motion and describes the safest way to produce with correct quality and quantity. Standardized Work describe the currently best-known way to conduct work. There is an expression in Toyota: “without standard there can be no kaizen”. The reason for this is that if improvements that is made in a process are not shared and documented with all shop floor operators, then we cannot ensure that they all do the same work and not slide back to their own personal but less efficient way. So, if there are no standards then there is no point in making improvements as it will anyway deteriorate. The first step in any improvement cycle

11

should be to create a clear and detailed standard which can be used as a baseline on which improvements can build (Womack & Jones, 1996). Standardized Work is sometimes criticized to take initiative and creativity away from the operators and turning them into “robots”. The standard should be made by the shop floor operator in collaboration with the supervisor. Experienced operators do not need the standard for each cycle as they the work well, but the supervisor will use the standard when making job observations to confirm that correct sequence job element times. If you watch someone doing a process, then it is difficult to understand if the operator is doing the job correctly and in correct time. If you have a standard in your hand you can clearly see if the correct steps are followed. If the sequence is not followed or if there is any gap in the timing that indicates that there is a problem somewhere, the standard creates problem-sensitivity, which is the underlying mechanism in TPS.

2.1.3.3. Jidoka

Jidoka is also known as “Built-in-Quality”. It was the first pillar of TPS and was one of the inventions from Sakichi Toyoda. Jidoka is Japanese and means automation with a human touch. The basic principle is that when a problem occurs the machine automatically stops and signal that there is a problem. The consequence is that no defect parts will be produced, this eliminates scrap cost and reduce the risk of mixing good and bad parts. Additional benefit is that we do not need someone to supervise each machine and hence productivity is improved (Fujimoto, 1999).

2.1.3.4. Just in Time (JIT)

JIT is one of the principles along with Jidoka, it was introduced by Kichiiro Toyoda for pragmatic reasons since they could not afford large amount of capital tied up in inventory. JIT focuses on three items: Continuous flow, Pull System and Takt Time (Ohno, 1988).

Continuous flow is relating to keeping the flow of products or services flowing in a steady state. By keeping a flow, stagnation is avoided, and stagnation will result in excess inventory which increases lead time. There are several basic means to keep continuous flow such as streamlining, multi-process handling and one-piece production. Pull system is connecting different production processes together, to only produce what is needed by the downstream process. Kanban has been successfully used to achieve this. Takt Time (TT) is translating customer demand into production demand. It is determined by how many parts the customer requires and available production process time. As an example, the customer needs 20 parts per day and production runs 400 minutes per day meaning TT is 20 minutes.

The level of JIT is measured through lead time which can be broken down in processing time and stagnation time. To reduce the lead-time, the waste which is creating the stagnation time should be addressed. With shorter lead times it also becomes easier to see problems. This will then enforce problem-solving in order to make sure that the problem will not recur and affect the lead time (Japan Management Association, 1986).

12

2.1.4. Waste elimination and Muri, Mura and Muda (3M)

When applying the Toyota Way values and TPS, the outcome will be that one starts to see the problems better which increases the visibility of the waste (Fujimoto, 1999). When talking about waste (or Muda in Japanese) it is important to first mention Overburden (Muri) and Unevenness (Mura).

Overburden (Muri) is the first priority for improvement. To overburden a shop floor operator is poor respect and from a practical perspective there is an increased risk for injury and the work is not sustainable over time. Unevenness (Mura) is the second priority for improvements. Unevenness can be physical, but it can also be in time where workload is high in the morning and low in the afternoon. Waste (Muda) is the third priority for improvement when Muri and Mura is reduced or eliminated. There are seven wastes with an eight-waste added in recent times relating to unused creativity of the shop floor operators:

1 Over-production 5 Over-processing 2 Inventory 6 Waiting 3 Defects and rework 7 Transport 4 Motion

2.2. CI implementation barriers

In this section articles found to specifically address barriers to CI, have been covered. A brief summary of the articles and their findings have been followed by a more general discussion and conclusion at the end of this section.

According to Bessant & Caffyn (1997), CI in all aspects of the business is crucial for an organization’s ability to meet challenges brought about a turbulent and dynamic environment. The authors also argue that CI is fundamentally about a behavioral process that involves both learning and unlearning (Bessant & Caffyn, 1997). This essentially highlights the importance of adaptability and to embrace continuous learning in order facilitate CI. Hence, any factors that impede the process of acquiring and learning the key set of behaviors needed to sustain CI can be seen as a barrier to an organizations’ CI-capability. During 1992-1997, the Continuous Improvement Research for Competitive (CIRCA) team at the University of Brighton developed a CI-capability model, describing CI from a set of key behaviors and routines as pre-requisites for long time success. For a complete outline of those set of behaviors the reader is referred to the study by Bessant & Caffyn (1997).

McLean & Antony (2014) investigated CI activities in the manufacturing industry and the reasons for their failures. Through a structured literature review, the authors found several factors as to why the activities fail and grouped them into eight themes (motives & expectation; organizational culture & environment; the management leadership, implementation approach; training; project management; employee involvement level; and feedback & results). Derived from the presented themes, McLean et al. (2015) continued the study and implemented a step-model to CI and grouped the themes into the different levels. Under “the management leadership theme”, several other authors conclude that the lack of managerial support and inadequate supervision constitutes substantial barriers to CI (Garcia-

13

Sabater & Marin-Garcia, 2011; Lodgaard, et al., 2016). Moreover; limited employee knowledge and motivation (Lodgaard, et al., 2016); difficulty in breaking traditional mindset and facilitating the right behaviors (Bhuiyan & Baghel, 2005; Caffyn, 1999); and lack of supporting organizational culture (Caffyn, 1999; McLean, et al., 2015; McLean & Antony, 2014) were also identified as potential barriers to organizations CI-capabilities.

A summary of CI-barriers as identified by several authors are given in Table 1. For the purpose of simplifying and structuring the review, the barriers have been divided into three categories, i.e., managerial factors (relating to the managers’ behaviors), employee factors (relating to the employees’ behaviors) and environmental factors (relating to organizational culture, methods and technologies). As observed from the studied articles, behavioral aspects are of critical importance to sustain CI, both in terms of managerial and employee factors. Arguably however, there is a lack of research focusing on the environmental factors and its potential impact on the conditions for CI. No articles were found to particularly outline environmental factors coupled to e.g., technological complexity as a potential barrier to CI. Considering that rapid changes in environmental factors (such as with the introduction of I4.0) may reinforce the importance of both managerial behavior and employee and factors; the topic ought to be extremely relevant. For example, as new manufacturing technologies are introduced, employee knowledge may be limited in terms of how to conduct improvements of the newly introduced systems, which would arguably have negative impacts on the conditions for CI. The categories outlined in Table 1 are interlinked as both managerial factors and employee factors are affected by a change in environmental factors and vice versa.

Table 1: Barrier to organization’s CI capabilities as found in literature.

Managerial factors Employee factors Environmental factors (Lodgaard, et al., 2016)

Lack of support and supervision from managers.

Limited knowledge and motivation for performing necessary activities.

(Garcia-Sabater & Marin-Garcia, 2011)

Managers need to be involved and continuously follow-up the process of CI. Preferably, there should be specially appointed CI-managers.

(Bhuiyan & Baghel, 2005)

Difficulty in breaking traditional mindsets and adapt to new ones.

(Caffyn, 1999) Behaviors of individuals and groups that are misaligned with those found in the CI Capability Model, developed by the CIRCA team.

Unlearning detrimental behaviors and make facilitating behaviors routines.

Lack of supporting culture, facilitating the correct attitude and behaviors.

(McLean, et al., 2015)

Lack of management support and commitment Lack of feedback and results

Pursue CI with the wrong motives and expectations Inadequate training to perform CI. Involvement level (time allocation & role conflict).

Organizational culture impeding the willingness and ability to change.

14

2.3. Background of Industry 4.0

Since the dawn of the industrial era, technological leaps have led to paradigm shifts within the manufacturing industry. These so-called shifts, or post-named “industrial revolutions”, are generally a result of social, economic and political changes (Lasi, et al., 2014). The first industrial revolution began in late 18th century triggered by the mechanization of the manufacturing systems using water and stream power; and the second industrial revolution began in late 19th century and was characterized by mass production through the utilization of electric energy (Rojko, 2017). Moreover, following the introduction of automation through computerization and the introduction of information technologies in the 20th century, the third industrial revolution emerged. (Xu, et al., 2018). Common for the various revolutions are the significant changes they imposed to the manufacturing environment. In the aftermath of the revolutions, it is safe to say that the changes have led to significant improvements in productivity and quality. Building on the third industrial revolution, the fourth one is now emerging. However, unlike the previous revolutions, the fourth is evolving at an exponential rather than a linear phase (Schwab, 2016). Figure 4 illustrate the previous industrial revolutions as to put the fourth one into context.

Figure 4: The four industrial revolutions (b.telligent, 2020).

A quick search reveals that the fourth Industrial revolution is often referred to as Industry 4.0, abbreviated as I4.0, especially in context of manufacturing. Concepts surrounding I4.0 was introduced in Germany 2011 and portrays an industrial age characterized by a high degree of data exchange, autonomy and automation (Agostini & Filippini, 2019). Such an industrial configuration can only be the results from a purposefully formulated strategy realized over time and can potentially enable an increased competitiveness in the form of increased business robustness and flexibility (Piccarozzi, et al., 2018). Presently, I4.0 is no longer just a future trend, but rather a strategical goal that manufacturing companies are actively trying to realize (Xu, et al., 2018). Although its definition is currently being debated, industry and academia have started to provide conceptual I4.0 framework definitions in combination with investigations into its many benefits. Common for the various definitions of I4.0 is the portrayal of a “future factory” that utilizes state-of-the-art technologies (Jabbour, et al., 2018). Arguably, companies which do not strive towards I4.0 can hardly call themselves innovative by today’s standards. Moreover, there is today no definition of a

15

particular degree of implementation constituting a final I4.0 stage. Instead, from the perspective of each individual player, there is merely a higher or lower degree of adoption of I4.0 technologies relative to other competitors. Figure 5 illustrates a typical I4.0 technology centered framework as defined by Saturno et al. (2018); encompassing technologies and notions that often finds their way into the concept of I4.0.

Figure 5 Industry 4.0 framework (Saturno, et al., 2018).

At first glance, one might think that I4.0 decreases the need for human operators within industry following the increased automation and autonomy of the production line. However, Schneider (2017) discusses the human-centered perceptive of I4.0, derived from the increased importance of the human-machine interaction in an environment surrounded by intelligent collaborative machines and cyber-physical assistance systems. Such a perspective bypasses the “dark-factory” perspective characterized by little to no human intervention in the production chain by focusing on the importance of the challenges related to cooperation and networking; not only between humans, but between humans and machines. From the studied articles within the I4.0 domain it becomes abundantly clear that I4.0 brings about several well-defined benefits such as: increased productivity, customizability, quality as well as the reduction of overproduction and waste (Mohamad, 2018). To the contrary, the Challenges and Risks are somewhat diffuse, partly due to novelty of the I4.0 topic.

2.4. Industry 4.0 technology value drivers

According the Word Economic Forum and McKinsey & Company, I4.0 front-runner production sites have embraced three technology value drivers that are the principal drivers of the I4.0 transformation in manufacturing; i.e., Connectivity, Intelligence and Flexible automation (World Economic Forum, 2019). The value drivers include: “enabling relevant information to the right decision maker in real time”, “generate new insight and enable improved decision making through advanced analysis and artificial intelligence” and “utilizing new robotic technologies to increase safety, quality and productivity” (Goering, et al., 2018; World Economic Forum, 2019). The three value drivers are closely interlinked and work in conjunction and therefore serve as both enablers and disablers for each other. Figure 6 shows a summary of the three value drivers in accordance to the definition made by the World Economic Forum (2019).

16

Figure 6: The three value drivers within I4.0 (World Economic Forum, 2019).

2.4.1. Connectivity

A definition of I4.0 from a “Connectivity perspective” has been introduced by Varela et al. (2019) “The principles of I4.0 are the horizontal and vertical integration of production systems driven by real-time data interchange and flexible manufacturing to enable customized production” (Varela, et al., 2019, p. 2).

Horizontal boundaries of the firm are identified as the range of quantities and varieties of products and services a company produces; additionally, the vertical boundaries are defined as all activities found throughout its value-chain (Besanko, et al., 2017, p. 55 & 90). Thus, following the definitions of horizontal and vertical integration as outlined above, I4.0 implies increased Connectivity and cooperation across all dimension of the company, not necessarily limited to the production environment. Nevertheless, an increased Connectivity across the production borders, from supplier to customer, will inevitably alter the manufacturing and shop-flor conditions by increasing visibility and the flow of information. The Connectivity domain also concerns the increased Connectivity between the virtual space and the physical world through cyber physical system (CPS), an integrated system of computing, communications and control (Zhou, et al., 2015). In general terms, a cyber-physical environment is created through the combination of Information technologies (IT) and Operational technologies (OT) (AMFG, 2019). In such an environment, physical process can be controlled and monitored with the use of computer algorithms in a constant feed-back loop where the physical process affects the computations and vice versa.

The ability to connect digital devices and shop floor monitors with IT-systems through the Internet of Things (IoT) can enable relevant information to decision makers in the right time (Goering, et al., 2018). IoT also enables the tracking of objects through the production line in real-time by integrating Radio Frequency Identification Devices (RFID) and other sensors with the internet (Xu, et al., 2018). In short, RFID can identify, and track tags attached to objects with the help of radio waves. When connected to the internet, the RFID-reader can send the identification and positional information to wherever it is needed. As sensor technologies enables computers to observe and identify their environment without the need for human intervention (Achton, 2009), the sensor to internet connection enables the sensor-collected information to flow to other units in the factory. Thus, through IoT, a continuous flow of information from systems to actors, machines-to-machines as well

17

from devices to components can be established (Rojko, 2017). Applying IoT in an industrial context is often referred to as the Industrial Internet of things (IIoT) or the “Industrial Internet”. Other technologies that fall under the authors definition of Connectivity are Augmented Reality (AR) and Virtual Reality (VR). AR uses the integration of computer-generated information into the real word environment, often through the projection of information into the user’s field of vision that interacts with the surrounding. In the context of I4.0, AR can be implemented as assistant systems, creating a visual interface that can provide system information to the user in a real-time context. AR can potentially guide a user with unfamiliar tasks by visualization of information directly in the user’s spatial vision, aiding e.g., the assembly of new products (Volker, 2014). VR however, allows for the visualization of objects through semi or fully-immersed devices such as binocular headsets or “cave” VR. It opens a virtual interactable environment that can contain, e.g., visual information regarding operational conditions and virtual models of entire products, machines and factories (Kovar, et al., 2016). Common for both VR and AR, are their potential to contribute to increased visibility and connection between the virtual and physical world.

Table 2: Definitions of key technologies enabling the “I4.0 Connectivity value driver”. Technology Definition Augmented Reality (AR) Refers to the integration of computer-generated information into the real-world

environment. Most current AR-applications integrate computer graphics into the user’s view of the surroundings (Volker, 2014)

Cyber-Physical System (CPS) A term used to denote a system of Integrated computations and physical processes. Utilizing embedded computers and network monitors the physical process is controlled through feedback loops between the physical process itself and the computer calculations (Lee, 2008). Essentially an integration of IT and OT in various combinations (AMFG, 2019).

Internet of Things (IoT) Industrial Internet of Things (IIoT) Industrial Internet

The definition encircles the concept of a network of devices embedded with software and information sensing devices such as laser scanners, position systems, infrared sensors and Radio Frequency Identification Devices (RFID) which are connected to the internet (Zhou, et al., 2015). The integration of such sensors with the internet enables real-time tracking and identification of tags that can be attached to any object (Xu, et al., 2018).

Information Technologies (IT) “Technologies that store, process and transport information” (ETCIO, 2016)

Operational technologies (OT) “Hardware and software that detects or causes a change through the direct monitoring and/or control of physical devices, processes and events…” (Gartner, 2020)

Virtual Reality (VR) Allows for visualization of virtual objects. Full immersive VR-devices provide 3-dimensional virtual projections accessible to the user (Kovar, et al., 2016)

2.4.2. Intelligence

As machines start to communicate and sense their surroundings through IoT and CPS, they have the potential to become autonomous. However, to usefully process, learn from and subsequently act on the vast amount of sensor-collected data; Intelligence is needed. With the introduction of Artificial intelligence (AI), which refers to the simulation of human-like intelligence in machines, human “thinking” can be mimicked (Frankenfield, 2020). When discussing AI, the term “Machine Learning” (ML) is unavoidable. ML refers to the science of getting machines to learn and improve by feeding them data from observations and interactions in order to mimic human behavior (Faggella, 2020). In essence, ML can be viewed as an enabler to AI (McClelland, 2017). Hence, through ML algorithms

18

combined with real-time data capturing, machines can make their own decisions (Rojko, 2017). As AI has the potential to make sense of data and act thereafter, it can enable predictive maintenance, simulation-based forecasting and other optimization benefits. However, the full potential of AI in the production environment is yet to be captured (Goering, et al., 2018).

Advanced Analytics (AA) is an umbrella term for several methods used in predictive forecasting (AI included). In essence, AA is known as predictive analytics and can be used in decision making by predicting various outcomes (Bose, 2009). The AA umbrella also covers terms such as “big data” and “data mining” as well as various statistical and semantic analysis techniques. Data mining intersects the science of machine leaning and refers to the process of finding usable correlations, patterns and anomalies in data sets. Moreover, using AA against large data sets that are both structured, semi-structured and unstructured is referred to as big data analytics. In general, “big data” is a term that describes data sets that cannot be captured, managed and stored using traditional table-structured databases. Furthermore, big data also inherit characteristics such as high volume, variety and velocity in data (IBM, 2020).

Table 3: Definition of technologies enabling the “I4.0 Intelligence value driver”. Technology Definition Advanced Analytics (AA)

An umbrella term for different advanced analytics techniques, or tools, that can be used in combination with each other to gain insights by analyzing information as well as performing predictive analysis (Bose, 2009).

Artificial Intelligence (AI) Refers to the simulation of human-like intelligence in machines that are programmed to mimic human behavior (Frankenfield, 2020).

Machine Learning Machine leaning is a subpart of AI and can be described as “…the science of getting computers to learn and act like humans do, and improve their learning over time in autonomous fashion, by feeding them data and information in the form of observations and real-world interactions.” (Faggella, 2020).

2.4.3. Flexible automation

According to Miller (2017) three distinct evolutionary phases can be observed: “fixed automation”, “programmable automation” and “flexible automation”. The fixed automation is designed to produce a particular product configuration repetitively and efficiently. The next level of automation is called programmable automation, or manual changeover automation, and implies that there is room for configurability after implementation through reprogramming. However, the process of manual configuration is labor intensive and often results in significant down time (Miller, 2017). Thus, following the programmable automation is Flexible automation, which, according to the World Economic Forum is a principal value driver of I4.0 (World Economic Forum, 2019). Flexible automation, also called “soft” automation, offers a quick and seamless conversion of the process or machine, which allows manufacturers to produce a variety of products using e.g., a single adaptive machine (Miller, 2017). Thus, being “flexible” in a manufacturing context essentially implies having the capability of adjusting to different configurability’s within a relatively short time frame.

Subtracting methods, such as machining, as well as additive methods, such as 3D-printing or Additive Manufacturing (AM), are manufacturing methods both capable of Flexible automation. However, AM is the most novel of the two. It has the capabilities to decrease waste, enable customized production and reduced inventory. It is a method where material is added in a layer-by-layer fashion, rather than

19

subtracted from a predefined form as in the case of subtracting methods (Zhang & Jung, 2018, p. 2). This is generally done by melting or fusing the input material with the substrate material utilizing a high energy laser, layer-by-layer, after first digitally slicing a 3D-model to retrieve path-coordinates for the 3D-printer. The notion of Flexible automation also includes automated vehicles, advanced robotics and collaborative robots, also known as “cobots” (Goering, et al., 2018). There are four collaborative features for cobots: safety monitored stop, hand guiding, speed and separation monitoring and power and force limiting (Bélanger-Barrette, 2015). This implies that cobots can sense their environment, can work safely in the proximity of humans, stop automatically and slow down when necessary, as well as learn by e.g., path teaching while minimizing the impact of human mistakes.

In relation to programmable automated industrial machines, autonomous robotics, also known as “advanced robotics”, have the potential to make their own decisions and act accordingly. Advanced robotics have superior perception (through computer vision and sensors), integrability (through plug-and-play functionalities), adaptability (through advanced data-processing and cloud-services) and mobility (autonomous guiding) in comparison to conventional robots. Moreover, they also have decentralized Intelligence that allow them to make their own decisions which provides a basis for self-controlled manufacturing (Küpper, et al., 2019). Arguably, the integration of Intelligence in machines, as described in section 2.4.2, are a prerequisite for enabling Flexible automation.

Table 4: Definition of technologies enabling the “I4.0 Flexible automation value driver”.

Technology Description Additive Manufacturing (AM) Additive Manufacturing is also referred to as 3D-printing: “3D printing is a production

process by which an object is produced in an additive fashion, layer-by-layer.” (Zhang & Jung, 2018, p. 2)

Advanced Robotics Autonomous Robotics

Robotics with intelligent functions that can autonomously react and act on information. Advanced robotics inherit increased perception, integrability, adaptability and mobility in comparison to conventional robotics. (Küpper, et al., 2019)

Collaborative robots Cobots

A robot that works in collaboration with humans to produce or create something. According to ISO 10218 part 1 and part 2, there are four types of collaborative features for cobots: safety monitored stop, hand guiding, speed and separation monitoring and power and force limiting. (Bélanger-Barrette, 2015)

2.5. Industry 4.0 implementation barriers

Several authors have been found to problematize around I4.0 challenges on the road to I4.0 adoption (Glass, et al., 2018; Schneider, 2018; Machado, et al., 2019; Mohamad, 2018; Karadayi-Usta, 2019; Zhou, et al., 2015). Nevertheless, apart from the mentioned studies, the literature is scarce when considering conceptual contributions studying the Challenges and Risks following I4.0 technology diffusion.

