Dave Dominicus 10/12/2018
Double Degree Master Technology and Operations Management
(2017-2019)
Master’s Thesis
ByDave Dominicus
(S2338165, 160679704)
Exploring the potential of smart manufacturing for improving performance
measurement and management in SMEs
Supervisor: Dr. J.A.C. Bokhorst Second supervisor: Dr. A. Small
Word count: 15.081
University of Groningen Faculty of Economics and Business
Nettelbosje 2, 9747AE Groningen
Newcastle University Business School 5 Barrack Road, Newcastle upon Tyne,
Acknowledgements
First and foremost, I would like to thank my first supervisor dr. Bokhorst of the Rijksuniversiteit Groningen. His feedback was invaluable, providing guidance and steering me in the right direction when I was struggling. His critical reviews addressed both minor and major issues, leaving no stone unturned. I am grateful that he always responded quickly to my questions, even during weekends, ensuring that I could always continue my writing. Also, I would like to thank my second supervisor dr. Small of the Newcastle University Business School. He provided me with constructive and creative feedback, showing me new insights for major parts of this thesis. His comments showed me different perspectives on this research, allowing me to critically rethink my work. I would also like to mention that, especially with the draft thesis, the comments of dr. Bokhorst and dr. Small complemented each other very well.
I am also grateful to S. Tiggeloven of the Hogeschool van Arnhem en Nijmegen. He assisted me in finding suitable cases for this research and put me in touch with the management of each company.
I would also like to thank all companies involved in this research. Without hesitation they freed up time for the interviews during their busy schedules. They provided me with all the information I needed for the completion of this thesis.
I want to express my profound gratitude to my family, Arno, Anita and Sven, who supported me throughout my years of study. Their continuous encouragement to strive for my goals has been invaluable. Without them, I would not have accomplished what I have today.
To my friends I would like to say that I am grateful when they helped me take my mind off things during stressful times.
Abstract
For companies to remain competitive it is of fundamental importance that decisions are properly made. No matter if a decision is operational, tactical or strategic, they all contribute towards the success or failure of a business. Inherent to good decisions is accurate, relevant and up-to-date data. The process of gathering, analyzing, presenting, and acting on such data is the core focus of performance measurement and management (PMM). This process aims to optimize decision making, seeking to identify issues with the use of data and tackling them. Generally, companies face a range of issues in this process such as a lack of commitment, lack of feedback and many more. PMM can be split up into three phases: measurement (data collection), reporting (analyzing and presentation) and management (decision-making and acting), each with their own evaluation criteria in terms of effectiveness and efficiency. Especially small- and medium enterprises (SMEs), which are the driving force of the Dutch economy, tend to struggle with the aforementioned issues. The fourth industrial revolution, smart manufacturing, is making its way towards businesses. Although smart manufacturing comprises many technologies, this research focusses on three: Internet-of-Things (IoT), cloud computing and cyber-physical systems (CPSs). Key advantages of these technologies were identified and linked to the evaluation criteria of specific PMM phases. Currently, next to no attempts have been made to combine these two subjects, which is where the main contribution of this thesis lies. PMM and smart manufacturing technologies were discussed with 18 diverse interviewees at six different companies, providing many different perspectives. It became apparent that IoT can contribute directly to each PMM phase and cloud computing is able to enhance performance reporting. CPSs are currently not viable for SMEs, therefore no conclusions could be drawn for this technology. As this research is of explorative nature, further research should validate these results.