According to Bonekamp & Sure (2015), the transformation of the shop floor will lead to a decrease of low-skilled Standardized Work and in an increase in high-skilled work activities. Thus, the importance of education, continuous learning and training to adapt to the new environment will be essential (Bonekamp & Sure, 2015). A perspective that is also supported by Glass et al. (2018) who, through a literature review and statistical survey, studied manufacturers’ need for external support when adopting I4.0. The authors also found that missing standards constituted major problems for both Large Enterprises (LEs) and Small to Middle-sized Enterprises (SMEs). In fact, “missing international

20

norms”, “lack of compatibility of ports and interfaces” and “missing standardized data formats” were together with “missing skilled operators and know-how” the highest rated challenges for LEs and SMEs (Glass, et al., 2018). Following on the same line of arguments is Schneider (2018), who conducted a comprehensive literature review outlining the managerial challenges towards I4.0 adoption. The author also outline another highly rated challenge towards the utilization of I4.0, i.e., “creating acceptance for change” (Schneider, 2018). Moreover, the current research surrounding I4.0 arguably takes the viewpoint that I4.0 is a desired state for any industrial company. Although such a viewpoint is agreeable, it causes an overall blindsided perspective, forgetting of the potential Challenges and Risks facing companies after I4.0 implementation.

2.6. Interrelation between Lean and Industry 4.0

Following the perspective of Dombrowski et al. (2017), the interrelation between Lean and I4.0 can be divided into two perspectives, i.e., “Lean as an enabler towards I4.0” and “I4.0 as advancing Lean production systems” (Dombrowski, et al., 2017). The two perspectives are supported by the findings from a structured literature review by Buer et al. (2018). However, Buer et al (2018) also added a performance dimension and an environmental dimension to the cluster of identified literature. Although Buer et al. (2018) did not find a single article belonging to the latter dimension, it was assumed that the environmental factors, e.g., if the manufacturing environment is repetitive or non-repetitive, affects the potential to integrate Lean and I4.0 (Buer, et al., 2018). Under the performance dimension, the authors also found a total of ten articles empirically outlining positive cost, flexibility, productivity, reliability and reduced inventory benefits following such I4.0 implementation. However, neither Dombrowski et al. (2017) nor Buer et al (2018), discuss any potential downsides on Lean princples following I4.0 implementation.

Mayr et al. (2018) conducted a structured literature review on the interrelation between Lean and I4.0 and found no articles outlining negative performance correlations, but instead found two articles specifically outlining positive performance correlations. In the first one, by Sanders et al. (2016), it was investigated whether I4.0 could function as a catalyst for Lean. The article concluded that integrated communication and information systems aid Lean processes by reducing waste and increaseing productivity. Problems associated with JIT deliveries by suppliers such as incomplete goods’ shipping status and mismatch in quantity of transported goods and unexpected delays during transportation, were found to be potentially solved by IoT through item-tagging and wireless tracking. Moreover, decentralized decision making and real-time inventory tracking using RFID were argued to enhance JIT pull-production and continuous flow (Sanders, et al., 2016). However, beside the positive correlations of I4.0 on Lean, Sanders et al. (2016) also stated that: “there is a demand for further research to emphasize the importance on continuous improvement over the dimensions of Lean manufacturing”. In the second article concerning positive performance correlations between Lean and I4.0, Mrugalska & Wyrwicka (2017) argued that CI can be enhanced by smart products by tracing product-related data over the whole lifecycle and contribute to effective value stream mapping. Moreover, Mrugalska & Wyrwicka (2017) argued that smart machines and augmented operators can decreasing the time between failure occurrence and its notification, increased traceability and visibility and therefore aid the process of CI (Mrugalska & Wyrwicka, 2017). However, as the authors

21

also purposefully investigates I4.0 as a way of overcoming the implementation barriers of Lean, there was an overall lack of problematization surrounding the potential discrepancies between the two.

As industry starts to diffuse I4.0 technologies it is likely to meet resistance in various form. Such resistance constitutes barriers that poses various challenges for management. As decision makers, managers need to implement practices supportive of the I4.0 environment. According to Agostini & Filippini (2019), CI practices is of high importance for the successful implementation of I4.0 technologies. Using a mixed research method, the authors found three “clusters” related to the degree of I4.0 technology implementation for SMEs in Italy (adopters, beginners and non-adopters). Moreover, the article outlines statistically significant correlations between companies belonging to each of the cluster and the managerial and organizational practices present within the companies. As was concluded, the adopters share a higher degree of management and organizational practices surrounding areas such Lean on a process and supply chain level. The study further suggested that firms that implement I4.0 technologies more intensely are also those that share a high degree of Lean practices and Lean tool-utilization (Agostini & Filippini, 2019). Whether the observed increased managerial and organizational practices for adopters were a result of increased I4.0 technology implementation, or vice versa, was not discussed.

Utilizing an empirical methodology, Lorenz et al. (2019) investigated the relationship between Lean and digitalization. Through survey data from Swiss manufacturing firms triangulated with consecutive interviews, the authors found a positive correlation between digitalization maturity and the Lean maturity of firms. Following the two perspectives (as outlined by (Dombrowski, et al., 2017)), Lorenz et al. (2019) argued that digitalization and Lean can facilitate each other by; e.g, decreased shop floor complexity utilizing digital shop floor management; increased flow enhanced by digital planning techniques; increased transparency utilizing data mining and increased quality by applying AA. In the opposite perspective, Lean was argued to support digitalization by streamlining the process and ease the data-collection phase. Moreover, Lean was also argued to facilitate the diffusion of technologies that are supportive of Lean thinking, i.e., customer value and waste elimination (Lorenz, et al., 2019). The latter perspective, emphasis the importance of having established Lean practices before considering I4.0 adoption.

2.6.1. Lean as an enabler towards I4.0

The articles outlined in this domain are founded on a quote by Bill Gates: “The first rule of any technology used in a business is that automation applied to an efficient operation will magnify the efficiency. The second is that automation applied to an inefficient operation will magnify the inefficiency” (Buer, et al., 2018, p. 2934). A quote clearly illustrating the importance of Lean following increased digitalization and automation.

Leyh et al. (2017) conducted a literature study where 31 papers containing holistic I4.0 frameworks were analyzed with a Lean perspective. Their finding was that “it became obvious that the Lean principles were not often addressed in I4.0 models”. They also concluded that despite the fact that Lean is often viewed as a basis for I4.0 implementation, they were not integrated into the frameworks. The concluding remark was that there still are issues that can be viewed as unsolved or at least not adequately addressed. Therefore, further research was concluded to be necessary in

22

order to combine existing approaches (Lean) with additional key aspects of Industry 4.0” (Leyh, et al., 2017). Supporting the comments from Leyh et al. (2017) with regards to Lean being viewed as a basis for I4.0, Rossini et al. (2019) argued, through a study of 108 European manufacturing companies, that “higher adoption levels of I4.0 may be easier to achieve when Lean practices are extensively implemented”. Further it was indicated that “in opposition, when processes are not robustly designed and CI practices are not established, companies’ readiness for adopting novel technologies may be lower” (Rossini, et al., 2019).

In a study of 147 manufacturing companies in Brazil, Tortorella et al. (2019) evidenced that purely technological adoption (I4.0) will not lead to distinguished results. Lean practices help in the installation of organizational habits and mindsets that favor systemic process improvements; supporting the design and control of manufacturers’ operations management towards the fourth industrial revolution era (Tortorella, et al., 2019). Although the authors acknowledged that, since I4.0 is are relatively recent topic especially in developing economies such as Brazil, there is a need for in-depth case studies; studies that address common barriers or difficulties with Lean implementation and I4.0 implementation. This may allow a better understanding of the inherent challenges on implementing these approaches concurrently. The lack of empirical research in this emerging field provides ample opportunities for further investigation…”. Supporting Tortorella et al (2019) with regards to further research is a study conducted by Yin et al. (2018), which suggests that: “…detailed case studies that are rigorous, deep, and insightful to explain how to create, manage, operate, and maintain production systems (such as Lean) in the context of I4.0” are a recommended direction for future research (Yin, et al., 2018, p. 858). Yin et al (2018) also highlighted that due to the interdisciplinary characteristics of the three areas of business, engineering and information technology, studies in the past have been performed from different perspectives and hence a study of I4.0 from an information technology perspective might miss manufacturing engineering aspects such as CI.

In a study by Sony (2018), Lean is outlined as a contributor to the successful implementation of I4.0. The author proposed an integrative framework for horizontal, vertical and end-to-end engineering of I4.0 with the aid of iteratively performing five basic Lean principles: identify value requested by the customer, value stream mapping for each product, create continuous flow, establish pull and seek perfection. These Lean principles are originally proposed by Womack and Jones in 1996 (Womack & Jones, 1996). The author also suggested two research proposals aiming to investigate the hypothesis that self-regulated mechanism through “smart data” from products in end-to-end engineering integrations as well as the vertical integration of hierarchical subsystems can provide a culture of CI (Sony, 2018). Bold hypothesis as such, it indicates the need specifically investigate the impact of I4.0 “Connectivity” on the culture of CI.

2.6.2. I4.0 as an enabler towards Lean

Rüttimann & Stöckli (2016) adopted a more critical approach in terms of the I4.0 benefits in relation to the other authors conceptually studying the interrelation between I4.0 and Lean. Based on discusison following a presentation from Fertigungstechnisches Kolloquium organized by the Institute for Machine Tools and Manufacturing (IWF) of ETH Zürich; the authors suggested that I4.0

23

will not materialize as a revolution but rather as an incremental process that needs to be integrated into the existing Lean theory framework. They further suggested that any attempts to implement I4.0 without a basic understanding of fundamental manufacturing performance laws have a high probability to fail. In conclusions, the authors argued that I4.0 makes Lean production more flexible, but whether it makes it faster and more stable needs to be proven. An ending statement provides the authors’ perspective on radical implementation for Industry 4.0: “Industry should learn to walk first before it may dream of flying” (Rüttimann & Stöckli, 2016).

Wagner et al. (2017) investigated successfully implemented I4.0 projects within a large automotive company and created a impact matrix of I4.0 Connectivity (horizontal and vertical integration) and Intelligence (big data and analytics) on Lean principles from the input of 24 project leaders during a workshop. In conclusion, the authors stated that I4.0 stabilizes Lean processes and that the vertical and horizontal integration will have a large impact on the principle of CI (Wagner, et al., 2017). However, it was not outlined as to “how” CI is improved by I4.0 technologies impact on Lean principles. Moreover, the fact the authors choose sixteen I4.0 project leaders as participants in the workshop and only eight Lean experts may have resulted in a technocratic bias. The fact that the authors only investigated successful I4.0 projects further contributed to results aligned with the current I4.0 hype. Following the same line of arguments as Wagner et al. (2017), Kolberg and Zühlke (2015) concluded that Lean and I4.0 do not eliminate each other, but rather enhance value for the user when combined. By smart products, smart machines, smart planners and smart operators connected with smart watches and augmented reality; JIT and Jidoka principles can be reinforced. This can be partly realized through real-time updates of manufacturing cycle time and failure issues notifications (Kolberg & Zühlke, 2015).

Powell et al. (2018) also suggested that I4.0 could enhance JIT and Jidoka by better planning and more effective coordination with upstream supply through data analytics and by instantaneous signaling for overproduction, waste and replenishment through Connectivity and data analytics. They also suggested that enhanced continuous flow can be established by automatic scheduling adjustments through smart automation and Intelligence, and that automatic guided vehicles can reduce the unnecessary movement of operators and that operators’ capabilities can be enhanced through virtual tools (Powell, et al., 2018). Supporting the conclusions of Powell et al. (2018) were Romero et al. (2018), who conclude that new capabilities enabled by Intelligence and Connectivity can help eliminate waste. However, the authors also argued that the redundant use of digital capabilities of smart manufacturing technologies can give rise to digital waste (Romero, et al., 2018). Lastly, Mora et al. (2017) developed a theoretical implication matrix of Lean and I4.0 and found that JIT can be enhanced by improved process time through advanced analytics; continuous flow through augmented reality; and increased production and material data efficiency through connected machines. The implication matrix was filled by using existing applications found on publicly available information collected from websites and company brochures (Mora, et al., 2017). The findings were arguably positively biased towards the I4.0 impact on Lean practices which was not surprising considering the fact that the information contained in such sources is often biased for commercial purposes.

24

2.7. Summary of background and literature review

I4.0 is a difficult-to-grasp topic and as of today there exists no consensus around its true meaning. As a result, there are a wide range of different definitions and perspectives on I4.0 and its implications – implications which are often positively biased. In this thesis, I4.0 has been defined through three principal value drivers: Connectivity, Intelligence and Flexible automation. Lean however, is a proven strategy to achieve cost reduction through waste reduction by inducing “problem-sensitivity” into the organization which is a prerequisite for CI.

As stated by several authors, the literature lacks comprehensive frameworks which combines I4.0 solutions with Lean. However; conclusive from those articles found within the interdomain of Lean and I4.0, there seems to be an overall consensus that Lean and I4.0 are complementary rather than contradictory. An increased Connected, Intelligent and Flexibly automated manufacturing environment enable new possibilities that can potentially enhance Lean practices, e.g., JIT and Jidoka, which benefits problem-sensitivity. Lean can also ease the implementation of I4.0 by streamlining the process and by steering the focus to technologies that yield the most value to the customers and organization. However, it becomes apparent that the literature lacks a more realistic and critical approach to the interrelation between the two domains. Example of gaps in the studied literature are topics surrounding the impact on I4.0 on CI following the lack of standards and best practices; cyber-security issues; decentralization of decision making; inaccuracy in data interpretation; outsourcing of competence; and acceptance for change. These are some topics that are not part of the discussions in the current Lean-I4.0 related literature.

Several articles have been found, see Table 1, which study the barriers to CI within organizations; barriers that in this thesis have been divided into three factors (managerial, employee and environmental). However, none of the articles discusses the possible implications of an increased technology centered environment as a potential barrier to an organization’s CI-capabilities through these factors. In conclusion, the absence of the above-mentioned topics in the discussion surrounding the present Lean-I4.0 integration frameworks implies that the implementation Challenges and Risks associated with I4.0 adoption have been overlooked.

Table 5 displays an overview of the possible interrelation between CI and I4.0 by summarizing the main findings from the articles outlined in Chapter 2. I4.0 have been divided into its three value drivers as defined in section 2.4.1-2.4.3 and CI has been divided into two sections, i.e., “see problems”, related to problem-sensitivity built by Lean values and principles as discussed in section 2.1 and “solve problems”, related to problem-solving capabilities. The categories Standardized Work; Competence & mindset; Resources & time; and Supervision & follow-up are all viewed as independent of the I4.0 value drivers. Furthermore, the positive aspects of I4.0 on the conditions for CI are marked in green (+) and the negative aspects re marked in red (-). It is assumed that managerial and employee factors, outlined as barriers towards CI in Table 1, section 2.2 (here displayed under “solve problems”) might be reinforced by the introduction of the value drivers. Observably, the positive arguments under “see problems” dominate over the negative arguments (nine to one), which signal the absence of critical assessments concerning the Challenges and Risks that I4.0 might impose on the Lean principles. It can also be observed that “Connectivity” is the most investigated value driver. Not too surprisingly, considering that it constitutes a foundation for both

25

Intelligence and Flexible automation. The information contained in Table 5 will be referred to continuously in the discussion (Chapter 5) as the arguments are important factors to consider when critically assessing the I4.0 impact on CI.

Table 5: I4.0 value driver’s impact on CI as seen from current literature.

Industry 4.0 value drivers Connectivity Intelligence Flexible

automation

See

Prob

lem

s

Jidoka Real-time failure issues notifications (Kolberg & Zühlke, 2015)

Increased transparency using data mining & increased quality by applying AA for detection and prediction (Lorenz, et al., 2019)

JIT

Real-time cycle updates accessible to the operator enable continuous flow (Kolberg & Zühlke, 2015)

Decreasing the time between failure occurrence and failure notification through augmented reality and connected operators (Mrugalska & Wyrwicka, 2017)

Continuous flow through augmented reality, increased production and material data efficiency through connected machines (Mora, et al., 2017)

Incomplete goods’ shipping status, mismatch in quantity, unexpected delays during transportation can be solved by IoT through item-tagging and wireless tracking (Mayr, et al., 2018)

Increased digital waste through over processing of digital data & overburden of digital systems (Romero, et al., 2018)

Effective planning and waste signaling using data analytics (Powell, et al., 2018)

Decentralized decision making and real-time inventory tracking using RFID can enhance JIT pull-production and continuous flow. (Mayr, et al., 2018).

Automatic scheduling and quick production adjustments (Powell, et al., 2018).

Standardized Work

Lack of standards and missing best practices, “missing international norms”, “lack of compatibility of ports and interfaces” and “missing standardized data formats” (Glass, et al., 2018).

The transformation of the shop floor will lead to a decrease of low-skilled Standardized Work and in an increase in high-skilled work activities (Bonekamp & Sure, 2015).

Solv

e Pr

oble

ms

Competence & mindset

Learning and unlearning necessary behavioral patterns to sustain CI (Caffyn, 1999) Limited knowledge and motivation for performing necessary activities. (Lodgaard, et al., 2016) Difficulty in breaking traditional mindset and adapt to new ones (Bhuiyan & Baghel, 2005) Resistance towards technology diffusion decreases employee commitment (Schneider, 2018)

Resources & Time

Increase in maintenance costs can tie up capital for other investments (Rüttimann & Stöckli, 2016) Free up time by decrease unnecessary movement with automated vehicles (Powell, et al., 2018)

Supervision & followup

Lack of management feedback and involvement (McLean, et al., 2015; Garcia-Sabater & Marin-Garcia, 2011)

Lack of management support and supervision (Lodgaard, et al., 2016)

26

3. Methodology

In this chapter, the research methodology is reviewed and discussed. The chapter contains discussion of the research design including the benefits and shortcomings of the Delphi method (section 3.1), data collection and sample selection (section 3.2), operationalization (section 3.3), data processing (section 3.4), reliability (section 3.5), validity (section 3.6) and considered ethical research aspects (section 3.7). Details of the survey questions that has been used in both iterations of the data collection process can be found in Appendix A and Appendix D.

3.1. Research design

In this thesis, an explorative research design has been utilized. This implies that the aim of the presented research has been to clarify and deepen the insight of the studied phenomena rather than to derive precise and conclusive results. Thus, the conducted research has been aimed at shedding light on the ambiguities of the I4.0 impact on CI in order to lay a foundation for further discussions as well as to provide directions for future research. As explained by Saunders et al. (2019), an exploratory research design is particularly useful when aiming to clarify and deepen the understanding of an issue, phenomena or problem; especially when the nature of the research topic is unclear (Saunders, et al., 2019, p. 186). Hence, due to the novelty of the topic and because I4.0 is called out a priori of its realization, studies aiming to investigate the I4.0 adoption impact on managerial practices can only be exploratory as opposed to descriptive or explanatory. In the context of this research, the explorative nature has been highlighted through the research questions, R1, R2 and R3 (section 1.3) and through the chosen data collection method the Delphi survey and its constituting open-ended questions. Furthermore, a thematic data analysis strategy has been embraced when processing the collected data. This implies that categories and themes have been identified during the processing of the collected qualitative data. Thematic analysis is both systematic and flexible and facilitates the identification of key themes and patterns for further exploration; particularly useful for the thesis’ explorative research design (Saunders, et al., 2019, p. 651).

3.1.1. Delphi survey method

The Delphi method, also known as the Delphi technique, is a semi-structured and iterative survey method. It is suitable for explorative research purposes; thus, through the Delphi method, the future impact of I4.0 on CI has been explored through the collective Intelligence of a panel of experts. The main use of Delphi studies is to create future scenarios and since most companies are still in the early phases of their I4.0 implementation; this method has been deemed suitable.

The Delphi method was originally developed by Olaf Helmer and Norman Dalkey at the RAND Corporation in 1959 (Dalkey & Helmer, 1963). RAND Corporation ("Research ANd Development") is an US policy think tank created in May 1948 by the Douglas Aircraft Co. to conduct research for the US military (Rand Corporation, 2020). The original purpose of the Delphi method was to predict future technological capabilities of the armed forces. Several alternative approaches to make such predictions were initially tested, such as quantitative models and trend extrapolation but without

27

satisfactory results which led to the development of the Delphi method. The name Delphi is derived from the Oracle of Delphi in ancient Greece. Table 6 summarizes some of the benefits and shortcomings of the Delphi method.

Table 6: Benefits and Shortcomings of the Delphi method.

Benefits Shortcomings Suitable for explorative purposes Suitable for creating future scenarios Suitable for geographically dispersed

participants

No standard statistical tests Sampling and expert selection procedures

are not clearly defined Difficult to replicate

Using the Delphi method, respondents answer several rounds of questioners and the group response of each round is consolidated and feedback to each respondent before they answer the subsequent round of the questioner. Respondents are carefully selected based on their expert knowledge in the area being studied. The intended outcome is that the result of a group is more accurate compared to a pure compilation of individual views. Julsrud & Priya Uteng echos this “The methodology can be said to have foundations in the notion that several brains are better than one, especially in areas with a high degree of complexity and uncertainty” (Julsrud & Priya Uteng, 2015).

There are some key attributes to a Delphi studies:

Anonymity: The panel of experts are anonymous to each other before, during and after the process. This helps to prevent that anyone in the panel have any form of personality impact, such as dominating or authoritative, on the other participants. It also makes each expert freer to express themselves and to change and update their initial reply, there is less prestige in the answer provided and hence easier to gravitate towards stability.

Iterative: The input after the first round is collected and consolidated. Any information that is judged irrelevant is removed. This collection and filtering are done without any consultation of the experts and hence potential difficulties due to group dynamics or face to face discussions are eliminated.

Feedback: The consolidated input is provided to the experts and they are asked to reflect and provide input whilst all seeing the same information from the previous round. Anyone with an opinion deviating from the consolidated information will then either stand their ground but more likely reconsider their answer and reflect on the group replies and potentially update and revise his/her answer.

3.2. Data collection

The data has been collected through Delphi survey method, as described in the previous section. The survey consisted of three open-ended questions for the first iteration and three open-ended questions for the second iteration. The first three questions have been designed as to encourage participants to reason around both the potential positive and negative aspects of I4.0 impact on CI.

28

The follow-up questions for the second iteration have been designed on the basis of the findings from the data obtained in the first iteration. Thus, the second iteration encouraged the participants to problematize around the topic as well as to identify potential success factors to ensure minimum negative impact from I4.0 on CI. Through the iterative process of the Delphi survey, the researchers have directed and narrowed the research to clarify and deepen the understanding of emerging questions following the revelation of new data and insights; a process that is characteristic of an explorative research design (Saunders, et al., 2019, p. 187). The survey was attached as a word file and sent out using email to all participants. This was deemed the most affordable method in terms of time and resources for the purpose of reaching all participants. Table 7 outlines the different steps which was used to prepare and collect the data.

Table 7: Data collection steps.

Steps Description 01 Create the three open survey questions for the first iteration in cooperation with three

stakeholders (Appendix A) 02 Identify suitable participants in the four different categories described in section 3.2.1.

03 Create email template used to share the questioner (Appendix C)

04 Identify at least four participants to pilot the survey (Appendix B)

05 Share the first iteration questioner with the above four participants

06 Adjust the survey questions and email template based on the feedback from the piloting

07 Share the first iteration of questioner as an attached word file by email with the full sample population and allow a one-week time limit for reply

08 Send a reminder to people who did not yet reply after two days and after four days

09 Review and consolidate the answers provided, process the collected data by identifying themes according to the procedure described in section 3.4.

10 Review and update the questioner with the processed data from the first iteration along with three follow-up questions (Appendix D)

11 Share the consolidated answers from the first iteration with the people along with the second iteration questions (Appendix E)

12 Send a reminder to people who did not yet reply after two days and after four days 13 Review and consolidate the answers provided, process the collected data by identifying

gaps and themes 14 Decide based on the answer if a round three is required or if there are any further in-

depth interview required with anyone 15 Move into analysis and discussion phase

3.2.1. Sample selection

This study made use of individuals with knowledge about the investigated topics, i.e., Lean and I4.0. In the context of the study, the term “expert” has been avoided due to its unclear definition and resulting controversy (Hasson, et al., 2000); instead, the targeted participants have been referred to as “a panel of experienced individuals. Furthermore, in order to pinpoint suitable candidates to join

29

the panel, a purposive (also known as judgmental) non-probabilistic sampling method has been used, targeting participant based on two criteria’s: (A) their experience within the subject and (B) their present exposure to and involvement with the topic. Hence, the non-probabilistic sampling has been deemed to be the most suitable technique for this study; firstly, due to the absence of a clearly defined sampling frame, and secondly; because of the need to include subjective judgements based on the above-mentioned criteria’s (Saunders, et al., 2019, p. 315). The judgmental screening has been preceded by “snowballing” GKN and Volvo CE for suitable candidates, allowing participants to recommend other potential participants both within and outside the two organizations. In total, 33 participants have been found, from four different groups of categories:

1. GKN Aerospace (12) 2. Volvo Construction Equipment (12) 3. I4.0 Lighthouse companies (5) 4. Academia (post doctorial or more senior academic staff) (4)

There are presently no recommendations or guidelines concerning appropriate sample sizes for Delphi surveys. However, qualitative assessments as in this study, generally requires a smaller sample size in comparison to quantitative analysis. As a rule of thumb, the sample size needs to be sufficient to ensure data saturation, i.e., to ensure that the research questions are sufficiently addressed and that the studied phenomena is sufficiently described (Saunders, et al., 2019, p. 315). For this study, more than 15 replies in total has been deemed necessary to reach a sufficient saturation level.

3.3. Operationalization

This research has not been aimed at establishing casual relationships though statistical analysis between dependent and independent variables but has rather aimed at exploring and yielding insight into an unstudied topic. Thus, in the context of the conducted qualitative research, the term “variables” has been avoided. Instead, I4.0 has been referred to as the “independent concept” and CI as the “dependent concept”. The independent concept, I4.0, has been conceptualized into three value drivers (Connectivity, Intelligence and Flexible automation); whose definitions have been framed by a set of disruptive technologies as outlined in section 2.4. By dividing the abstract concept of I4.0 into three more distinctly defined value drivers, the authors have aimed to clarify and deepen its definition in order to further contrast the findings.

The dependent concept, CI, has been divided into “see problems” and “solve problems”. The first relates to the sensitivity to problems, i.e., the ability of an organization to see problems; built by Lean values and principles as described in section 2.1. The latter theme relates to the organizations’ ability to solve problems; built by, but not limited to, behavioral factors, resources, time and competence. Both themes can arguably be viewed as prerequisites to sustain CI throughout the organization. It is further assumed that the “improvement rate” can be affected by altering either the improvement level per activity or the frequency of improvement activities, or both. Moreover, both the improvement level and improvement frequency are assumed to be positively affected by enhanced problem-sensitivity and problem-solving capabilities and vice versa.

30

Three main dimensions of impact for each value driver have been identified based on nature of the participants’ responses, i.e., Possibilities, Challenges and Risks as outlined below. Figure 7 shows a graphical illustration of their impact on the rate of improvement. In this thesis, Possibilities have been defined as inherently positive, while Risks have been defined as inherently negative. Challenges have also been defined as negative because of the effort, resources and time needed to overcome them.

Possibilities: Include factors that positively impact the rate of improvement in relation to the baseline through enhanced problem-sensitivity or problem-solving capabilities, or both. In addition, Possibilities might result in a significant step-improvement; directly increasing performance in relation to the baseline before the I4.0 implementation.

Challenges: Include factors that need to be overcome in order realize Possibilities or to mitigate4 Risks. Thus, if a challenge is not overcome, it either results in a lost opportunity or in a failure to adequately mitigate a risk. Regardless, the improvement rate will follow the baseline, i.e., the challenge neither decreases nor increases the rate of improvement. Instead, the resulting increase or decrease in the rate of improvement following a challenge is related to the accompanying possibility or risk, not to the challenge itself.

Risks: Include factors that negatively impact the rate of improvement in relation to the baseline by undermining problem-sensitivity or problem-solving capabilities, or both.

Figure 7 Graphical illustration of Possibilities, Challenges and Risk in relation to a Baseline.

The operationalization of the research objectives through key independent and dependent concepts are illustrated in Figure 8. For a more through definition of the concepts of both CI and I4.0, the reader is referred to Chapter 2.

4 Mitigation: Implies reducing the probability and the negative impact of risk realization.

31

Figure 8: Operationalization of the collected data.

3.4. Data processing

This section outlines a description of the data-processing steps applied to the quantitative data obtained during the first and second iteration of the survey. The participants’ answers for each survey question observed in Appendix A and Appendix D (belonging to the three different value drivers) have been condensed into categories and themes according to the below described steps.

1. Create an overview of all the answers belonging to each independent value driver (Connectivity, Intelligence and Flexible automation) by manually sorting and copying them all into one document.

2. For each independent value driver, divide the constituting text into smaller text-strings and sections contacting key points and arguments while filtering out any unnecessary “filler text” not contributing to the participants’ line of thoughts.

3. Reduce the processed data further into single text-strings that capture the essence of the

argument whilst keeping it short and concise. This step induces a certain level of subjectivity as explained in section 3.6.

4. Sort each text-string into categories depending on their impacts and color code them accordingly, i.e., (green) for Possibilities, (orange) for Challenges and (red) for Risks as outlined in section 3.3. Note, for the processing of the data obtained in the second iteration the text-strings have instead been directly grouped into four categories (Purpose, Involvement of people, Competence and Implementation strategy).

5. Group arguments within each of the three categories into themes. For example, within the concept of Intelligence, three themes have been identified under the “Possibilities category”, i.e., perception, decision-support and autonomy. Each of the three sub-categories shall contain similarities in the participants’ line of reasoning.

The processed data according to the above steps can be seen in Table 9 to Table 14 in Chapter 4 together with an explanation of the identified themes.

3.5. Validity

In order to establish a high content validity, i.e., an adequate coverage of the research questions through the Delphi survey, careful formulation of the survey questions have been made using

32

feedback from three stakeholders with experience from CI and I4.0. Moreover, the survey was piloted by four participants to identify and correct any unclarities. During the piloting, the participants have been encouraged to give feedback on both the email formulation and the survey questions. This has been done in order to correct inaccuracies such as vague instructions and formulations as well as to minimize the risk of miscomprehensions. Regardless, whether the findings from the Delphi survey truly represent a valid ground for providing answers to the research questions should be subject for discussion. Since both CI and I4.0 are vague concepts that include various perspectives and definitions, there is a risk that validity is decreased following the deviation of participants’ answers from the intended measure. However, by purposefully narrowing the research and by clarifying any misconceptions during the second iterations, validity has been increased. Still however, as the participants can merely speculate about the future I4.0 impact on CI, it is nearly impossible to assure a high degree of validity. As concluded from a literature review by Hasson and Keeney (2011), it has to be accepted that “Delphi results do not offer indisputable fact and that instead they offer a snapshot of expert opinion, for that group, at a particular time, which can be used to inform thinking, practice or theory” (Hasson & Keeney, 2011). Moreover, as expressed by Professor Olof Helmer (one of the originators behind the Delphi method), research into the future “never will be, an exact science”. Thus; building on the above statements, it becomes apparent that it is inherently difficult to establish validity when studying future concepts, let alone measure it in the context of the chosen explorative research approach. Helmer (2011) further suggested that the main improvements of the Delphi method was related to “the selection of experts” and that it would be beneficial to have “greater flexibility in the number and nature of inquiry rounds” (Helmer, 1999). The selection of the experts in term of expertise and knowledgeability is something which Rowe & Wright, 1999 also emphasized and further concluded that “accuracy is improved over rounds as a consequence of the panel experts ‘holding-out’, while the less-expert panelists ‘swing’ towards the group average” (Rowe & Wright, 1999). Melander discusses the issues with trying to achieve consensus and argues that there is a risk that it hinders the development of more radical scenarios and hence consensus should not always be something which is strived towards in a study using the Delphi method (Melander, 2018). Steinert takes this even further and proposes an explorative research tool using the Delphi approach as to achieve dissensus (Steinert, 2009). In fact, Linstone and Turoff who wrote a seminal book on the Delphi method in 1975, reflects in a paper from 2011 that the need for consensus is a “persistent misperception” and instead the target is to achieve stability in the answers (Linstone & Turoff, 2011). Thus, fewer iterations and less focus on consensus is likely to encourage more radical scenarios and critical thinking while not necessarily decreasing the overall validity of the results. In order to facilitate validity to the extent possible, eight pitfalls as outlined by Tapio (2003) have been considered when setting up and conducting the Delphi survey (Tapio, 2003).

1. Biased selection of panelists 2. Disregarding organizations 3. Forgetting disagreements 4. Ambiguous questioners

5. Oversimplified structured inquiry 6. Feedback reports without analysis 7. Forgetting the arguments 8. Lack of theory (method understanding)

33

3.6. Reliability

The reliability of the Delphi survey as a data-collection method is generally weak due to the overall difficulty of replicating the research (Hasson, et al., 2000). As the Delphi survey provides a snapshot of expert opinions at a given time, replicative research (even targeting the same panel of experts used in previous studies) is unlikely to result in a similar outcome due to the dynamic nature of participant opinions and knowledge. Furthermore, as the collected data is filtered through the subjective opinions and knowledge of the authors before being sent back to the participants, further instability and inconsistency is induced; making the research even more difficult to replicate. Conclusive for the Delphi survey method is arguably that; the more iterations, the higher the reliability and vice versa. More iterations are likely to increase the expert consensus and the overall specificity of their answers, subsequently reducing the need for researcher interpretation and hence, researcher bias. Due to the purpose of minimizing the time for data-collection in this thesis, the authors have chosen to limit the survey to two iterations; placing less focus on panel consensus and detailed answers. This has induced a larger portion of researcher bias in the interpretation and processing of the data than if more iteration would have been used. Thus, because of the author’s relatively high involvement in the data interpretation and the data-processing, reliability was partly lost. However, to minimize the impact on reliability the results have been continuously discussed between the authors as to reduce any potential bias through consensus. Moreover, for the purpose of making the research easier to replicate and to grasp, a transparent line of reasoning in the data-processing has been outlined in section 3.4 and section 3.2. Furthermore, as the participants got the opportunity to answer the questions in their own time (in a closed space whenever they find it suitable to do so) participant bias could be decreased, as compared to if conducting structured or semi-structured interviews under stricter time constrains. One week to answer the questions between iterations was deemed sufficient to provide participant with the opportunity to sufficiently understand the questions and prepare their answers. Unfortunately, the establishment of a reliability measure through e.g., detailed comparison of similar data has not been possible due to absence of relevant studies.

3.7. Research ethics

The authors have embraced several ethical research considerations as listed by Vetenskapsrådet (2002) in their report “Ethical Principles within Humanistic and Social Research”. The considerations can be divided into four requirements that have served as a general guide for the conducted research. (Vetenskapsrådet, 2002)

Requirement of information: Each participant has been fully informed about the purpose of the study, their role as well as of how the collected data was handled. Moreover, the participants have been informed that participation was voluntary and that the participants reserved the right to withdraw from the study at any moment.

Requirement of consent: Each participant has confirmed that they agree to participate by answering and sending back the word document. The participants have also been informed

34

that they had the right to decide how long and on which terms they wished to participate under.

Requirement of confidentiality: The participants’ data has been kept confidential during and after the survey and no data has been published which could be traced back to any individual participant. The participants’ names have been kept anonymous before, during and after the survey.

Requirement of data usage: The participants have been informed of how their data was to be used. Moreover, the collected data has not and will not, under any circumstances, be subject to commercial use or other non-scientific purposes.

35

4. Results

In this chapter, the survey data in the form of text-strings containing participants’ key points and arguments are presented in addition to the survey response information. Section 4.1 contains information of the first survey iteration as well as the identified Possibilities, Challenges and Risks, and section 4.2 contains information of the second survey iteration and the identified success factors. The text-strings in Table 9 to Table 14 have been found by processing the raw-data according to the procedure described in section 3.4. Each text-string has been assigned a tag for later reference. For example, Connectivity Possibilities, number 1 in Table 9, is identified as CP01. Same numbering is used for Connectivity Challenges and Risks, CC01 and CR01. Same naming is also used for Intelligence (i.e. IP01, IC01 & IR01) and Flexible automation (i.e. FP01, FC01 & FR01).

4.1. First iteration survey

For the first iteration, the survey has been shared with 33 individuals resulting in 20 responses which corresponds to a response rate of 61%. Details can be seen in Table 8. The responses from the first iteration under the four main categories can be seen in Figure 7.

Table 8: Response rate details for the first iteration. Sent Received Rate GKN 12 7 58% Volvo CE 12 6 50% Lighthouse 5 4 80% Academics 4 3 75% Total 33 20 50%

Data with regards to number of years of experience within Lean and I4.0 was collected and a box plot of each of them can be seen in Figure 9 and Figure 10. The median years of experience of Lean was 11 years and the median experience of I4.0 was three years. This is an expected distribution as Lean is well established since a long time, whereas I4.0 is a much more recent topic.

Figure 9: Number of years of experience of Lean.

Figure 10: Number of years of experience of I4.0.

As observed in Figure 11, the experience in Lean was lowest (6.1 years) in GKN whereas it was high (>10 years) in the other three categories. This is mainly due to the type of industry where GKN Aerospace is operating, i.e., low volume and high variation, which has traditionally not been heavily

36

exposed to Lean initiatives. The level of experience of I4.0 within GKN (2.9 years) and Volvo CE (3.2 years) was relatively low as would be expected from “non-adopters”, but what was interesting is that the stated experience from the Lighthouse companies was also relatively low at 4.3 years. This gave rise to questions whether the right people in the Lighthouse companies have been targeted or if there were other conditions within the Lighthouse companies which allows them to rapidly move to a “Lighthouse status” such as culture, organization, competence or if perhaps their business environment is more suitable for I4.0 implementation.

Figure 11: Number of years of experience of participants per organization.

4.1.1. Connectivity

Figure 12 shows the number of text-strings found containing the participants’ positive and negative arguments for the Connectivity value driver’s impact on the conditions for CI. As can be observed, 69% of the answers were identified as positive (Possibilities) and 31% as negative (Challenges and Risks). This is indicative of an overall positively weighted attitude towards the implementation of Connectivity and its impact on CI in the manufacturing environment. In total, 54 arguments have been found, 37 as positive and 17 as negative.

Figure 12: Summary of positive and negative input on the impact of Connectivity on CI.

37

Each identified text-string under the “possibility category” for the Connectivity value driver can be observed in Table 9. Each text-string has been assigned a tag (CP01-37) for reference. Responses related to Possibilities have been sub-divided into three main themes as seen below. Nine responses have been classified under Collect data, 16 responses under Visualize data and 12 responses under Use data.

Collect data: Relates to the collection of data in an I4.0 manufacturing environment. In a traditional factory, there is already a lot of data collection taking place, both through manual and automatic means. However, in the future I4.0 environment, it is believed that this data volume will significantly increase, both with regards to the type of data collected but also its frequency and resolution. (CP01-CP09)

Visualize data: Relates to “how” the collected data is visualized. Large amounts of data are pointless unless it can be properly processed and visualized. (CP10-CP25)

Use data: Relates to “how” the collected and visualized data is used. This is discussed more in depth under the Intelligence value driver, section 4.1.2. (CP26-CP37)

Table 9: Possibility with introducing Connectivity in the manufacturing environment.

Possibilities (+)

Colle

ct d

ata

CP01 Enhanced Connectivity offers more automated ways for data collection and analysis instead of manual collection and analysis of equipment and production data.

CP02 Transparently and quickly gather quantitative data regarding improvements and development of processes.

CP03 Connectivity will bring precision into performance measurement. CP04 More direct, available and accurate data on the evolution of tolerances. CP05 By automatic data collection we can improve the conditions for other parts of the manufacturing

environment where it may be more difficult to introduce IoT and where man is needed. CP06 Instantaneously have and provide access to the needed information to improve to anyone

around the world and from anywhere in the world, thus potentially increasing the speed and quality of improvement.

CP07 Connectivity offers better solution than handwritten LTA charts, value stream mapping etc, data can also be converted into transparent and standardized dashboards and turned into action.

CP08 Faster implementation of CI since higher Connectivity allows for increased real time monitoring and two-way communication.

CP09 Data collection which is accurate and provide the level of quality and reliability that other methods fail to achieve.

38

Visu

alize

dat

a CP10 Current state vs ideal state will be clearer. CP11 Collect and visualize live data and trends so everyone sees the same picture. CP12 Real time data visualization will offer possibilities to find waste and hence improvement

opportunities. CP13 Provides a way of visualizing if we are winning or losing which is a fundamental part of building a

productive culture. CP14 Connectivity allows for increased real time visualization and improvement by two-way

communication systems and hence faster implementation. CP15 Connectivity will support better fact-based decision making. CP16 Access to comprehensive, up-to-date production information, together with a complete historical

picture, will take the guesswork out of CI activities. CP17 Data visualization will enable more in-depth analysis and hence create better process

understanding so that variation and waste can be addressed. CP18 Data of bottlenecks will be available in a more direct and accurate way. CP19 Bottlenecks and quality issues will be more visual. CP20 Avoidance of problems as KPI’s will be visualized before problems occur. CP21 Value Streams and processes could be better visualized, better connected, and metrics data

more easily accessed and interpreted. CP22 Connectivity will visualize the improvement opportunities and hence provide CI opportunities on

the shop floor. CP23 Real time data visualization will improve the daily shop floor meetings since they will be less

about the number and more about creating motivation and enthusiasm. CP24 Connectivity will change how data and information is visualized to the shop floor operators. CP25 Data visualization will enhance shop floor operators’ ability to perform by making information

more visual and easier to understand such as picking operations or warehouse operations.

Use

data

CP26 Connectivity will allow systems to self-manage towards better performance CP27 The use of connected factory will allow a better understanding of correlations and root causes

and hence improve problem-solving CP28 Problems will be discovered faster CP29 The risk of the human factor will decrease with use of connected data CP30 Data driven problem-solving which will be accelerated compared to today CP31 Data analysis will be more in-depth and comprehensive with better understanding of processes

and variables influencing variation CP32 Rapid response using data across the enterprise but also with external partners enabling rapid

response to changes in the environment CP33 Suggestion submission system for CI using video file from smart phone directly linked to the

process managing improvement suggestions CP34 Faster suggestion system for CI CP35 Occupational health and safety will be positively impacted by Connectivity due to sensors and

smart devices CP36 The use of Connectivity will enable collection of data which can be used for different types of

optimization to improve productivity which was not possible before CP37 The sample availability of data reduces the time to evaluate and implement process changes

39

Table 10 illustrates identified inertia tendencies for the Connectivity value driver. The Challenges sub-category has been divided into two themes, Value adding and Engagement & Change as displayed below. In total, nine responses have been classified as Challenges.