Table of contents
1. Introduction ... 1
2. Theoretical background ... 3
2.1. Small- and medium enterprises ... 3
2.2. Performance measurement and management ... 3
2.2.1. PMM hierarchy ... 5
2.2.2. Shop floor measures ... 6
2.2.3. Current IT-usage in PMM ... 7 2.2.4. PMM evaluation ... 9 2.3. Smart manufacturing ... 10 2.3.1. Internet-of-Things ... 11 2.3.2. Cloud computing ... 13 2.3.3. Cyber-physical systems ... 15 2.4. SME limitations... 17 2.5. Research question ... 18 3. Methodology ... 20 3.1. Research design ... 20
3.2. Case description and selection ... 21
3.3. Data collection... 21
3.3.1. Protocol ... 21
3.3.2. Interviews ... 22
3.4. Data analysis ... 23
3.5. Reliability and validity ... 25
4.2.5. Management support ... 34
4.2.6. Lack of resources ... 35
4.2.7. Lack of feedback ... 35
4.2.8. Time lag ... 35
4.3. Smart manufacturing and PMM ... 36
4.3.1. Internet-of-Things ... 36
4.3.2. Cloud computing ... 38
4.3.3. CPSs... 40
4.4. Smart manufacturing adoption in SMEs ... 40
4.4.1. Focus on internal processes ... 40
4.4.2. Lack of knowledge ... 41
5. Discussion ... 42
5.1. PMM... 42
5.1.1. Foundation ... 42
5.1.2. Company issues ... 43
5.2. Smart manufacturing contribution ... 45
5.2.1. Data accuracy ... 45
5.2.2. Automation ... 46
5.2.3. Availability of data ... 46
5.2.4. Time lag reduction ... 46
5.2.5. Transparency ... 47 5.2.6. Cloud computing ... 47 5.2.7. CPSs... 48 5.3. Theoretical implications ... 48 6. Conclusion ... 51 6.1. Main findings ... 51
6.2. Limitations and future research ... 52
7. References ... 54
8. Appendix ... 59
8.1. SME interview... 59
List of figures
Figure 2.1: Performance pyramid ... 5
Figure 2.2: Generic performance measures ... 6
Figure 2.3: System interaction between hierarchical levels ... 8
Figure 2.4: Applications of IoT ... 12
Figure 2.5: Cloud computing aspects... 13
Figure 2.6: Data and information flows in CPSs ... 15
Figure 2.7: 5C architecture ... 16
Figure 2.8: Conceptual model with proposed relationships... 19
Figure 4.1: PMM processes at company A ... 27
Figure 4.2: PMM processes at company B ... 29
Figure 4.3: PMM processes at company C ... 30
Figure 4.4: Measured performance in company C... 32
Figure 5.1: Empirical model ... 50
List of tables
Table 2.1: Evaluation criteria per PMM phase ... 10Table 2.2: 5C architecture related to PMM ... 17
Table 2.3: Smart manufacturing linked to PMM ... 17
Table 3.1: Interviewee summary ... 23
Table 3.2: Coding tree... 24
Table 4.1: Mentions of main PMM issues ... 36
Table 4.2: Mentions of IoT advantages ... 38
1. Introduction
Companies currently operate in environments which become more sophisticated in terms of complexity and competitiveness. Increased globalization allows companies to cooperate with more diverse partners being geographically dispersed. On the other hand, globalization opened international markets, forcing companies to compete with new market entrants. Companies must be able to adapt to the changing market requirements in this dynamic environment (Pekkola et al., 2016). Nudurupati et al. (2011) state that to proactively respond to these challenges, management requires up-to-date and accurate performance information. Performance measurement and management (PMM) allows companies to gather such data essential to making operational, tactical or strategic decisions. Therefore, PMM is of fundamental importance to business success.
Marchand and Raymond (2008) state that performance measurement has evolved from merely being a toolkit to a complete management system. This shift shows that PMM does more than simply capturing and processing data, it also evaluates this data and subsequently supports decision-making. This is made possible by advances in information technology (IT) such as data-warehousing, data-mining, expert systems, web-based technologies and artificial intelligence technologies (Marchand and Raymond, 2008). Nudurupati et al. (2011) stress the importance of IT in PMM, as a lack of it results in 1) significant time and investments being spent on data collection, analysis and reporting, 2) out-of-date and irrelevant information, and 3) large amounts of measures which are difficult to manage on paper. Neglecting IT can cause serious issues for PMM, thus emphasizing the necessity.
and processes that must be managed simultaneously. Smart manufacturing can provide a streamlined flow of information, which can significantly improve business performance of small- and medium enterprises (SMEs) as they generally must be flexible and tend to be reactive (Moeuf et al., 2018). Even though smart manufacturing is an emerging form of production and the body of knowledge is rapidly expanding (Kusiak, 2018), the real benefits and requirements for SMEs are not fully known (Moeuf et al., 2018).