Value adding: Relates to Challenges associated with the collection of data and the value that it brings to the organization. (CC01-04)

Engagement & Change: Challenges associated with the engagement in CI among primarily the shop floor operators as well as aspects relating to how change is driven and implemented. (CC05-09)

The Risks sub-category has been divided into two themes, Digital waste and Rigidity as displayed below. A total of eight responses have been classified as Risks.

Digital waste: Relates to Risks associated with an overflow of unused data which might prevent visibility of the useful data. (CR01-03)

Rigidity: Risks associated with the rigidity of the production environment due to cost and complexity of moving sensors and the associated competence to make these changes. (CR04-08)

Table 10: Inertia tendencies derived from Challenges and Risks associated with Connectivity.

Inertia tendencies (-) Challenges Risks

Value adding CC01 How to handle, process and use the data

collected to make sure shop floor operators CI activities can be benefited.

CC02 To use the collected data to drive CI activity systematically.

CC03 Implement too quickly. CC04 Connectivity used in isolation with low

complete business or customer focus

Engagement & Change CC05 Low involvement of shop floor operators in CI

activities using Connectivity data. CC06 Saturation and exaggeration of Connectivity

technology losing “old fashion” problem-solving and CI.

CC07 Old production equipment could cause Connectivity problems and used as an excuse to not do CI.

CC08 Less go and see in the production flow. CC09 Peoples’ competence and skills could be used

as an excuse to not do CI

Digital waste CR01 Drown in data without knowing why it is

collected or how to use it. CR02 Too much data presented which does not

support the production flow CR03 Data visualized incorrectly and insights

misinterpreted.

Rigidity CR04 Less involvement from shop floor operators

due to lack of expertise in Connectivity might lead to loss of engagement

CR05 Sensors and actuators make the production environment more “static” even if technical assistance from IT department is available due to cost of change.

CR06 Data visualized to create a blame culture rather than inspiration and coaching.

CR07 The shop floor operators will need technical assistance from IT department to implement small simple and cheap improvements due to Connectivity hardware/software

CR08 Absorbed on Connectivity data use and loose holistic overview of step changes in manufacturing technology outside I4.0.

40

4.1.2. Intelligence

Figure 13 shows the number of text-strings found containing the participants’ positive and negative arguments for the Intelligence value driver’s impact on the conditions for CI. As can be observed, 61% of the text-strings have been identified as positive (Possibilities) and 39% as negative (Challenges and Risks). In total, 51 text-strings have been found, 31 positive and 20 as negative; indicative of an overall positively weighted attitude towards the implementation of Intelligence and its impact on CI in the manufacturing environment.

Figure 13: Summary of positive and negative input on the impact of Intelligence on CI.

Each identified text-string under the “possibility category” for the Intelligence value driver can be observed in Table 11. Each text-string has been assigned a tag (IP01-31) for reference. Responses related to Possibilities have been sub-divided into three main themes as displayed below. 15 responses have been classified under Transparency, seven under Decision support and nine under Autonomy.

Transparency: Relates to the Possibilities following enhanced visibility of problems and their solutions, both pre and post problem occurrence. This is made possible by generating data-driven insights from applying AA and ML on data-streams from an increasingly connected manufacturing environment. (IP01-15)

Decision support: Relates to Possibilities with “intelligent support” provided by AI and supportive AA-tools, enhancing problem-solving capabilities by supporting decision makers, facilitating faster and more accurate decisions. Decision support naturally follows from an increased transparency of problems and their solutions. (IP16-22)

Autonomy: Relates to the decreased human intervention by complementing and replacing human intelligence and by enabling self-governance, which can partly relieve the manufacturing from human errors. Transparency and Decision support are enablers for autonomous manufacturing. (IP23-31)

41

Table 11: Possibilities with introducing Intelligence in the manufacturing environment.

Possibilities (+) Tr

ansp

aren

cy

IP01 AI can substitute humans in the data analysis phase which may yield insights into higher levels of aggregated data.

IP02 AA and deep learning can yield insight into unknown problems and ways of how to solve them. IP03 AA can be used to investigate and improve areas of manufacturing processes that was not

possible before. IP04 AI can help us know what we do not know today. IP05 AI can find more advanced correlations. IP06 Machine learning and AI can solve problems that was previously unsolvable as it allows for

non-linear optimization. IP07 Enhanced understanding of how to improve efficiency and how to reduce problems. IP08 AA can provide pathways towards effective value streams and shorter lead times by increased

insight. IP09 AA and AI help us be faster and more agile by yielding insights and by speeding up problem-

analysis and problem-solving. IP10 Acceleration of CI through data-driven problem-solving. IP11 AA, AI and machine learning can help processing large amounts of data quickly to find root-

causes of past problems and to predict future problems and wastes. IP12 AA and AI can reduce the time to analyze and study production systems which will help

accelerate and focus CI to those activities with the biggest impact. IP13 AA can help manufacturers to predict outcomes and reduce errors. IP14 AI can be used to analyze the process, physically or digitally, or both. IP15 AA and AI can show the performance of an entire production system.

Decis

ion-

supp

ort

IP16 Enables the making of the right kind of decisions IP17 Faster analysis enabled by AA and AI, drives faster and more accurate decisions. IP18 Aiding the discovery of invisible patterns and deviations allows for more informed and fact-

based decision-making and choices. IP19 In the long-term AI can provide decision support. IP20 Faster decision-making and decreased need of management approval. IP21 Assistance from advanced analytics and AI decision-support tools for root cause analysis which

makes problem-solving easier. IP22 By increased data-analysis capabilities, employees can easier deduce the next course of action

in order to reduce waste and to optimize the process.

Aut

onom

y

IP23 Reduce human trial and error processes. IP24 Data analysis and AI can prevent errors such as forgetfulness and loss of concentration. IP25 Decreased need for operator intervention which decreases process variations and increases

process capabilities. IP26 AI to provide rapid fulfilment of human tasks with a higher quality outcome. IP27 AI can provide lower production cost by replacing humans in the manufacturing environment. IP28 Built-in CI driven solutions driven by machine learning and supervised by humans. IP29 Enabling self-managed systems which can optimize production. IP30 The use of big data and AA can potentially reduce mental overburden and product variability

by increasing design commonalities and effectiveness. IP31 Objective solutions through AI based analytics drives standard ways of working.

42

Table 12 illustrates identified inertia tendencies for the Intelligence value driver divided into several themes. The Challenges sub-category has been divided into two specific themes, Trust & Understanding and Education & Training as displayed below. A total of nine responses have been classified as Challenges.

Trust & Understanding: Relates to Challenges associated with gaining trust for intelligent solutions and their capabilities, understanding their potential and limitations, establishing clear targets for proper supervision as well as with sharing data externally to realize the potential of AA. (IC01-06)

Education & Training: Challenges stemming from the mindset, specific training and education needed for IT, Engineering and manufacturing personnel in order to implement and realize the potential of the intelligent value driver. (IC07-09)

The Risks sub-category has been divided into two themes, Data quality & Application and Ownership & Participation as displayed below. A total of 11 responses have been classified as Risks.

Data quality & Application: Risks derived from applying AI on data of inadequate quality, cyber-security issues and other risks following improper implementation, usage and supervision of its development. (IR01-08)

Ownership & Participation: Risks associated with the higher need for experts to realize even smaller improvement and with humans becoming less involved in the manufacturing process. (IR09-11)

Table 12: Inertia tendencies derived from Challenges and Risks associated with Intelligence.

Inertia tendencies (-) Challenges Risks

Trust & Understanding IC01 Difficult to implement improvements derived

from low trust and little understanding. IC02 Increasingly complex solutions require extra

attention in how to present results in order to gain trust from users.

IC03 Difficulty in showing the value of data-driven decision making.

IC04 Difficulty to define and therefore to introduce Intelligence into the manufacturing environment.

IC05 Well understood targets and scenarios need to be defined beforehand. Collection of information regarding exceptions and big deviations of results from intelligent systems then need to be followed up and resolved in relation to those targets.

IC06 To realize its potential, companies need to take advantage of “open-innovation”, i.e., to share data outside of the company borders.

Education & Training IC07 Specific education and training needed to

realize potential. IC08 High demand for engineering experts to

Data quality & Application IR01 Applying Intelligence on low quality data

following poor data-collection can yield unwanted results.

IR02 Too much data provided by AI and AA can bring frustration before decision making.

IR03 Difficulty in knowing if the machine is correct in its prediction.

IR04 AI is used to confirm what experts already know, making AI redundant while potentially decreasing objectivity.

IR05 Rapid growth can be damaging because reflection and proper implementation are overlooked.

IR06 When not considering organizational change and learning into its integration, Intelligence can be destructive.

IR07 Intelligence being a goal instead of a means. IR08 Compromised data-security as data is shared

within and outside the company borders. Ownership & Participation IR09 As AA and AI becomes a more integrated part

of production, “experts” are needed to make even small changes.

43

realize potential (within data and IT) as well as IT-personnel with engineering knowledge.

IC09 Management will need to change their mindset about CI as self-managed systems becomes a reality.

IR10 Benefits are limited to a selected few and high-tech companies that have adequate resources.

IR11 The machine operator becomes removed from the improvement work which might decrease the overall problem-sensitivity.

4.1.3. Flexible automation

Figure 14 shows the number of text-strings found containing the participants’ positive and negative arguments of the Flexible automation value driver’s impact on the conditions for CI. As can be observed, 63% of the text-strings have been identified as positive (Possibilities) and 37% as negative (Challenges and Risks). In total, 38 text-strings have been found, 24 positive and 14 as negative; indicative of an overall positively weighted attitude towards the implementation of Flexible automation and its impact on CI in the manufacturing environment.

Figure 14: Summary of positive and negative input on the impact of Flexible automation on CI.

Each identified text-string under the “possibility category” for the Flexible automation value driver can be observed in Table 13. Each text-string has been assigned a tag “FP1-24” for reference. Responses for Possibilities relating to Intelligence were sub-divided into two main themes as displayed below. 13 responses have been classified under flexibility and 11 responses have been classified under automation.

Flexibility: Relates to the Possibilities brought on by AM and advanced robotics to respond quicker and more efficiently to the dynamic nature of the environment. Flexibility is facilitated partly through the integration of Intelligence into machines and robots. (FP01-13)

Automation: Relates to the Possibilities brought on by increased automation of intellectual as well as repetitive tasks, decreasing human intervention and workload. This can decrease waste and enable additional time for CI activities and faster trial and error processes. (FP14-24)

44

Table 13: Possibilities with introducing Flexible automation in the manufacturing environment.

Possibilities (+)

Flex

ibili

ty

FP01 Faster and more efficient production will be enabled. FP02 More complicated design will be able to be produced. FP03 Mass customization and single piece flow can become a reality. FP04 Co-location, location at the point of need and JIT deliveries can be realized. FP05 AM can decrease material waste and increase productivity by faster manufacturing closer to the

customer. FP06 Flexible responses to a changing environment and to customers’ needs opens up CI

opportunities. FP07 Try-storming new ideas and testing potential improvements will become easier. FP08 Possibilities to conduct CI activities remotely from distance. FP09 Decreased lead time for testing of new products and processes by using e.g., 3D-printed

prototypes that can help simulation of tolling and the assembly processes needed. FP10 Allows faster proof of concepts and testing of improvement ideas which de-risk the

implementation and shortens the CI-cycle. FP11 AM reduces the amount of time from idea to implementation by quick manufacturing of

fixtures and prototyping. FP12 Decreased need of supplier material and equipment can open up for the possibility to efficiently

work with cost and quality issues that before was not possible. FP13 AM can secure optimal solutions for the manufactured part instead of sub-optimizing solutions

per manufacturing step.

Auto

mat

ion

FP14 The use of advanced robotics and cobots will facilitate the elimination of waste as the processes will be easily observed and repeatable.

FP15 When variability cannot be limited, or reduced to levels of minimal impact, cobots can help to “pick up slack”.

FP16 Autonomous vehicles could be beneficial for internal logistics by moving parts through the factory and between sections with more precision and reliability.

FP17 Yield more quick opportunities to reduce and eliminate the 3Ms. For example, advanced robotics and cobots will ease the mental and physical overburden.

FP18 Further automation of heavy repetitive tasks that are bad for ergonomics and that do not present intellectual changes for people.

FP19 Advanced robotics fully automates production yielding low to basically non-existent manual set-up time.

FP20 Possibility to empower the operators and low-level management so that they can give more ideas to solve their daily problems and be more efficient.

FP21 Increased automation results in more time for CI-work as it shapes the overall work of the operator to something more than pushing buttons.

FP22 Increased automation can help organizations to directly improve operator safety, improve quality and reduce costs.

FP23 Cheaper production as less operators are needed. FP24 3D-printing can speed up the prototyping process and change the way a product or component

can be manufactured and designed. Table 14 illustrates identified inertia tendencies for the Flexible automation concept divided into several themes. The Challenges sub-category has been broken down into two specific themes, Knowledge and Strategy & Standards as displayed below. A total of eight responses have been classified as Challenges.

Knowledge: Relates to the Challenges in understanding correlations following increased technical complexity, understanding applications for collaborative robots and involving operators in more knowledge demanding CI-activities. (FC01-05)

45

Strategy & Standards: Relates to Challenges in establishing and maintaining industrial standards, establishing a clear strategy as well as to set adequate targets. (FC06-08)

The Risks sub-category has been broken down into two themes, Excessive solutions and Rigidity as displayed below. A total of 6 responses have been classified as Risks.

Excessive solutions: Risks that Flexible automation becomes redundant as it is used where it is not needed and where less costly solutions could prove as suitable or better alternatives, resulting in over-engineered solutions in sub-optimized processes. (FR01-03)

Rigidity: Risks associated with an increased rigidity of the manufacturing systems. Contrary to flexibility; rigidity makes systems more difficult and costly to improve and can potentially accumulate more problems for the users than it temporarily solves. (FR04-06)

Table 14: Inertia tendencies derived from Challenges and Risks associated with Flexible automation.

Inertia tendencies (-) Challenges Risks

Knowledge FC01 The complexity in understanding why

somethings happens will increase with autonomy and automation.

FC02 New unknown challenges can arise that will affect improvement efforts and learning.

FC03 No clear applications have been found for cobots in low volume and low “manual labor” environments.

FC04 As operators and assemblers will be affected to a large extent by Flexible automation, they need to be involved earlier in the CI-process.

FC05 Achieve adequate technology readiness level of metal 3D-printing.

Strategy & Standards FC06 To keep up practices and standards for

world-wide industrial groups when creation and evolvement of the process is different in each of the plants.

FC07 Automation strategy must align with the business and operation strategy (so far, few organizations have been capable of using advanced automation and provide clear return of investments).

FC08 Difficulty in setting clear targets and purposes with the technology to measure improvements.

Excessive solutions FR01 With a poorly designed factory layout and

production line configuration, automation does not solve problems, it instead creates new ones.

FR02 Activity should only be undertaking if it solves a well-defined problem. Having the technology might tempt people in to using it when it is not actually needed.

FR03 The wide adoption of Flexible automation in the manufacturing environment will not eliminate process variation or wastes in production flow.

Rigidity FR04 It may complicate things and being more

engineering developed and dependent can create more cost on problem solutions.

FR05 Important to purchase and install systems that have the possibility to be improved (CI built-in). Rigid systems will likely be too costly to improve.

FR06 Less manual work reduces the team member’s ability to make improvements.

46

4.2. Second iteration survey

For the second iteration, a second survey has been shared with the respondents of the first iteration, i.e., a total of 20 individuals. The survey can be seen in Appendix D. Responses have been received from 10 individuals resulting in a response rate of 50% (details can be found in Table 15)

Table 15 Response rate details for the second iteration.

Sent Received Rate GKN 7 2 29% Volvo CE 6 5 83% Lighthouse 4 2 50% Academics 3 1 33% Total 20 10 50%

The main purpose of the second iteration has been to share the summary of the results of the first iteration. For this purpose, an executive summary document has been created consisting of three pages, as seen in Appendix E, showing summaries of the result illustrated in Table 9 to Table 14. Based on this summary, the respondents have been asked to identify success factors for the implementation of I4.0; factors that can ensure Possibilities to be embraced, Challenges to be overcome and Risks to be mitigated without impacting CI negatively.

4.2.1. Success factors

This section outlines the success factors for I4.0 implementation in relation to CI as identified by the participants in the second round of the survey. Each text-string displayed in Table 16 to Table 19 have been found by processing the raw-data from the second iteration according to the procedure described in section 3.4. 4 different categories of success factors have been identified, i.e., Purpose, Involvement of people, Competence and Implementation strategy. The summary of the responses can be seen in Figure 15. Each category is presented in section 4.2.2 to section 4.2.5, respectively.

Figure 15: Summary of responses from the second iteration.

47

4.2.2. Purpose

Table 16 shows the success factors categorized into the “Purpose” category. In short, the category concerns the success factors related to the overall purpose of embarking on an I4.0 transformation as to reach the potential of I4.0 while minimizing the negative impacts on the conditions for CI. The success factors are listed under its constituting value driver; Connectivity Success factor (CS01-03), Intelligence (IS01-03) and Flexible automation (FS01-03).

Table 16: Success factors related to “Purpose”.

Success factors CS01 It has to solve a real problem for the affected workforce. Emphasize the benefits for shop floor

operators and others. CS02 Define clear value of the Connectivity before implementation. CS03 People tend to connect everything and collect all data possible, but this does not make sense.

IS01 Having clear responsibilities and goal settings. IS02 Having an idea on business case (cost for implementation/competence development vs effects of

implementation) on what technological implementation can bring as opposed to not implementing.

IS03 Find well-defined problems that can be resolved using “Intelligence”, i.e., make it a means to solve problems instead of making it a goal itself.

FS01 Ensure that the implementation has been thoroughly assessed with the involvement of all areas

of the business, including the shop floor. FS02 Do not apply Flexible automation thinking on cost savings, but rather on value added to the

customer. FS03 Assuming a strong CI culture exists, the success factors for Flexible automation would be having a

clearly visualized and agreed upon ideal state for the business and end to end flow.

4.2.3. Involvement of people

Table 17 shows the success factors categorized into the “Involvement of people” category. In short, the category concerns the success factors related to how people and stakeholders should be involved in the I4.0 transformation as to reach the potential of I4.0 while minimizing the negative impacts on the conditions for CI. The success factors are listed under its constituting value driver; Connectivity (CS04-11), Intelligence (IS04-09) and Flexible automation (FS04-05).

48

Table 17: Success factors related to “Involvement of people”.

Success factors CS04 Create an understanding of the value of the collected data for all stakeholders and how it will be

used. CS05 Foster a culture which confirms deviations with real data and facts such as go and see, discussing

with shop floor operators etc. CS06 Provide authority for shop floor operators to act on collected data and reduce hierarchies. CS07 Selection & classification of data (required collection and management) through opinions from

operators and experts. CS08 Be open minded and curious to find the best data. CS09 Foster a pragmatic culture where data is questioned to decrease the risk of misinterpretation. CS10 Ensure reliability of data by confirming correlation between operators and experts. IS04 Realize that the greatest Intelligence is found with the people who do the job currently; hence,

learn from those who work with the process and find how out they would approach it differently. IS05 Involve operators to keep them motivated for CI and involve humans in decision making. IS06 Keep human involvement and expertise, people are the greatest asset for CI IS07 Prevent that creative people who are the driving force behind CI, are pushed away from the

working area as Intelligence is integrated. These considerations between mathematically correct intelligence versus ownership has to be made.

IS08 People need to clear understand the purpose and logic when implementing AI, only then they will build trust for the proposed solutions.

IS09 Make sure AI is used for finding out the things we do not know, i.e., to find answers for questions that has not been raised yet, instead of confirming what is known.

FS04 Involve operators in decisions making. FS05 Make sure managers and operators understand what the engineers are implementing.

4.2.4. Competence

Table 18 shows the success factors categorized into the “Competence” category. In short, the category concerns the success factors related to training and knowledge as a means to reach the potential of I4.0 while minimizing the negative impacts on the conditions for CI. Each success factors are listed under its constituting value driver; Connectivity (CS12-18), Intelligence (IS10-14) and Flexible automation (FS06-07).

Table 18: Success factors related to “Competence”.

Success factors CS11 Training and competence development are critical to engage shop floor operators. CS12 Develop shop floor operator training for CI based on Connectivity. CS13 Have a clear strategy for competence development of both management and shop floor operators. CS14 Train everyone in statistics to support data analysis and create understanding of process capability. CS15 Develop understanding of correlations between input data. CS16 Develop understanding of casual relationships between input and output data. CS17 Develop a good process understanding.

49

IS10 Acquire and foster the needed data science skills within the organization. IS11 Make data scientist work closely with manufacturing. IS12 Consider competence development on both management and employee level before, during and after

implementation. IS13 Keep the pilot team small and open-minded, forward thinking members who can see the vision and

positives of successful AA/AI. IS14 Align IT-personnel with the CI-work. FS06 Develop human competence alongside the increase of Flexible automation. FS07 Assign multi-skilled teams to work on Flexible automation in order to include all functions beneficial to

CI-work.

4.2.5. Implementation strategy

Table 19 shows the success factors categorized into the “Implementation strategy” category. In short, the category concerns the success factors related to having an adequate business case and action plan of how to integrate I4.0 while minimizing the negative impacts on the conditions for CI. It describes how to approach the integration and what to pay attention to when implementing each value driver. Each success factors are listed under their constituting value driver; Connectivity (CS18-38), Intelligence (IS15-22) and Flexible automation (FS08-14).

Table 19: Success factors related to “Implementation strategy”.

Success factors CS18 Make a step by step implementation approach. CS19 Having a clear production strategy that includes both I4.0 pilot implementation and roll out. CS20 Do not spend more money on Connectivity than it gives back. CS21 Clear business case including cost of implementation and competence development vs benefits. CS22 Use visualization to make data more intuitional. CS23 Make sure that data is easily visualized. CS24 Development of data analysis tool and securing reliability: Utilization of various analysis tools, Data

simulation based on case study. CS25 Collect data across the whole extended value chain (from supplier to end customer). CS26 Be careful of how much data that is gathered and presented, if data is gathered and presented that is

not of relevance, CI can be impacted. CS27 Continuously review sampling speed and frequency as well as data quality and range. CS28 Data collection has to be reliable. If the data is incorrect or unstable CI can be impacted. CS29 Define clear process and decision points as well as review frequency of the data collection. CS30 That the integrity of the data is validated, and cross checked depending on the criticality, cost, impact,

or repercussions of data inaccuracy as a risk to success. CS31 Ensure that data collection is stable. CS32 Apply edge computing to minimize lag and potential data loss. CS33 Implement technology that actually works which often means simpler technology. CS34 Use Connectivity that is flexible and easily adapted. CS35 Make sure that IT and OT are aligned. CS36 Have a single source to go to for data to ensure that all data and trends are up to date. CS37 Reduce reaction lead time by using a common system for all information. CS38 Ensure data safety (cyber security) and use separate networks and storage depending on data type.

50

IS15 Roll out the digital strategy following a gradual process which allows for time to learn from mistakes and grow. Carefully plan, slowly do, check and act fast. Repeat.

IS16 Utilize the Lean toolkit to fully understand the value stream and its processes as they run in reality, not as they are described in the ISO or ME documentation. Then subsequently design AI to be implemented into the processes.

IS17 Establish a process by which the AA/AI is reviewed, validated, verified. This could be through either a project, by project validation based on a cost, risk, impact type of matrix; by a system of period review; or both in combination.

IS18 Starting small with pilot projects and things where the AA/AI, if incorrect, can either be easily caught, and/or has low impact.

IS19 Develop Standardized Work and define and continuously upgrade response scenarios. IS20 Threat the problem being solved holistically and integrate all aspects of People – Process – Technology. IS21 Keep the flexibility by avoiding resolving only simple and repetitive processes. IS22 Review the complex processes that are outside of the AI capabilities in order to scope and understand

how to integrate them. FS08 In regard to cobots, AGV/VGV’s, & additive manufacturing, it is a must to have a successful CI-program

based on the “proper thinking way”, principles, methods, and tools. FS09 Do not automate a bad process, it will merely lead to an automated bad process. FS10 Start with heavy and dirty processes that are requested by the workforce to be automated and maybe

not the most beneficial processes. FS11 Large volumes and bad ergonomics operations are first candidates for automation and very easily

received, but when complexity and need for flexibility appears, usually the automation is not the best solution.

FS12 Define what “flexibility” truly implies for the manufacturing by scoping its limitations and possibilities for each user-case.

FS13 Understand the lead time for changes in each automation implementation, how does that impact previous and next process, in case of changes how quickly we can react to adapt in manual vs automated set up.

FS14 Work with establishing standard processes.

51

5. Analysis and Discussion

In this chapter the empirical findings are analyzed and discussed. The three research questions (R1, R2 and R3) have provided a basis for the analysis, and the findings have been compared and elaborated on in relation to the theoretical framework. The analysis and discussion following the below sections are reflective of the authors own knowledge and perspective of the research findings. Therefore, the discussion should be viewed as the authors’ subjective generalizations of the participants’ findings concerning I4.0’s Possibilities, Challenges, Risks as well as success factors. Moreover, instead of critically reviewing the content of each text-string the discussion has served as to clarify and to reason around them.

5.1. Industry 4.0 Possibilities, Challenges and Risks

The discussions following sections 5.1.1-5.1.6 have been founded on the text-strings observed in Table 9 to Table 14. Each text-sting has been referred to directly in the text to support the line of reasoning made by the authors. For example, Connectivity Possibilities, number 1 in Table 9, is identified as CP01 when discussed to correlate the discussion with the analysis findings. Same numbering is used for Connectivity Challenges and Risks, for example CC01 and CR01. Same naming is also used for Intelligence (i.e. IP01, IC01 & IR01) and Flexible automation (i.e. FP01, FC01 & FR01). The discussions concerning the Possibilities for each value driver have been divided into the identified themes, whilst the discussion concerning the Challenges and Risks, follow a more general discussion.

5.1.1. Connectivity Possibilities

Collect data: Faster data collection due to it being automatic and automated is highlighted in several text strings (CP01, CP02, CP04, CP05, CP06 and CP08). Words such as quick (CP02), direct (CP04), instantaneous (CP06) and real time (CP08) were also words used to describe this. A future smart factory will most likely be able to provide most of the Connectivity data in real-time which is something which is supported by Kolberg & Zühlke (2015). That the automated data collection will be more accurate since no risk for manual errors or operator misinterpretation of settings and gauges is also found in several text strings (CP03, CP04 and CP09). Words as reliable (CP09) and precise (CP03) were also used to describe this. The fact that Connectivity enables automated, high quality data collections means that manpower (which is currently used to collect and process this data) can be freed up and deployed in other areas in the manufacturing flow to work on CI, especially in areas which might be more difficult to digitalize (CP05). The word transparent is also something which was re-occurring under collect data (CP02, CP06 and CP07). The participants see Possibilities in that the collected data will lead to a more transparent data management where data is easily available to everyone regardless of position in the company or geographical location. Easily available data is supporting Jidoka where visualization of performance gaps is important as to create problem-sensitivity; the more people that see the gap the larger possibility there is that action will be taken. Increased data transparency is something which is supported by Lorenz et al. (2019). Data collection is seen as a means of standardizing the way data is collected which will lead to, for example, better

52

value stream maps as compared to them being made manually (CP07) which will support JIT improvements. Overall the feedback and input with regards to Collect data seemed quite tangible for most respondents since this is something which is already done in many places, and the ratio between positive and negative input was quite balanced with 56% positive answers. This balanced ratio between positive and negative for collecting data is interesting especially when the responses were much more positive with regards to visualization (76% positive) and data use (71% positive). Could one explanation be that there is a more balanced opinion for activities which are relatable and the further into the future something takes place, such as using Connectivity data, there is a tendency to lose this balanced view and become overly positive? Visualize data: Data visualization is seen as being a key driver of being able to understand current state vs ideal state clearer, i.e., see the gap (CP10, CP12, CP13, CP14, CP16 and CP22). This will highlight abnormalities and hence, improvement opportunities will be more obvious which will benefit problem-sensitivity. This data might today be limited to front line shop floor operators but with the visualization of the collected data this will be visible to a larger population. Visualization will support trend analysis (CP11 and CP20) which might be useful in catching a deviation before a control limit is reached, i.e., enable proactive action before it becomes a problem which supports Jidoka. The benefits of proactive actions is something which is echoed by Kolberg & Zühlke (2015) as well as by Mrugalska & Wyrwicka (2017). Better visualization is also believed to enable deeper analysis since less time is spent on collecting and understanding the data and hence more time can be dedicated to the problem analysis. This will enhance the understanding so that deeper conclusions can be drawn (CP16 and CP17), i.e., better problem-solving. Another way to see this is that better visualization will help to turn data into facts (CP15 and CP21). Standardized daily board meetings with data automatically updated and visualized will also free up time in the daily meetings, time which can then be used to motivate and engage the work force (CP23). This will potentially change the way that shop floor operators see the data collected (CP24) and enhance the operator’s possibility in understanding and acting on the data (CP25). Bottlenecks becoming more visual was also pointed out as a potential where actions can then be addressed in managing the bottlenecks as well as actively work with the bottlenecks to reduce them and their impact (CP18 and CP19). Use data: The use of data stemming from Connectivity is discussed in more detail in section 5.1.3. It is clear from the responses that many see the use of Connectivity data as the pre-requisite for the self-managing and self-optimizing smart factories (CP26 and CP36). AI and ML is managing and optimizing the smart factory using the data collected, eliminating potential operator errors (CP29), and in general support the continuous production flow and JIT, as mentioned by Mora et al. (2017). The use of data is also believed to result in faster problem-solving were the collected and visualized data will enable better understanding of different variables in the process flow and how they are causing variation (CP28 and CP30). Building on this, it is also believed that correlations of different variables will be easier to see (problem-sensitivity) and more data driven and hence root cause analysis and problem-solving will be better and faster. It can also enable problem-solving of issues which today are un-solvable such as non-linear process parameter correlations between different

53

process steps (CP27 and CP31). Overall, Connectivity is hence contributing significantly to both seeing problems and problem-solving. A further benefit of Connectivity is potentially that the lead time in evaluating and implementing process changes will be reduced as better data is available faster along with trend analysis (CP37). Improved shop floor operator improvement suggestion scheme (CP33 and CP34) was also something which was pointed out as use of Connectivity. Positive contributions to occupational health and safety due to the array of sensors in the manufacturing environment was also highlighted (CP35). The use of data will most likely also give birth to new business models where the data mining in one location or part of a factory can be shared across the whole enterprise or even further with customers and suppliers (CP32). One example of this could be plastic injection plant sharing process parameters and output results (e.g. cycle times and quality yield) with the external plastic injection machine supplier. This supplier can then propose more frequent parameter improvements to improve productivity and reduce cost (e.g. shorter cycle time and improved quality yield). This can then further be rapidly shared with other plastic injection plants which the equipment supplier supply.

5.1.2. Connectivity Challenges and Risks

Challenges: A challenge relating to Connectivity is how to make sure that the data is collected in a way that the shop floor operators will benefit from it (CC01) and hence remain engaged (CC05). It is especially important that the shop floor CI activities can be driven or guided by the data (CC02). The saturation of data is also a challenge where “old fashion” problem-solving might be more effective (CC06). The competence of the shop floor operators is highlighted as a challenge which is common across all three I4.0 value drivers (CC09) (see 5.3.3 for Competence related success factors). During the implementation phase many Challenges are identified, such as too fast implementation resulting in that more data is made available than what the organization have resources of competence to act on (CC03). It was also pointed out that the rapid access to data, even from remote locations, might lead to less go and see on the shop floor and hence many relevant inputs might be overlooked (CC08). During the implementation phase, initially only parts of the production flow might be connected and hence a complete production flow might be overlooked (CC04). Old production equipment which is lacking modern information technology interfaces might be a challenge and used as an excuse to not pursue Connectivity (CC07). Risks: Data saturation as a result of excessive data collection (CR01) and visualization (CR02) would become a Risk if not under control. Having too much data and using it in a non-productive way is something which Romeo et al. (2018) also identified as a Risk in their research relating to increased digital waste from connected systems (Romero, et al., 2018). Collecting data which is not used will add cost due to the cost of collection and storage, but more importantly, the useful parts of the data might be lost in the overwhelming amount of data which can cause incorrect interpretation and hence missed or even incorrect insights (CR03). An increased amount of sensors etc. will also increase the maintenance cost of the production process, both professional maintenance and autonomous operator maintenance, which is something which Rüttimann et al. (2016) also highlights.

54

The way the collected data is used is also a potential Risk as seen by some of the respondents were the increased speed and transparency of data is used to create a blame culture rather than using it for CI and to motivate and engage the shop floor operators (CR02). In a smart factory, where each operator “badge-in” on the specific station it becomes easier to track and monitor individual performance. Any slightest tendency to use this information incorrectly is likely to have a quick counter reaction from individual shop floor operators as well as trade unions etc. The Lean core value of Respect for People is an important pre-requisite to prevent any such tendency. There is a fear among the respondents that the shop floor environment will be more static due to Connectivity since even small changes might result in sensors having to be moved, re-calibrated and perhaps even involve software updates and changes of user interface (CR05). This new and additional complexity will also require additional competence and skills among the shop floor operators. To overcome this lack of skills, initially the IT department can support but their resources will be limited and the barrier of having to ask for support, might lead to a slower CI rate (CR07). There is further a Risk that the additional cost due to the higher complexity will require stronger motivation such as Benefit/Cost (B/C) ratio of proposed changes and it might also mean that there will be less trial and error activity on the shop floor. The worst-case scenario would be if external support would be needed where the organization gets locked up with a dependence of external expert support which is seen as to further aggravate the reduction in CI rate. If the shop floor operators become dependent on either internal or external support then there is a Risk that the CI rate further drops due to a lack of engagement and involvement when “everything is anyway always too complicated, and the B/C ratio is never good enough” (CR04). A significant investment in Connectivity in the existing production equipment is also believed to result in a Risk that the sunken cost is too high and that the organization prefer to keep existing production process and make incremental changes rather than investing in new breakthrough technology, “difficult to kill your darlings” (CR08).

5.1.3. Intelligence Possibilities

Transparency: According to the participants, AI and AA have the potential to increase “perception” by complimenting human Intelligence and contributing to insight into previously unknown problems as well as to their potential solutions (IP01-02). The increased perception made possible by AA can help manufacturers to investigate and hence, improve areas of manufacturing that was not possible before (IP03). In essence, AI and machine learning can help manufacturers know what they do not know today (IP04). Moreover, by using AI and machine learning to find advanced correlations and utilize non-linear optimization, manufacturers can enhance their understanding of how to reduce problems and increase efficiency, providing pathways to more effective value streams and shorter lead-times (IP05-08). The ability to predict events and to see the unknown can accelerate CI-driven solutions within the manufacturing environment, speeding up the problem-detection and problem-solving processes (IP09-12); potentially predicting outcomes with a high degree of accuracy and thus preventing, for example, quality issues before they occur (IP13). This is synonymous to “waste-signaling” as outlined by Powell et al. (2018). “Perception” enabled by the Intelligence value driver is arguably at the very essence of CI as it both enhances an organization’s problem-sensitivity and decreases the time for problem-detection and root-cause analysis by providing data-driven insights.

55

For example, Jidoka practices can be directly benefited by event-prediction which allows for the problem to be avoided as the operator can make necessary adjustments before the machine needs to be shut-down. Similar arguments are provided by Lorenz et al. (2019) who argue that the built-in quality can be improved by increased process transparency following the application of AA for problem detection and prediction (Lorenz, et al., 2019). Perception is arguably the core-theme of the Intelligence value driver as it has the potential to provide insight into physical and digital processes; insights that was not possible before (IP14-15).

Decision-support: Derived from an increased perception, AA and AI can provide the foundation necessary to make fast and accurate decisions; implying that CI can be accelerated as the focus is aimed towards the most value-generative activities and problem-solving processes (IP16-18). In the long run, AI has the potential to further facilitate decision support (IP19), enabling faster decision making and decentralized control, decreasing the need of management approval (IP20). This follow the line of arguments outlined by Mayr, et al., (2018) that decentralized control can facilitate flexibility and productivity (Mayr, et al., 2018). In essence, AA and AI decision-support tools simplifies root-cause analysis and accelerates the CI-work by making the problem-solving phase easier (IP21). Decision support also facilitates waste-reduction and optimization of the process which enhances CI (IP22).

Autonomy: Increased perception and autonomous decision-making arguably sets the foundation for autonomy in the manufacturing environment. As AI gradually compliment and eventually replace human Intelligence, the need for human intervention decreases. This subsequently decreases the process variability (Waste – defects and rework) as human trial and error process can be avoided and as human faults such as forgetfulness and loss of concentration no longer interfere with the manufacturing process (IP23-25). Rapid fulfilment of tasks with high quality outcomes can become a reality as humans are gradually replaced (IP26); decreasing the cost of manufacturing (IP27). In the long-run, the CI-work might partly or solely be performed by machines as built in CI-solutions driven by machine learning can become a reality (IP28). This can potentially enable intelligent self-improved manufacturing systems that do not need human intervention in order to detect, nor solve problems (IP29). Moreover, increasing support from intelligent systems can also reduce mental and physical overburden (Muri) by increasing design commonalities and effectiveness (IP30). Furthermore, AI can provide objective solutions which drives standard ways of working (IP31), essential to CI. The above arguments suggest that the intelligent environment can increase an organizations’ agility, facilitating quicker responses and problem-solving which might be beneficial for the overall CI-conditions. Self-managed systems, enabled by IOT, CPS and AI combined might decrease the overall need for human labor as they make the manufacturing environment more autonomous. This is contrary to the arguments by Schneider (2018), who bypasses the “dark factory” perspective and instead suggest that the need of supervision and human-machine interaction will increase with I4.0. However, whether a dark-factory is beneficial for the CI-environment is left to be proven. Nevertheless, increased autonomy could arguably yield more time for humans to work with CI within areas where technologies are not beneficial.

56

5.1.4. Intelligence Challenges and Risks

Challenges: Several Challenges associated with Intelligence can be derived from its vague definition and the generally low trust and reliability of intelligent systems (IC01-02). Exactly what Intelligence implies in practice and what value it will generate for the manufacturing environment is difficult to pinpoint (IC03-04). Hence, potential improvement activities can be lost due to difficulties in showing the value of data-driven decisions stemming from AI solutions. The utopic reality of self-managed systems is far from being realized. Therefore, Intelligence; wherever it is introduced, will arguably have to be supervised. Supervision requires well understood scenarios and clear targets and any deviations in relations to those targets need to be followed up and resolved (IC05); which arguably poses additional Challenges for manufacturers. As argued by Lodgaard et al. (2016), the lack of managerial supervision of employees constitute significant barriers towards sustaining CI (Lodgaard, et al., 2016), and intelligent systems is no exception in this regard. Furthermore, in order to realize the full potential of Intelligence, data need to be shared across borders (also known as open innovation), which requires adequate data security as well as organizational integrity (IC06). Making a manufacturing environment intelligent and realizing its true potential, also requires specific training and education, which stresses the demand for engineering experts (IC07-08). Additionally, following increased autonomy and decentralized decision making, management will need to change their mindset about CI as activities are carried out unsupervised by self-managed systems (IC09). As known from literature, difficulties in breaking old mindsets and subsequently embracing new ones are argued as being some of the main barriers towards CI (Caffyn, 1999; Bhuiyan & Baghel, 2005).

Risk: AA and AI cannot be used without data and the quality of the data is of utmost importance for the end-results (IR01). Feeding intelligent systems support with faulty and inadequate data Risk creating fragile and unsteady systems which can have direct negative impact on the built-in quality (Jidoka). Moreover, error in prediction and forecasting stemming from feeding decision support tools with low quality data can blur perception, deceiving decision-makers into making the wrong kind of decisions. This could result in the creation of problems rather than to contribute to the ability to solve them. While unfiltered data can clearly obstruct the problem-solving phase and bring frustration into decision making (IR02), the question of whether the machine will be right in its prediction, even with adequate data feeding, is left unanswered (IR03). There is also Risks that AI is used to confirm what is already known (IR04), increasing cost but yielding no additional benefits to the process of problem detection and problem-solving. Additionally, discussions concerning the danger associated with AI cannot be unheard; rapid growth within this area can be detrimental if the proper implementation and reflection is overlooked (IR05), not just to the conditions for CI, but for the whole organization. An all increasing dynamic environment requires the ability to quickly adjust and to respond to external changes; and although Intelligence can facilitate decision-making by forecasting and increased insight, self-managed systems that incorporates organizational change and learning are far from being realized. As a result, there is a Risk that intelligent machines (at least in its infancy) need to be constantly retrained as new data emerges (IR06); creating a time-lag between problem identification and problem-solving. Further Risks include Intelligence becomes a goal instead of a means (IR07); directing time and resources towards efforts of achieving Intelligence rather than to use it for the purpose of continuously improving which decreases the problem-solving capabilities. Moreover, as data is shared across borders, externally and internally, the security risks

57

being compromised as data might find its way to unauthorized personnel (IR08). Hence, to facilitate adequate data-security the systems might become more difficult to access and interpret, which could potentially impede improvement efforts. Moreover, as AA and AI become a more integrated part of production, “experts” will be needed to perform even small changes (IR09). This Risk making intelligent system more rigid as specific training and education are needed to improve them. There is also a Risk that companies need to outsource improvement activities to the suppliers of intelligent systems and to consultants specialized in intelligent solutions. This can arguably decrease the overall problem-sensitivity as ownership of the CI-process is partly lost, and the benefits are instead concentrated to a selected few and high-tech companies (IR10). Increased autonomy also removes the operator from the improvement work which Risks affecting CI-work negatively (IR11), considering humans are an essential part in identifying and conducting CI-activities. Moreover, autonomous systems can also result in less involvement from management, negatively affecting Go and See which overall decreases the problem-sensitivity as human involvement is decreased. Less involvement and support form management following increased autonomy have been previously argued by several authors to negatively affect the conditions for CI (Lodgaard, et al., 2016; Garcia-Sabater & Marin-Garcia, 2011; McLean, et al., 2015).

5.1.5. Flexible automation Possibilities

Flexibility: In relation to today’s semi-automated systems, Flexible automation through advanced robotics, offers increased adaptability and faster and more efficient production enabled by quicker changeover times and intelligent reactions (FP01). Such a flexible configuration of the manufacturing environment facilitates mass customization and single piece flow (FP02); design opportunities (FP03); and JIT-deliveries (FP04). Moreover, the manufacturing can take place closer to the customer or at the point of need which reduces unnecessary waste in the form of unnecessary transportations (FP05). Moreover, flexible responses to the dynamics of the environment can potentially open up new CI-opportunities not visible by today’s more rigid manufacturing systems (FP06). It also contributes to the problem-solving capabilities by decreasing the CI-cycle by faster proof of concepts, testing of improvement ideas, faster production of prototypes and simulation of tooling (FP07-11). Moreover, AM could potentially open up the possibilities to work with quality and cost-reductions where not previously possible as it brings ownership over the complete assembly by decreasing dependency on external suppliers (FP12). It can also secure the optimal product solution without sub-optimizing solutions for each manufacturing step (FP13). In summary, flexibility is a target image of the smart factory as it allows for flexible response to changing environments and customer needs. This is partly realized as advanced robotics are integrated with Intelligence as argued by Powell et al. (2018) to be a key factor in facilitating quick production adjustments (Powell, et al., 2018).

Automation: First off, it should be noted that automation, per say, is not a new phenomenon. However, with the dawn of new technologies, such as advanced robotics that offers enhanced perception, integrability, adaptability and mobility; automation can be reinforced and introduced where previously not possible. Especially in high volume production, advanced robotics can reduce manufacturing variability and eliminate waste as the process will become more easily observed and repeatable (FP14). Moreover, when variability cannot be further reduced by automation, cobots can

58

help manufacturers to “pick up slack” (FP15). Elimination of unnecessary movement by transportation done by autonomous vehicles are other potential benefits that could benefit waste reductions (FP16). In addition, the mental and physical overburden of operators can be significantly reduced by the automation of heavy physical repetitive tasks as well as non-intellectual tasks that were not viable or possible to automate before (FP17-18). Moreover, as repetitive and non-intellectual tasks are further automated, shop floor operators receive more time for CI-work which can enhance the problem-solving capabilities as well as empowering operators and low-level managers (FP20-21). The technologies associated with Flexible automation have the potential to directly decrease cost, improve quality as well as shop floor operator safety (FP22-23). New CI-opportunities can also be found as AM changes that way a product can be manufactured as well as decrease the time from idea to implementation by faster prototyping (FP24).

5.1.6. Flexible automation Challenges and Risks

Challenges: Advanced robotics and additive manufacturing will increase the complexity in understanding why something happens as non-linear relationships are introduced into the manufacturing environment (FC01). This could lead to unknown Challenges that will impede CI-efforts and learning (FC02). Thus, when problems arise (not if they arise), it might be inherently difficult to solve them. Other Challenges of Flexible automation include finding suitable applications in low-volume and low-manual labor environments commonly found in the aerospace sector (FC03). Cobots have been investigated for years, by no clear applications has been found in such environments. Although increased problem-solving capabilities and perception through Intelligence might facilitate problem-sensitivity and problem-solving capabilities; to see, understand and solve problems might prove to become very challenging following the introduction of Flexible automation. Thus, operators and assemblers need to be involved earlier in the CI-cycle; analyzing, modeling and planning in order to solve occurring problems (FC04). Moreover, to reach the full potential of technologies such as 3D-printing (especially metal 3D-printing), extensive qualification and verification need to be performed in order to guarantee sufficient quality of manufactured products. This time-consuming process is not only challenging itself but might temporarily tie up resources and capital which can impede CI-efforts (FC05).

Establishing and sustaining standard ways of working in an increasingly technologically complex environment will challenge adopters of Flexible automation (FC06). Thus, before the best-known way to conduct work is established, it might prove difficult to identify CI-activities as well as to challenge the status quo. Moreover, implementing advanced automation is costly, and its integration must align with business and operational strategy in order to provide clear return on investments. As of today, few companies have been capable of providing clear returns on investments of advanced automation technologies (FC07). Furthermore, difficulty in setting clear targets and KPIs for novel technologies might prove challenging for manufacturers to establish (FC08). These Challenges are particularly important to overcome since KPIs are needed for conducting successful performance management as well as to continuously review and improve the technologies (FC08).

Risks: Flexible automation might magnify existent inefficiencies if introduced into a poorly designed factory layout and production line configuration (FR01). This argument follows the statement of Bill

59

gates’ second rule of technology used in business, implying that ”...automation applied to an inefficient operation will magnify the inefficiency” (Buer, et al., 2018, p. 2934). Although, such a perspective depicts automation as potentially detrimental if applied to inefficient processes; automation can still increase efficiency by decreasing or removing waste. It would also be unrealistic to suggest that all inefficient processes are to be made efficient prior to automation; particularly as the question; “when is the process efficient enough?” is unanswerable. Nevertheless, the Risk of magnifying different types of wastes as a result of inefficient and “wasteful processes” being automated, is imminent. For example, implementing costly and complex advanced robotics systems in an inefficient assembly for the purpose of decrease the number of manual steps that could have otherwise been reduced or removed, will only make those “wasteful” steps completed faster. Hence, it is important to remember that: automation does not remove a task, but instead removes the human element necessary for its completion. Thus, if the process is characterized by unnecessary movements, defects and reworks and over-processing; the automation will only magnify those wastes. It is therefore of utmost importance that the technology solves a well-defined problem before being implemented (FR02). “Riding the hype” and implementing technology solutions were simpler and cheaper solutions could have worked, is always a Risk if not adequately assessing the different alternatives and the technology impacts beforehand. Furthermore, a wide adoption of advanced robotics, cobots and AM will not eliminate process variation or wastes in the production flow (FR03), but at their best, reduce them. There is also a Risk that the rigidity of the systems offset the benefits from the increased flexibility, e.g., as the manufacturing will become more engineering developed and therefore dependent, improvements can be more difficult to implement (FR04). If the purchased systems do not have built-in improvement possibilities, rigidity might further offset the benefits of Flexible automation (F05). Following the line of reasoning in section 5.1.4, there is a Risk that CI need to be outsourced as activities are too difficult and complex to perform in-house due to built-in rigidity and lack of competence. Outsourcing of improvement activities decreases the ownership of the process and negatively affect both the problem-sensitivity as well as the problem-solving capabilities. Furthermore, less involvement of manual shop floor operators due to automation and outsourcing reduce the team member’s ability to perform and to see potential improvements, negatively affecting the rate of CI (CF06).

5.2. Discussion summary first iteration

Most of the answers provided for all value drivers have been positive, 69% for Connectivity, 61% for Intelligence and 63% for Flexible automation; indicative of an overall positively weighted belief of the impacts of I4.0 on the conditions for CI. While the results are arguably a reflection of the current I4.0 hype, they also highlight the overall difficulty in critically assessing the potential impacts from technologies that have not yet made their way into manufacturing (at least not beyond the piloting scale into large-scale rollout). Moreover, in combination with I4.0’s generally vague definition and its announcement before its realization, industry and academia tend to customize its definition as to constitute a salvation from the current limitations and problems associated with today’s manufacturing systems, which further adds to the I4.0 hype.

Nevertheless, the participants have identified several Challenges and Risks for each value driver. Challenges that need to be overcome in order to reach the direct step-improvement and

60

improvement rate potential post I4.0 implementation, and Risks that can negatively affect the rate of improvement post-implementation. Table 20 shows a summary of the themes identified for each I4.0 value driver and its constituting Possibilities, Challenges and Risks. The themes are described in more detail in section 4.1.1, section 4.1.2 and section 4.1.3.

Table 20: Summary of themes of Possibilities, Challenges and Risks derived from the value drivers’ impact on the conditions for CI.

Connectivity Intelligence Flexible automation

Possibilities 69% Collect data Visualize data Use data

61% Transparency Decision support Autonomy

63% Flexibility Automation

Challenges 17% Value adding Engagement &

Change 17%

Trust & Understanding

Education & Training

21% Knowledge Strategy &

Standards

Risks 14% Digital waste Rigidity 22%

Data quality & Application

Ownership & Participation

16% Excessive

solutions Rigidity

As have been derived from the above discussion and results, the introduction of I4.0 offers a wide range of possibilities for the CI-conditions by positively impacting problem-sensitivity and problem-solving capabilities. However, not without accompanying certain Challenges and Risks. Figure 16 illustrates a discretized improvement over time diagram, highlighting the potential impacts of the Possibilities and the Risks. Figure 15 (A) shows the “realized step-improvement” versus the “potential step-improvement” and Figure 15 (B) shows different rates of improvement post I4.0 technology implementation.

As described in section 3.4, Challenges are factors that need to be overcome in order to either realize the improvement potential or to mitigate the risks. Therefore, the challenge itself neither decreases nor increases the rate of improvement, but Instead, the resulting increase or decrease in the rate of improvement following a challenge is related to the realization of the accompanying possibility or risk. For example, overcoming Challenges related to, e.g., providing necessary education and training to operators or establishing trust and understanding for I4.0 technologies and its solutions might help adopters realize the technology potential while minimizing the risk impacts.

Risks can, if not mitigated, decrease the rate of CI post technology diffusion by negatively impacting problem-sensitivity and problem-solving capabilities. This might result in a decrease of the frequency of improvements as well as lower levels of improvements per conducted activity. For example, lower degrees of Ownership & Participation following outsourcing of improvement activities and decentralization of decision making; and increased Rigidity, derived from increased cost and difficulty of changing systems might decrease the rate of improvement. Such a scenario is illustrated in Figure 16 (B, red). Therefore, it is important for companies to assess whether the potential step-improvement following I4.0 implementation as seen in Figure 15 (A), risks being offset by the increase in inertia tendencies, a scenario illustrated in Figure 1, section 1.2. Thus, in order to reach the full potential of I4.0 without negatively affecting the conditions for CI, companies should

61

focus on overcoming Challenges and to mitigate the Risks instead of blindly focusing on the mere possibilities.

Figure 16: Example of Possibilities’ (green) and Risks’ (red) impact on the conditions for CI.

R1 concerns how the introduction of I4.0 affects the conditions for CI with regards to rate of improvement, and R2 concerns the potential factors which can decrease the rate of improvement following an increased Connected, Intelligent and Flexibly automated manufacturing environment. Based on the above discussion, it can be concluded that I4.0 implementation through its main three value drivers: Connectivity, Intelligence and Flexible automation, provides numerous Possibilities for CI; Possibilities that can increase the rate of improvement post I4.0 implementation by positively affecting the conditions, i.e., the problem-sensitivity and problem-solving capabilities. However, the introduction of the I4.0 value drivers also accompany Challenges and Risks as discussed extensively above. Challenges that require efforts, resources and capital in order to be overcome and Risks that, if realized, decrease the rate of improvement.

5.3. Success factors

The below sections summarize and discusses the findings of the second iterations of this study which focus on success factors around the three different value drivers of I4.0 (Connectivity, Intelligence and Flexible automation) with regards to CI. These success factors can be found in Table 16 to Table 19 in Chapter 4.

5.3.1. Purpose

The category “Purpose” deals with the success factors relating to “why” an organization should embark on an I4.0 transformation journey and how it relates to their CI efforts. To understand the value of I4.0 up-front (CS02), to make sure that real business needs are addressed and that the benefit and value of addressing those business needs are known (CS01, IS02, FS02), are all seen as important factors. By just connecting everything that can be connected and starting to collect random data will not benefit anyone (CS03) and it will be detrimental due to the cost of doing so (hardware, software, storage etc.). Connecting everything with also add additional complexity

62

(maintenance etc.) and potential alienation of the shop floor operators when their needs are not addressed and when capital is wasted rather than driving their CI-ideas. Knowing the value that I4.0 will bring (IS03) means that the implementation will address a true business need and not just drive implementation for the sake of implementation. A good example is one of the Lighthouse respondents, where the I4.0 solutions implemented were driven by the broad range of part numbers where manual management would have made the business unviable. Equally important to understanding the need, is having clear targets of what the implementation should provide or accomplish (IS01). With a clear target the gap before and after implementation can be seen, success can be judged and a clear visualization between current state and ideal state after implementation can be supported (FS03). When consolidating the Purpose of an I4.0 implementation, it is highlighted that preparation activities should involve all stakeholders and especially the shop floor operators so that they understand the benefits for the business and for themselves (CS03) and that future CI activities are not negatively impacted when some people might feel that they were not involved in the implementation (FS01). This will be further discussed under “Involvement of people”.

5.3.2. Involvement of people

To make sure that CI remains successful even after an I4.0 implementation it is critical to involve the most important resource – People. Successful I4.0 is very much like successful change management and involvement of all stakeholders and especially shop floor operators who will be facing and using this new technology every day is crucial (CS04, CS07, CS08, CS10, IS04-IS08, FS04, FS05). The best knowledge is often found with the shop floor operators (IS04) and by involving them in the decision making, ownership is built (IS07) which support future CI activities (IS05). Involving the shop floor operators will also make sure that the implemented solutions are logical (IS08) and that the data is correctly selected and classified (SC07) as well as aligning everyone (CS04, FS05). This is best done through open minded discussions on the shop floor (CS05, CS08). During commissioning and after implementation it is important to have an open mind so that the data is challenged and confirmed (CS05, CS09, CS10) in collaboration with shop floor operators. If the preparation work prior to and during implementation was done correctly, there should be good ownership that allows an open discussion as to clarify correlations (CS10). Doing this will ensure that the data and knowledge extracted thanks to I4.0 can be used to solve problems rather than only confirming things already known (IS09). Good change management and involvement of people will further support this use of the capabilities of I4.0 as trust have been established in the technology (IS08). To ensure that the trust in the technology and ownership which have been built up and is sustained during daily operations, it is important to give authority to the shop floor operators to act on the data collected and the decision support provided (CS06).

5.3.3. Competence

Employees’ competences will be an even more important part in the process of identifying and solving problems in the future I4.0 factory. Specific education and training will be critical to engage

63

shop floor operators (CS11) and specific CI-training based on Connectivity needs to be implemented (CS12). It is therefore important that manufacturers have a clear strategy for competence development of both managers and shop floor operators prior to the implementation (CS13). Specifically, training in variation and statistics should be provided in order to support data analysis and to create understanding of process capabilities as Connectivity increases the flow of data (CS14). To facilitate CI following increased Connectivity, companies need to provide opportunities for employees to equip themselves with an understanding of correlations between process input data; causal relationships between output and input data as well as a solid process understanding (CS15-17). As for Intelligence and Flexible automation, skills particularly centered around data science need to be fostered within the organization to establish ownership by preventing outsourcing and external dependency when conducting CI (IS10, FS06). An alternative is to make data scientists work closer to manufacturing to facilitate knowledge transfer and problem-sensitivity (IS11). Competence developments needs to be considered before, during and after the implementation on both managerial and employee level (IS12). Concerning the pilot projects, the teams should be kept small with members that are open-minded and that understands the vision of Intelligence in order to facilitate effective change management and to provide AI and AA solutions that can benefit CI (IS13). Multi-skilled teams, with competences ranging from process experts, data scientist and shop floor operators need to be assigned to work specifically with the integration of Intelligence and Flexible automation to include functions beneficial to everyday CI-work (FS08). It is also important to align IT-personnel with the CI-work to facilitate a closer cooperation between manufacturing and IT (IS14).

5.3.4. Implementation strategy

Constituting the foundation for the successful integration of the other value drivers, Connectivity is of outmost importance. Thus, Connectivity needs to be implemented in such a way as to ensure an adequate data reliability and validity. In order to minimize its integration Risk, a step-by-step approach should be embraced (CS18), where parts of the manufacturing being connected can be thoroughly assessed as benefits and downsides can be measured prior to further implementation. However, a clear implementation strategy is also needed that includes, not only I4.0 piloting and a step-by-step approach, but also its large-scale roll-out (CS19). According to a report published by McKinsey & Company, a success factor for escaping pilot-purgatory is: “establish a clear vision for Digital Manufacturing and a phased road map to get there” (McKinsey & Company, 2018; World Economic Forum, 2019). This adds to the weight of the above arguments regarding the importance of having clear strategies and road maps to reach a successful I4.0 implementation. Moreover, as for all technology integrations, the implementation should be founded on a solid business case that includes the implementation costs as well as the cost of competence development needed to sustain CI (CS20-21). Proper data-visualization and analyze tools need to be developed prior to roll-out in order to utilize the complete value of the collected data and to ensure its reliability and visualization (CS22-24). While Connectivity should also be integrated to allow for data collection across the complete value-chain, from supplier to customer (CS25), care must be exercised to not overload systems with excess data not generating value (CS26). Thus, the data-quality, sampling speed and frequency should be continuously reviewed (CS27), and a reliable data-connection needs to be

64

ensured (CS28-29). It is of outmost importance that the data collection is stable and that the integrity of the data is validated and cross-checked in regard to its criticality, impact and cost (CS30-31). Moreover, when increased data collection stresses the bandwidth5, edge computing6 could be prioritized over cloud computing7 as to decrease data loss and system lag (CS32). As a general guideline, technology that is flexible and easily adaptable shall be prioritized before more complex solutions (CS33-34). IT and OT shall also be aligned as any discrepancies can cause malfunctions and unwanted results (CS35). This further stress the importance of a careful step-by-step integration of Connectivity as to ensure an adequate input to the OT governing the process that are to benefit from an increase flow of information. It is further important to integrate a common system for data collection and processing in order to facilitate a better overview and reaction lead-time (CS36-37); while ensuring data safety by using separate networks and storages for different types of data (CS38).

As for the Connectivity value driver, Intelligence need to be implemented step-by-step following a gradual roll-out and guided by a digital strategy that allows for sufficient time to learn from mistakes and grow (IS15). A suggestion is to use the Deming cycle, used for carrying out changes of processes and products were activities are performed in an iterative manner – Plan, Do, Check and Act (PDCA). PDCA is a well-used Lean tool for carrying out CI (Skhmot, 2017), which goes to say that not only Lean values and principles are of importance when integrating I4.0 (Leyh, et al., 2017; Rossini, et al., 2019), but Lean methods and tools might also come in handy during its implementation. Utilizing the Lean-toolbox to fully understand the value-stream and its processes is also pointed out as a success factor, necessary before implementing AI-based solutions (IS16). This follows the line of arguments provided by Sony (2018), who argued that methods such as value-steam mapping can facilitate the I4.0 integration process (Sony, 2018). Considering the many Risks associated with the improper introduction of Intelligence in the manufacturing environment (see section 4.1.2); it is important to start small, using piloting projects and by establishing processes by which AA and AI can be reviewed, validated and verified in order to minimize any negative CI-impacts (IS17-18). It is therefore important to establish Standardized Work and response scenarios to be able to review, validate and verify the system as well as to be able to detect problems (SF19). Furthermore, any problems that are to be solved using Intelligence needs to be treated holistically, integrating all aspects of people, process and technology (IS20). This is particularly important in order to gain trust and momentum for its implementation and to minimize the negative outcomes on the conditions for CI. In regard to the aspects of people (as observed from Table 1 and discussed in section 2.2) managerial and employee factors are of vital importance for the success of CI. Lack of managerial support, supervision and commitment as well as employee commitment and willingness to change; all constitute significant barriers to CI (Lodgaard, et al., 2016; Garcia-Sabater & Marin-Garcia, 2011; McLean, et al., 2015). Focusing on applying Intelligence on complex processes rather than simple and repetitive ones can keep some of the flexibility (IS21). Thus; therefore, complex processes that are “outside” of the 5 Bandwidth: Describes the maximum data transfer rate of a network or internet connection, i.e., how much data that can flow through a specific connection at given time (Christersson, 2012). 6 Edge computing: Computing that is done at the “edge”, i.e., in close proximity of the source of the data rather than relying on the “computer cloud” of one of several data centers (Miller, 2018). 7 Cloud computing: Computing that is done remotely over the internet, often in a commercial providers’ data center (Knorr, 2018).

65

current AI-capabilities need to be thoroughly reviewed in order to understand how to integrate them with Intelligence (IS22). Before the implementation of Flexible automation, a successful CI-program based on the Lean values and principles as outlined in section 2.1 as well as methods and tools, should be established (FS08). However, this is not a success factor only for the Flexible automation value driver but can arguably be applied on both Intelligence as well as Connectivity. This success factor is supported by several authors, who argue that Lean facilitates the I4.0 implementation process (Leyh, et al., 2017; Rossini, et al., 2019; Tortorella, et al., 2019; Sony, 2018; Agostini & Filippini, 2019). Following the same line of arguments, it is important to not automate an already “bad” process (FS09). Thus, before integrating Flexible automation, wasteful processes need to be removed and ”effectiveness through simplicity” shall be set as a primary focus by embracing Lean values and principles prior to integration. To show respect for people and to decrease overburden, the focus should be to automate the “heavy” and “dirty” processes as requested by the shop floor operators instead of focusing on the most beneficial processes from a monetary perspective (FS10). Large volume processes and bad ergonomic operations are often first candidates for Flexible automation; however, if simpler solutions could provide the same results as if automating the process, these should be prioritized (FS11). It is important to define what flexibility truly implies for the manufacturing environment by scoping its possibilities and limitations for each case (FS12) and to understand how the lead-time is affected (FS13). Such a thorough assessment can mitigate the Risks concerning, e.g., Rigidity that might offsets the flexibility of automated systems. Moreover, as for the other value drivers, standardized processes shall be established to drive improvement activities (FS14).

5.4. Discussion summary second iteration

The below paragraphs will attempt to summarize the findings around success factors as broken down in the four categories Purpose, Involvement of people, Competence and Implementation Strategy; starting with a more generic discussion. The main takeaways are bolded for the readers’ convenience. The second iteration of the survey presented an executive summary of the findings of the first iteration to the participants (see Appendix E). The summary showed that the overall response was positive but highlighted the possibility that CI might be negatively impacted by I4.0 implementation. The survey questions in the second iteration then went on to ask for input on success factors for I4.0 implementation with regards to CI, i.e., factors that should be considered during an I4.0 implementation as to ensure that CI is not negatively impacted. The data collected in the second iteration has been post processed similar to the data from the first iteration as described in section 3.4. In total, 74 text strings have been extracted, 38 of these regarding Connectivity, 22 regarding Intelligence and 14 regarding Flexible automation. These have been grouped into four different categories: Purpose, Involvement of people, Competence and Implementation Strategy. The details of this breakdown can be seen in Figure 15. When post processing the data from the second iteration, it has been noticed that some of the success factors mentioned were not always directly relating to the success of CI during and after an I4.0 implementation, but some could rather be considered as

66

generic success factors for an I4.0 implementation journey. Both goes to some extent however, hand in hand, because if the I4.0 implementation turns out unsuccessful, it is also likely that the Lean and CI implementation journey will be less successful. It shall be noted that, success factors are themselves, closely related to the presented Challenges (Table 10, Table 12 and Table 14, Chapter 4) as some also need effort, time and resources to be overcome. However, the success factors have been defined as guiding principles that can help to minimize the negatives from I4.0 implementation by overcoming Challenges and to mitigate the Risks. Nevertheless, some success factors can arguably be seen as Challenges and vice versa. The category Purpose reinforces that any implementation of any tool regardless of I4.0 and CI should always have a purpose. The implementation itself of the tool cannot be the purpose but rather the value that the tool brings should be the purpose. There is a considerable hype around I4.0 and a clear Risk that executives will suffer from “fear of missing out” and initiate I4.0 implementation without a clear understanding of the value and specially to fail to share this understanding with everyone involved, including the shop floor operators. Some of the lighthouse companies are almost not realizing that they are implementing I4.0 as they are so focused on addressing their business needs and it is only when they step back that they are realizing that they have been implementing I4.0. To have a clear understanding of the purpose is an important prerequisite when implementing I4.0 as well as to clearly understand how the implementation will impact CI. The second category, Involvement of people, highlights the importance of involving people across the business and value chain in the preparation, implementation and usage of the I4.0 value drivers; again, both for generic success in implementing I4.0 as well as ensuring successful CI activity after implementation. The shop floor operators know the production equipment better than anyone else, they know what happens and which data that might be interesting to collect and use, failure to involve them may therefore result in sub-optimal implementation as well as reduced CI-activity in the post implementation environment. Basic things like ownership will suffer if new tools are pushed out in the manufacturing environment without proper communication, anchoring and collaboration. The third category, Competence, is highlighted as a key category both before, during and after an I4.0 implementation as to ensure positive impact on CI. To be able to effectively conduct the communication, anchoring and collaboration, the competence level across the whole company needs to be improved so that everyone, and especially the shop floor operators, can be engaged in the discussions concerning CI-activities. To only use external competence would be detrimental to CI-activities in the post implementation environment and it would also reduce the effectiveness of the implemented tools. If the shop floor operators would have to solicit external support or even internal IT support for any small improvement that they want to implement, then the inertia will be high, and the improvement rate Risks being reduced. The fourth category, Implementation strategy, discusses more pragmatic aspects relating to an I4.0 implementation journey in general and its relation to CI specifically. From the success factors collected it quickly becomes clear that successful Lean implementation will increase the chance of success due to several reasons. A manufacturing environment where Lean has been applied since

67

several years tend to be more stable and repeatable, and hence, all I4.0 value drivers will be easier to implement. It is also likely that the performance improvement obtained will also be higher than in an unstable and erratic environment. If a culture of CI and problem-solving is already established, then change is a natural part of normal work and hence the change that I4.0 means will be easier to manage with a lower threshold for acceptance. As touched upon in section 2.1.2, without a culture that reflects the Toyota Way values; the right Lean thinking will not flourish and instead there will be a focus on specific Lean tools which in time will be difficult to sustain (Suzaki, 1987). Following the same line of reasoning: I4.0 itself cannot benefit and sustain CI without the proper culture reflecting the Toyota Way values, since I4.0 by definition, is just a set of disruptive technologies, tools and methods used to enhance the manufacturing capabilities, not a set of values and principles. Finally, Lean tools will most likely be of use both in the preparation of the implementation and afterwards to directly support success. Examples mentioned were value steam mapping in the preparation and gradual PDCA during implementation rollout. In line with the recommendations the emphasis is to focus on the small, simple and low-risk improvements where possible, rather than the larger and complicated improvements. This is in accordance with the very definition of CI as a core-value, section 2.1.2. Echoing the findings under Purpose it is clear that Implementation should have a vision and road map based on a solid business case in which Challenges and Risks are thoroughly assessed and accounted for prior to implementation. The success factors under Implementation strategy also includes a range of more pragmatic aspects ranging from data analysis tools and visualization to data stability and sampling parameters. These success factors are seen as more relating to success of I4.0 implementation and less to its relationship to CI, although as already mentioned, the two goes hand in hand. Figure 17 displays the themes of success factors as a means to realizing the I4.0 Possibilities, ensuring that Challenges are overcome and that Risks are mitigated in order to not affect CI through negatively impacting problem-sensitivity (see problems) and problem-solving capabilities (solve problems). The visibility and focus of I4.0 adopters are often centered around the various Possibilities that I4.0 implementation entails. However, submerged under the surface of what is visible, are numerous Challenges and Risks that this study have highlighted and provided insight into.

Figure 17 Success factors as a means to ensure CI, realizing Possibilities whilst overcoming Challenges & mitigating Risks.

68

R3 concerns the success factors which would prevent a decrease in the rate of improvement as a consequence of I4.0 implementation. Based on the above discussions, it has been shown that the success factors can be grouped into four categories; Purpose, that concerns success factors related to having a clear purpose of embarking on the I4.0 transformation journey, i.e., having thoroughly answered “why” a technology should be implemented. Involvement of people concerns success factors related to how people and stakeholders should be involved in the I4.0 transformation to reach the full potential. Competence includes success factors related to the education and training as well as the knowledge needed to reach the potential of I4.0 while minimizing the Risks of negatively impacting CI. Finally, Implementation strategy concerns the success factors related to, e.g., having an adequate action plan and a solid Lean foundation prior to implementation.

5.5. Summary of discussion

To condense the findings from the first and the second iteration discussed in section 5.2 and section 5.4, respectively; three figures have been created to illustrate the main takeaways for each different value driver. Figure 18, illustrate a summary of the Possibilities, Challenges and Risk for the Connectivity value driver, including the main themes and success factors. Figure 19 illustrate a summary for Intelligence and Figure 20, a summary for Flexible automation.

Figure 18 Summary for Connectivity.

69

Figure 19 Summary for Intelligence.

Figure 20 Summary for Flexible automation.

70

6. Conclusions

This chapter outlines a short summary of the results, answers to research questions, theoretical and practical research contribution, limitations and improvement potential as well as guidance for future research.

6.1. Summary

I4.0, also known as the fourth industrial revolution, is dawning upon the manufacturing industries. The revolution is characterized by the diffusion of novel technologies into the manufacturing environment. Common for the various revolutions prior to I4.0 are the significant changes they all have imposed on the manufacturing environment. However, contrary to the previous revolutions, I4.0 is yet an understudied topic and its implementation consequences are both unexplored and unclear (Piccarozzi, et al., 2018; Schneider, 2018). From published studies, an overwhelmingly technology centered and positive perspective on I4.0 have been embraced, overlooking the potential Challenges and Risks facing adopters (Mohamad, 2018; Karadayi-Usta, 2019; Schneider, 2018).

Lean, with its origins in the Japanese automotive production and Toyota, is a well-studied topic and is broadly seen as the most commonly adopted manufacturing philosophy since several decades (Shah & Ward, 2003). One of the core values of Lean is Continuous Improvement (CI) which means that “everything can always improve”. CI is about the many small, simple and cheap improvements in which everyone is involved in, every day. The importance of CI cannot be understated as it is a necessity for all existing organizations subject to competition (Bicheno & Holweg, 2008; Sanders, et al., 2016).

There is a lack of comprehensive frameworks which combines Lean values and principles with I4.0 solutions (Kolberg & Zühlke, 2015; Leyh, et al., 2017; Wagner, et al., 2017). Moreover, the impact on CI following I4.0 implementation is yet to be explored by researchers, despite the relevance and interdependency of the two subjects (Buer, et al., 2018; Mayr, et al., 2018; Sanders, et al., 2016). To explore the Possibilities, Challenges and Risks of I4.0 implementation on CI, the authors have performed a Delphi survey containing open-ended questions, targeting Lean and I4.0 experts within the manufacturing industry. The survey has been conducted in two iterations were the first iteration allowed participants to confer how the conditions for CI would be affected post I4.0 adoption, outlining both positive and negative aspects. For the second iteration, participants have been encouraged to identify success factors on the basis of the Challenges and Risks identified in the first iteration; success factors that can ensure a good progress of CI post I4.0 implementation. In an attempt to enable the I4.0 concept to be subjected to a critical review, the scope of the I4.0 definition has been narrowed into specific value drivers (Connectivity, Intelligence and Flexible automation) that each encompass a set of key technologies.

The survey raw-data has been processed into text-strings containing the participants’ key points and arguments, which have been subsequently grouped into categories and themes. In total, 64% of the identified text-strings provided for all value drivers were positive, indicative of an overall positive belief in the impacts of I4.0 on the conditions for CI through increased problem-sensitivity and

71

problem-solving capabilities. While the results were arguably a reflection of the current I4.0 hype, they also highlighted the overall difficulty in critically assessing the potential impacts from technologies that have not yet made their way beyond the piloting stage into large-scale rollout. The participants also identified several Challenges (18%) and Risks (18%) for each value driver that might increase the inertia post-technology diffusion by negatively affecting CI. In relation to these, the participants also identified numerous success factors to ensure that the rate of improvement is not negatively affected by the implementation of each value driver.

6.2. Answers to research questions

R1: How does the Introduction of I4.0 affect the conditions for continuous improvement with regards to the rate of improvement?

I4.0 provides numerous Possibilities for CI within the manufacturing environment; Possibilities that can increase the rate of improvement post I4.0 implementation by positively affecting the conditions, i.e., the problem-sensitivity and problem-solving capabilities. This is done by, for example, facilitating data collection and data visualization through Connectivity; by applying Intelligence to facilitate data-driven insights and forecasts and by applying Flexible automaton to realize Lean principles such a JIT. However, the introduction of the I4.0 value drivers also accompany Challenges and Risks. The Challenges need to be overcome in order to reach the full potential of the value drivers’ positive impacts as well as to minimize their potential negative impacts on the conditions for CI. Not overcoming a Challenge would either results in loss of a possibility or in the failure of mitigating a risk. Examples of Challenges include providing necessary training and education to operators as well as inducing trust and understanding for intelligent solutions. Contrary to the Challenges, the Risks are inherently negative. The realization of Risks might decrease the rate of improvement post I4.0 adoption by negatively impacting the conditions for CI, resulting in both a lower improvement level per activity and a lower frequency of improvement activities. Thus, as Possibilities are inherently positive for the CI-conditions, Risks negatively affects the conditions for CI. Examples of identified Risks are outlined in the below paragraphs.

R2: What are the potential factors which can decrease the rate of improvement following an increased connected, intelligent and flexibly automated manufacturing environment?

Below paragraphs outline a short summary of the identified Risks associated with the implementation of the three value drivers. The Risks need to be mitigated in order to not negatively impact the rate of improvement.

In terms of the Connectivity value driver, two themes have been found; Digital waste and Rigidity. The first theme encompasses the Risks associated with data saturation, incorrect data visualization and data misinterpretation. The second theme includes Risks associated with a lower involvement of shop floor operators due to lack of expertise and an overall increased rigidity due to increasing cost of change.

For the Intelligence value driver, two themes have been found. Data quality & Application and Ownership & Participation. The first relates to Risks derived from applying intelligent solutions on

72

data of inadequate quality in combination with cyber-security issues and other Risks following improper implementation, usage and supervision of AI development. The latter includes Risks associated with an increased need for experts to realize even small improvement.

Finally, two themes were found for the Flexible automation value driver: Rigidity and Excessive solutions. The first related to the Risks associated with an increased difficulty and cost of changing systems due to increased complexity and interdependency. The latter includes Risks that solutions becomes redundant as it is used where it is not needed and where less costly solutions could prove as suitable or even better alternatives.

R3: What are the success factors preventing a decrease in the rate of improvement as a consequence of I4.0?

Four categories of success factors were identified: Purpose, Involvement of people, Competence and Implementation strategy. Purpose concerns success factors related to having a clear purpose of embarking on the I4.0 transformation journey, i.e., having thoroughly answered “why” a technology or a set of technologies should be implemented. Involvement of people concerns success factors related to how people and stakeholders should be involved in the I4.0 transformation as to reach the potential of I4.0. Competence includes success factors related to education and training as well as knowledge as a means to reach the potential of I4.0 while minimizing the Risks of negatively impacting CI. Finally, Implementation strategy concerns the success factors related to, for example, having an adequate action plan and Lean foundation for the integration of I4.0.

6.3. Theoretical contribution

This research has contributed to the necessary discussion about the Possibilities, Challenges and Risks of the I4.0 value drivers’ impact on the conditions for CI in the manufacturing environment. To the extent of the authors’ knowledge, this has not been previously done despite the importance of CI for corporate success and the growing relevance of the I4.0 topic. Even though some studies exist within the interdomain between Lean and I4.0 (section 2.6), no studies have been found that examine I4.0’s impact on CI, which has been called out in several other articles (Buer, et al., 2018; Mayr, et al., 2018; Sanders, et al., 2016).

In addition to the identified Possibilities, which reinforced the positive correlations between Lean and I4.0 observed in previous studies (section 2.6), relations between the present literature found on barriers to CI (section 2.2) and between I4.0, have been drawn and argued for throughout the discussion. The study has also added to the small pool of articles studying I4.0 from a management perspective, which distinguish it from the overwhelmingly technology-centered perspective observed in the current I4.0 literature (Piccarozzi, et al., 2018; Schneider, 2018). Furthermore, the literature summary presented in section 2.7, in combination with the discussions presented in section 5.1, can serve as a starting ground for a comprehensive Lean-I4.0 framework as called for by several authors (Kolberg & Zühlke, 2015; Leyh, et al., 2017; Wagner, et al., 2017).

73

6.4. Practical contribution

With emphasis on the Challenges and Risks, the authors have tried to cut through the noise of the ongoing I4.0 hype. As such, this research has introduced an alternative perspective that sets it apart from the overwhelmingly positive discussions surrounding I4.0 as observed in other studies (Dombrowski, et al., 2017; Buer, et al., 2018; Mayr, et al., 2018; Mohamad, 2018; Schneider, 2018; Karadayi-Usta, 2019). As companies are likely to be blinded by the overly positive rhetoric originating from consultancy firms and other stakeholders that would benefit from selling I4.0 related products and services; such a contribution ought to be warmly welcomed by all current non-adopter companies. Moreover, the importance of CI for the manufacturing industries cannot be understated, which is why a realistic exploration of the I4.0 impact on the conditions for CI needed to be investigated.

As this study has contributed to an increased visibility of the Challenges and Risks, I4.0 adopters may now include them as factors in their I4.0 business case and thereby approach the implementation more critically and realistically. The findings can also be used as a foundation for discussions prior, during and post I4.0 implementation. It can assist the process of identifying and embracing the Possibilities as well as identifying, overcoming and mitigating the Challenges and Risks associated with different I4.0 value drivers.

In summary, the practical contribution has been derived from a more realistic understanding that in addition to the positive aspects, there might also be negative aspects related to I4.0 implementation. In addition, specific Challenges and Risks have been outlined together with success factors that can potentially guide companies in their I4.0 implementation journey; realizing its Possibilities without negatively impacting CI.

6.5. Research limitations and improvement potential

Examining a topic that has not previously been thoroughly explored induces uncertainties in the research’s validity and reliability, as discussed in section 3.5-3.6. It is also important to keep in mind that this study has been exploratory; as such, no key variables or relationships have been defined and no conclusive results can therefore be derived. Instead, the I4.0 impact on CI have been assessed by collecting and interpreting qualitative data which have yielded insights rather than conclusions. As the interpretation of the data has been both judgmental and biased, the results should be subject to a comprehensive critical review. Moreover, the sample size was overall small, and the findings can therefore not be interpreted for a generalized population.

Due to the relatively complex nature of the subjects involved in this research, Lean/CI and I4.0, it would in hindsight have been more appropriate to do this study using interviews rather than sending out a survey. A survey has been chosen as a means of saving time as well as allowing the participants sufficient time to reflect on the questions and replying at their convenience. An interview study would however have had the benefit of more in-depth framing of the questions as well as more elaborate answers and would also have allowed for the interviewee to ask questions. For example, some of the identified success factors can arguably be viewed as trivial due to their “obvious” nature. Hence, interviews could have allowed for more thorough evaluations of each success factor by

74

directly asking participants to further elaborate on their answers. This could potentially have increased the overall value and relevance of the success factors as well as differentiated them more from the Challenges. The main questions could still have been shared in advance of the interview to allow time for reflection by the recipients prior to the actual interview. Interviews would also have allowed for a more thorough outline of the two subjects and different scenarios in the context of manufacturing to minimize misunderstanding and misinterpretation. Further improvement potential could also have been realized by focusing more on lighthouse companies with regards to their I4.0 implementation. It would have been interesting to have more experts from lighthouse companies involved due to their overall experience on the I4.0 transformation scale. It would also have been interesting to try to capture their I4.0 journey and their associated metrics, and from that explore the impact on CI. The unforeseen outbreak of Covid-19 resulted in research limitations as both the response rate was impacted (especially at Volvo CE due to short term unemployment) and the data-gathering phase was therefore significantly prolonged. This limited the Delphi survey to two iterations and no convergence in expert opinions could be resolved. One more iteration could have been conducted focusing on the Challenges and Risks in particular; categories which were not saturated. The Risks in particular should have been further investigated and clarified to yield more distinct themes. The time to follow up any ambiguities in participant’s responses from the first round was also limited and required the authors to interpret the answers. So, instead of identifying success factors, Challenges and Risks could have been further investigated. This would arguably have increased the quality of the research.

6.6. Future research

This study has merely touched the surface of the numerous Possibilities, Challenges and Risks that companies can encounter on their I4.0 transformation journey in relation to CI. Future research should aim at further substantiate the findings in this study. More specifically, future research should aim at reaching a higher degree of data saturation, particularly focusing on clarifying the potential downsides of the I4.0 implementation on CI, i.e., the Challenges and Risks. Moreover, as I4.0 technologies are more widely adopted, researcher should try to quantify the Risks impact by conducting quantitative studies. The studies should preferably be longitudinal in order to capture the potential differences in CI-rates, pre and post implementation. This could be done from a more general perspective by studying the impacts from each individual I4.0 value driver, or from a more detailed perspective, e.g., on a case-to-case basis by assessing specific technologies separately or jointly.

75

Bibliography

Achton, K., 2009. That 'Internet of Thing' Thing. [Online] Available at: https://www.rfidjournal.com/articles/view?4986 [Accessed 10 March 2020].

Agostini, L. & Filippini, R., 2019. Organizational and managerial challenges in the path toward Industry 4.0. European Journal of Innovation Management, 22(3), pp. 406-421.

AMFG, 2019. Industry 4.0: 7 Real-World Examples of Digital Manufacturing in Action. [Online] Available at: https://amfg.ai/2019/03/28/industry-4-0-7-real-world-examples-of-digital-manufacturing-in-action/ [Accessed 16 February 2020].

b.telligent, 2020. Industry 4.0. [Online] Available at: https://www.btelligent.com/en/portfolio/industry-40/ [Accessed 5 March 2020].

Bélanger-Barrette, M., 2015. Blog Robotic. [Online] Available at: https://blog.robotiq.com/what-does-collaborative-robot-mean [Accessed 6 February 2020].

Benitez, G. B., Lima, M. J. d. R. F., Lerman, L. V. & Frank, A. G., 2019. Understanding Industry 4.0: Definitions and Inisights from a Cognitive Map Analysis. Brazilian Journal of Operations & Production Management, pp. 192-200.

Besanko, D., Dranove, D., Shanley, M. & Schaefer, S., 2017. Economies of Strategy. Hoboken: John Wiley & Sons.

Bessant, J. & Caffyn, S., 1997. High-involvement innovation through continuous improvement. International Journal of Technology Management, pp. 7-27.

Bhuiyan, N. & Baghel, A., 2005. An overview of continuous improvement: from the past to the present. Management Decision, pp. 761-771.

Bicheno, J. & Holweg, M., 2008. The Lean Toolbox. Buckingham: Picsie Books.

Boer, d. & Enno, 2018. Five Steps to Get Out of Pilot Purgatory. [Online] Available at: https://www.industryweek.com/technology-and-iiot/article/22026267/five-steps-to-get-out-of-pilot-purgatory [Accessed 17 March 2020].

Bonekamp, L. & Sure, M., 2015. Consequences of Industry 4.0 on Human Labour and Work Organisation. Journal of Business and Media Psychology, pp. 33-40.

Bose, R., 2009. Advanced analytics: opportunities and challenges. Industrial Management & Data Systems, pp. 155-171.

Brettel, M., Friederichsen, N., Keller, M. & Rosenberg, M., 2014. How Virtualization, Decentralization and Network Building Change the Manufacturing Landscape: An Industry 4.0 Perspective. International Journal of Mechanical, Aerospace, Industrial, Mechatronic and Manufacturing Engineering, 8(1).

76

Buer, S.-V., Strandhagen, Jan, O. & Chan, F. T. S., 2018. The link between Industry 4.0 and lean manufacturing: mapping current research and establishing a research agenda. International Journal of Production Research, pp. 2924-2940.

Burkett, D., 2017. Digitisation and Digitalisation: What Means What?. [Online] Available at: https://workingmouse.com.au/innovation/digitisation-digitalisation-digital-transformation [Accessed 23 February 2020].

Caffyn, S., 1999. Development of a continuous improvement self-assessment tool. Journal of Integrated Manufacturing Systems, pp. 1138-1153.

Christersson, P., 2012. TechTerms - The Tech Terms Computer Dictionary. [Online] Available at: https://techterms.com/definition/bandwidth [Accessed 9 March 2020].

Dalkey, N. & Helmer, O., 1963. An experimental application of the DELPHI method to the use of experts. Management Science, 9(3), pp. 458-467.

Dombrowski, U., Richter, T. & Krenkel, P., 2017. Interdependencies of Industrie 4.0 & Lean Production Systems - a use cases analysis -. Procedia Manufacturing, pp. 1061-1068.

ETCIO, 2016. Convergence of IT & OT: a new business imperative to fortify information security. [Online] Available at: https://cio.economictimes.indiatimes.com/news/digital-security/convergence-of-it-ot-a-new-business-imperative-to-fortify-information-security/53374920?redirect=1 [Accessed 26 March 2020].

Faggella, D., 2020. EMERJ. [Online] Available at: https://emerj.com/ai-glossary-terms/what-is-machine-learning/ [Accessed 16 February 2020].

Frankenfield, J., 2020. What Is Artificial Intelligence (AI)?. [Online] Available at: https://www.investopedia.com/terms/a/artificial-intelligence-ai.asp [Accessed 16 February 2020].

Fujimoto, T., 1999. The Evolution of a Manufacturing System at Toyota. New York: Oxford University Press.

Garcia-Sabater, J. P. & Marin-Garcia, J. A., 2011. Can we still talk about continuous improvement? Rethinking enablers and inhibitors for successful implementation. International Journal of Technology Management (IJTM), Volume 55.

Gartner, 2020. Operational Technology (ot). [Online] Available at: https://www.gartner.com/en/information-technology/glossary/operational-technology-ot [Accessed 5 February 2020].

Glass, R. et al., 2018. Identifying the barriers to Industrie 4.0. Procedia CIRP, pp. 985-988.

Goering, K., Kelly, R. & Mellors, N., 2018. McKinsey & Company. [Online] Available at: https://www.mckinsey.com/business-functions/mckinsey-digital/our-insights/the-next-horizon-for-industrial-manufacturing [Accessed 16 February 2020].

77

Hasson, F. & Keeney, S., 2011. Enhancing rigour in the delphi technique research. Technological Forecasting & Social Change, 78(9), pp. 1695-1704.

Hasson, F., Keeney, S. & McKenna, H., 2000. Research guidelines for the Delphi survey technique. Journal of Advanced Nursing, pp. 1008-1015.

Helmer, O., 1999. The past and future of futures research. Technological Forecasting & Social Change, 62(1), pp. 33-35.

IBM, 2020. IBM. [Online] Available at: https://www.ibm.com/analytics/hadoop/big-data-analytics [Accessed 25 February 2020].

Jabbour, A. B. L. d. S. et al., 2018. When titans meet – Can industry 4.0 revolutionize the environmentally sustainable manufacturing wave? The role of critical success factors. Technological Forecasting & Social Change, Volume 132, pp. 18-25.

Japan Management Association, 1986. Kanban Just in Time at Toyota. Portland: Productivity Press.

Julsrud, T. E. & Priya Uteng, T., 2015. Technopolis, shared resources or controlled mobility? A net-based delphi-study to explore visions of future urban daily mobility in Norway. European Journal of Futures Research, 3(1), pp. 1-13.

Karadayi-Usta, S., 2019. An Interpretive Structural Analysis for Industry 4.0 Adoption Challenges. IEEE Transactions on Engineering Management, pp. 1-6.

Kinnear, C. & Roodt, G., 1998. The Development of an Instrument for Measuring Organisational Inertia. Journal of Industrial Psychology, pp. 44-54.

Knorr, E., 2018. What is cloud computing? Everything you need to know now. [Online] Available at: https://www.infoworld.com/article/2683784/what-is-cloud-computing.html [Accessed 6 April 2020].

Kolberg, D. & Zühlke, D., 2015. Lean automation enabled by Industry 4.0 Technologies. IFAC-PapersOnLine, pp. 1870-1875.

Kovar, J. et al., 2016. Virtual Reality in Context of Industry 4.0. s.l., IEEE.

Küpper, D. et al., 2019. BCG. [Online] Available at: https://www.bcg.com/publications/2019/advanced-robotics-factory-future.aspx [Accessed 16 February 2020].

Lasi, H. et al., 2014. Application-Pull and Technology-Push as Driving Forces for the Fourth Industrial Revolution. Business & Information Systems Engineering, pp. 239-242.

Lee, A. E., 2008. Cyber Physical Systems: Design Challenges, Berkeley: University of California.

Leyh, C., Martin, S. & Schaffer, T., 2017. Industry 4.0 and lean production - A Matching Relationship? An analysis of selected Industry 4.0 models. PRAGUE, FEDCSIS, pp. 989-993.

Lillrank, P. & Noriaki, K., 1989. Continuous Improvement: Quality Control Circles in Japanese Industry. s.l.:Center for Japanese Studies, University of Michigan.

78

Linstone, H. A. & Turoff, M., 2011. Delphi: A brief look backward and forward. Technological Forecasting & Social Change, 78(9), pp. 1712-1719.

Lodgaard, E., Ingvaldsen, J. A., Aschehoug, S. & Gamme, I., 2016. Barriers to Continuous Improvement: Perceptions of Top Managers, Middle Managers and Workers. Procedia CIRP, Volume 41, pp. 1119-1124.

Lorenz, R. et al., 2019. Lean and Digitalization—Contradictions or Complements?. Advances in Production Management Systems. Production Management for the Factory of the Future pp 77-84, pp. 77-84.

Louw, G. J. & Martins, N., 2004. Exploring Inertia in a Typical State Organisation. Journal of Industrial Psychology, pp. 55-62.

Machado, C. G. et al., 2019. Industry 4.0 readiness in manufacturing companies: challenges and enablers towards increased digitalization. Procedia CIRP, pp. 1113-1118.

Mayr, A. et al., 2018. Lean 4.0 - A conceptual conjunction of lean management and Industry 4.0. Procedia CIRP 72, pp. 622-528.

McClelland, C., 2017. The Difference Between Artificial Intelligence, Machine Learning, and Deep Learning. [Online] Available at: https://medium.com/iotforall/the-difference-between-artificial-intelligence-machine-learning-and-deep-learning-3aa67bff5991 [Accessed 18 March 2020].

McKinsey & Company, 2018. Digital Manufacturing – escaping pilot purgatory, s.l.: Digital McKinsey.

McLean, R. & Antony, J., 2014. Why continuous improvement initiatives fail in manufacturing environments? A systematic review of the evidence. International Journal of Productivity and Performance Management, pp. 370-376.

McLean, S. R., Antony, J. & Dahlgaard, J. J., 2015. Failure of Continuous Improvement initiatives in manufacturing environments: a systematic review of the evidence. Total Quality Management & Business Excellence, pp. 219-237.

Melander, L., 2018. Scenario development in transport studies: Methodological considerations and reflections on delphi studies. Futures, Volume 96, pp. 68-78.

Miller, J., 2017. The drivers of flexible automation. [Online] Available at: https://www.controleng.com/articles/the-drivers-of-flexible-automation/?fbclid=IwAR1XkTgz6LXZ-m0MHH2OrqOnA28Z0mXgg_s657VCuzbXXU1pM7co2x72hsU [Accessed 5 March 2020].

Miller, P., 2018. What is edge computing?. [Online] Available at: https://www.theverge.com/circuitbreaker/2018/5/7/17327584/edge-computing-cloud-google-microsoft-apple-amazon [Accessed 17 April 2020].

Modig, N. & Åhlström, P., 2015. This is Lean. Stockholm: Rheologica Publishing.

79

Mohamad, M., 2018. Challenges and Benefits of Industry 4.0: An overview. International Journal of Supply and Operations Management, 5(3), pp. 256-265.

Mora, E., Gaiardelli, P., Resta, B. & Powell, D., 2017. Exploiting Lean Benefits Through Smart Manufacturing: A Comprehensive Perspective. Hamburg, Springer.

Mrugalska, B. & Wyrwicka, M. K., 2017. Towards Lean Production in Industry 4.0. Procedia Engineering, pp. 466-473.

Nosalska, K., Piątek, Z. M., Mazurek, G. & Rządca, R., 2019. Industry 4.0: coherent definition framework with technological and organizational interdependencies. Journal of Manufacturing Technology Management.

Ohno, T., 1988. Toyota Production System, Beyond large-scale production. New York: Productivity Press.

Ohno, T., 2013. Taiichi Ohno's Workplace Management. New York: McGraw-Hill.

Pagliosa, M., Tortorella, G. & Ferreira, J. C. E., 2019. Industry 4.0 and Lean Manufacturing - A systematic literature review and future research directions. Journal of Manufacturing Technology Management.

Piccarozzi, M., Aquilani, B. & Gatti, C., 2018. Industry 4.0 in Management Studies: A Systematic Literature Review. MDPI - Sustainability, pp. 2-24.

Powell, D. et al., 2018. Towards Digital Lean Cyber-Physical Production Systems: Industry 4.0 Technologies as Enablers of Leaner Production. Seoul, Springer.

Rand Corporation, 2020. History and Mission. [Online] Available at: https://www.rand.org/about/history.html [Accessed 22 February 2020].

Rojko, A., 2017. Industry 4.0 Concept: Background and Overview. International Journal of Mobile Technologies, 11(5).

Romero, D. et al., 2018. Digital Lean Cyber-Physical Production Systems: The Emergence of Digital Lean Manufacturing and the Significance of Digital Waste. Seoul, Springer.

Rossini, M., Costa, F., Tortorella, G. L. & Portioli-Staudacher, A., 2019. The interrelation between Industry 4.0 and lean production: an empirical study on European manufacturers. The International Journal of Advanced Manufacturing Technology, p. 3963–3976.

Rowe, G. & Wright, G., 1999. The delphi technique as a forecasting tool: Issues and analysis. International Journal of Forecasting, 15(4), pp. 353-375.

Rüttimann, B. G. & Stöckli, M. T., 2016. Lean and Industry 4.0 - Twins, Partners, or Contenders? A Due Clarification Regarding the Supposed Clash of Two Production Systems. Journal of Science Service and Management, pp. 485-500.

Sanders, A., Elangeswaran, C. & Wulfsberg, J., 2016. Industry 4.0 Implies Lean Manufacturing: Research Activities in Industry 4.0 Function as Enablers for Lean Manufacturing. Journal of Industrial Engineering and Management, pp. 811-833.

80

Saturno, M., Pertel, V. M., Deschamps, F. & Eduardo de, F. R., 2018. Saturno, M., Pertel, V. M., Deschamps, F., & EdProposal for new automation architecture solutions for industry 4.0. Logforum, p. 14 (2).

Saunders, M. N., Lewis, P. & Thornhill, A., 2019. Research Methods For Business Students. 8th Edition ed. Harlow: Pearson.

Schneider, P., 2018. Managerial challenges of Industry 4.0: an empirically backed research agenda for a nascent field. Managerial Science, Volume 12, pp. 803-848.

Schwab, K., 2016. The Fourth Industrial Revolution: what it means, how to respond. [Online] Available at: https://www.weforum.org/agenda/2016/01/the-fourth-industrial-revolution-what-it-means-and-how-to-respond/ [Accessed 6 February 2020].

Shah, R. & Ward, T. P., 2003. Lean manufacturing: context, practice bundles, and performance. Journal of Operations Management, pp. 129-149.

Skhmot, 2017. Using the PDCA Cycle to Support Continuous Improvement (Kaizen). [Online] Available at: https://theleanway.net/the-continuous-improvement-cycle-pdca [Accessed 16 February 2020].

Sony, M., 2018. Industry 4.0 and lean management: a proposed integration model and research propositions. Production & Manufacturing Research, pp. 416-432.

Steinert, M., 2009. A dissensus based online Delphi approach: An explorative research tool. Technological Forecasting & Social Change, 76(3), pp. 291-300.

Suzaki, K., 1987. The New Manufacturing Challenge. New York: The Free Press.

Tapio, P., 2003. Disaggregative policy delphi: Using cluster analysis as a tool for systematic scenario formation. Technological Forecasting & Social Change, 70(1), pp. 83-101.

Tjahjono, B., Esplugues, C., Aeres, E. & Pelaez, P., 2017. What does Industry 4.0 mean to Supply Change?. Procedia Manufacturing, pp. 1175-1182.

Tortorella, G. L., Giglio, R. & van Dun, D. H., 2019. Industry 4.0 adoption as a moderator of the impact of lean production practices on operational performance improvement. International Journal of Operations & Production Management, Volume 39, pp. 860-886.

Toyota Institute, 2001. The Toyota Way. Toyota city: Toyota Institute.

Wagner, T., Herrmann, C. & Thiede, S., 2017. Industry 4.0 impact on Lean Production Systems. Procedia CIRP, pp. 125-131.

Varela, L. et al., 2019. Evaluation of the Relation between Lean Manufacturing, Industry 4.0, and Sustainability. MDPI - Sustainability, 11(5), pp. 1-19.

Vetenskapsrådet, 2002. Forskningsetiska principer inom humanistisk-samhällsvetenskaplig forskning, Stockholm: Vetenskapsrådet.

Volker, P., 2014. Augmented Reality in the Smart Factory - Supporting Workers in an Industry 4.0. Environment. s.l., IEEE.

81

Womack, J. P. & Jones, D. T., 1996. Lean Thinking. New York: Free Press.

Womack, J. P., Jones, D. T. & Roos, D., 1990. The Machine that Changed the World. New York: Free Press.

World Economic Forum, 2019. Fourth Industrial Revolution. Beacons of Technology and Innovation in Manufacturing, Geneva: World Economic Forum.

Xu, L. D., Xu, E. L. & Li, L., 2018. Industry 4.0: state of the art and future trends. International Journal of Production Research, 56(8), pp. 2941-2962.

Yin, Y., Stecke, K. E. & Li, D., 2018. The evolution of production systems from industry 2.0 through industry 4.0. International Journal of Production Research, pp. 848-861.

Zhang, J. & Jung, Y.-G., 2018. Additive Manufacturing: Materials, Processes, Quantifications and Applications. Oxford: Elsevier Inc.

Zhou, K., Liu, T. & Zhou, L., 2015. Industry 4.0: Towards Future Industrial Opportunities and Challenges. Zhangjiajie, IEEE.

82

A. Appendix – Delphi survey 1st iteration

Impact of Industry 4.0 (I4.0) on Continuous Improvement (CI) Background information Please state your current work title: Please state your current employer (company/research institution): Please state your number of years of experience with Lean/CI: Please state your number of years of experience with Industry 4.0: Survey questions The three questions below circles around the impact on the future conditions for continuous improvement following the introduction of I4.0 into the manufacturing environment; more specifically, the introduction of “Connectivity”, “Intelligence” and “Flexible automation”. As participants, you are encouraged to outline both positive and negative aspects of the I4.0 impact on CI. Try to be as comprehensive as possible (there is no restriction on the number of words) and it is ok to not answer if you do not have an opinion. A more detail description of I4.0 and CI can be found on the next page. Question 1: In what way do you think the conditions for continuous improvement will be affected following an increased “Connectivity” in the manufacturing environment, i.e., through the introduction of Internet of Things (IoT) and Cyber Physical Systems (CPS)? Answer 1: Question 2: In what way do you think the conditions for continuous improvement will be affected by the introduction of “Intelligence” in the manufacturing environment, i.e., with the introduction of advanced analytics and artificial intelligence? Answer 2: Question 3: In what way do you think the conditions for continuous improvement will be affected by the introduction of “Flexible automation” in the manufacturing environment, i.e., with the introduction of collaborative robots, 3D-printing as well as other advanced and autonomous automation? Answer 3:

83

Continuous improvement Refers to the core value of the Toyota way which states that: “everything can always improve” and “everything can always improve more than what we think”. Continuous improvement is about the many small, simple and cheap improvements in e.g., health and safety, quality, delivery and cost, which everyone is involved in, every day. For this project we are focusing especially on continuous improvement relating to productivity. Industry 4.0

Connectivity Definition Augmented Reality (AR) Refers to the integration of computer-generated information into the real-world

environment. Most current AR-applications integrate computer graphics into the user’s view of the surroundings.

Cyber-Physical System (CPS) A term used to denote a system of Integrated computations and physical processes. Utilizing embedded computers and network monitors the physical process is controlled through feedback loops between the physical process itself and the computer calculations. Essentially an integration of IT and OT in various combinations.

Internet of Things (IoT) Industrial Internet of Things (IIoT) Industrial Internet

The definition encircles the concept of a network of devices embedded with software and information sensing devices such as laser scanners, position systems, infrared sensors and Radio Frequency Identification Devices (RFID) which are connected to the internet. The integration of such sensors with the internet enables communication, real-time tracking and identification of tags that can be attached to any object.

Information Technologies (IT) Technologies that store, process and transport information.

Operational technologies (OT) Hardware and software that detects or causes a change through the direct monitoring and/or control of physical devices, processes and events.

Virtual Reality (VR) Allows for visualization of virtual objects. Full immersive VR-devices provide 3-dimensional virtual projections accessible to the user.

Intelligence Definition Advanced Analytics (AA)

An umbrella term for different advanced analytics techniques, or tools, that can be used in combination with each other to gain insights by analyzing information as well as performing predictive analysis.

Artificial Intelligence (AI) Refers to the simulation of human-like intelligence in machines that are programmed to mimic human behavior.

Machine Learning (ML) Machine leaning is a subpart of AI and can be described as “the science of getting computers to learn and act like humans do, and improve their learning over time in autonomous fashion, by feeding them data and information in the form of observations and real-world interactions”.

Flexible automation Definition Additive Manufacturing (AM) Additive Manufacturing is also referred to as 3D-printing which is a production

process by which an object is produced in an additive fashion, layer-by-layer, rather than subtracted form a predefined form.

Advanced Robotics & Autonomous Robotics

Robotics with intelligent functions that can autonomously react and act on information. Advanced robotics inherit increased perception, integrability, adaptability and mobility in comparison to conventional robotics.

Collaborative robots Cobots

A robot that works in collaboration with humans to produce or create something. There are 4 types of collaborative features for cobots: safety monitored stop, hand guiding, speed and separation monitoring and power and force limiting.

84

B. Appendix – Piloting email

Email template for guinea pigs Title: MBA research survey (PILOTING): The impact of I4.0 on Continuous improvement Dear Sir/Madam You have been previously contacted regarding participating in a piloting of a survey connected to a study that Johan Wollin, Global Director of Volvo Production System at Volvo Construction Equipment, and I, Joel Larsson, Method Developer at GKN Aerospace, are conducting. The research questionnaire will be sent out early next week and we would therefore like to share our draft email and the questionnaire with you as to get improvement suggestions before the send-out. If you see any improvement potential, then please let us know. Moreover, with your consent, we would also like you to participate in the research by answering the attached questioner to the best of your ability. Looking forward to your improvement suggestions and your answers. Thank you very much! Johan w and Joel

85

C. Appendix – Survey email template 1st iteration

Dear Sir/Madam Johan Wollin, Global Director of Volvo Production System at Volvo Construction Equipment, and I, Joel Larsson, Method Developer at GKN Aerospace; are currently undertaking research aiming to investigate the impact of Industry 4.0 adoption on the conditions for continuous improvement within the manufacturing industry. The research is a part of a master’s thesis in the final year of a Master of Business Administration from Blekinge Institute of Technology.

You have been chosen to be part of a panel of experts within the field of Lean, Industry 4.0, or both; and we would be immensely grateful if you would like to participate in this study by answering the three questions found in the attached word file before the 6th of April. The questions will be followed by three focus questions at a later stage as the survey consist of two iterations. This implies that we will consolidate the answers from all participants in the first round and then send them back to you; hence, providing you with the possibility to update your answers from the first iteration after observing the consolidated answers from the entire panel of participants. By answering the questions attached in the word document and sending them back to us, you consent to participate and to that you have understood the following:

I am participating in a research study. I have been given an explanation of the research I am about to participate in and I know what

my participation entails. My participation in this research is voluntary, and I am free to withdraw at any time without

giving any reason. My identity cannot be linked to my data and all information I give remain anonymous. My identity will be kept anonymous before, during and after the research. All data will be kept confidential and not used for commercial purposes.

If you feel that you are not the right person in your organization to be participating in this study, please do not hesitate to forward this to someone you find more suitable. For any questions regarding the research please contact: Johan.wollin@volvo.com or Joel.larsson@gknaerospace.com Sincere gratitude for your interest in our research, Johan Wollin & Joel Larsson

86

D. Appendix – Email template 2nd iteration

Dear Sir/Madam

Thank you very much for participating in the research concerning the impact of Industry 4.0 (I4.0) on the conditions for Continuous Improvement (CI). Please find attached an executive summary of the answers from the first round. Apart from the majority of the results being positive, several Challenges and Risks were identified. This is why we would like to challenge you with final round of questions to identify potential success factors for I4.0 implementation; factors ensuring a good progress of CI in a future “smart factory”. The below Figure attempts to illustrate a potential negative outcome on the rate of CI following radical improvement by technology diffusion. two scenarios are compared, A (left) and B (right).

Scenario A shows a rate of improvement using only CI.

Scenario B shows initial progress using CI succeeded by a radical improvement from the implementation of I4.0 related technologies. However, if the organization do not overcome the associated Challenges, nor mitigate the Risks, the implementation might result in a lower overall performance as a result of lower CI rate afterwards.

With reference to the above illustration; Question 1: Can you list success factors that would ensure “Connectivity” to be successfully implemented without having negative impacts on CI? Question 2: Can you list success factors that would ensure “Intelligence” to be successfully implemented without having negative impacts on CI? Question 3: Can you list success factors that would ensure “Flexible automation” to be successfully implemented without having negative impacts on CI? We would be immensely grateful to have your input. You may provide your answers directly in this email and send it back to us, preferably no later than the 27th of April. For any questions regarding the research please contact: Johan.wollin@volvo.com or Joel.larsson@gknaerospace.com

Sincere gratitude for your interest in our research,

Johan Wollin & Joel Larsson

87

E. Appendix – Executive summary for 2nd iteration

A pdf containing information from the first iteration was compiled and shared for the second iteration. The information presented in the below summary constitutes the original send-out and has therefore not been updated to account for changes made after the second iteration.

88

89