Many SMEs struggle with PMM processes (Pekkola et al., 2016), further emphasizing the relevance of this research. Marchand and Raymond (2008) state that advanced IT solutions can increase PMM effectiveness and efficiency, but little to no research verifies the effects of smart manufacturing on PMM. Combining the novel concept of smart manufacturing with PMM processes is therefore where the academic contribution of this thesis lies. This research aims to identify the main advantages of smart manufacturing and how it can influence specific PMM processes. This had led to the devolvement of the research question – if PMM can play such an important role in business success, how can smart manufacturing then contribute to this key process?
2. Theoretical background
This chapter will first outline the scope of this research. Next, it will provide an in-depth analysis of the concepts PMM and smart manufacturing with its associated technologies. PMM and smart manufacturing are reviewed separately as next to no connection currently has been made in the literature, but interrelationships are proposed to combine these concepts. Throughout the smart manufacturing section, these proposed interrelationships between the corresponding technologies and PMM will be elaborated upon. Then, common issues concerning SMEs that must be considered are outlined. This chapter will conclude with a conceptual model of the proposed relationships and the research question.
2.1. Small- and medium enterprises
This thesis specifically aims at SMEs rather than large companies. The European Commission defined SMEs as companies with fewer than 250 employees, annual turnover below 50 million euros and a balance sheet equal to or less than 43 million euros (Mkbservicedesk.nl, 2018). With over 99% of companies being an SME in the Netherlands these companies represent 60% of the national GDP and 70% of the employment opportunities in 2017 (Mkbservicedesk.nl, 2018). PMM is recognized as being critical to effective and efficient management of any business through control and correction by reporting current level of performance and comparing it to the desired level (Melnyk et al., 2014). Since SMEs play such a vital role in the economy of the Netherlands it is important to understand how they can effectively and efficiently adopt and utilize PMM. Additionally, since IT is inherently related to the success of PMM (Nudurupati et al., 2011), the next generation of IT (smart manufacturing) can potentially contribute to PMM. However, it is important to understand the key differences between SMEs and large companies that might restrict the use of IT, PMM and/or smart manufacturing. These differences will be elaborated upon at the end of this chapter. First, a detailed description of the most important PMM aspects are presented.
2.2. Performance measurement and management
Strategic Measurement and Reporting Technique (SMART), Performance Prism, the Performance Measurement Matrix, Balanced Scorecard (BSC) and many more (Nudurupati et
al., 2011). Most of the work on performance measurement today has been influenced by the
latter, the BSC (Bitici et al., 2012), which will be elaborated upon later this section. These frameworks have in common that they require performance measures. Neely et al. (1995, p. 1) provide the most adopted definition for performance measurement:
“The process of quantifying the efficiency and effectiveness of an action”
Performance measurement was considered a tool to support management in decision-making, however, nowadays it plays a crucial role in continuous improvement of performance and achieving strategic goals (Marchand and Raymond, 2008). As performance measurement has evolved from a mere tool to a performance management system, three different phases can be identified. PMM comprises phase 1) measurement, phase 2) reporting and phase 3) management, where performance reporting can be defined as (Radnor and Barnes, 2007, p. 10):
“Providing an account, and often some analysis, of the level of input, activity or output of an
event or process usually against some form of target”
And performance management (Radnor and Barnes, 2007, p. 10):
“An action, based on performance measures and reporting, which results in improvements in
behavior, motivation and processes and promotes innovation”
The phases cover the entire process of identifying issues, collecting data to verify the issue, transforming the data into information, presenting the information and taking corrective (or preventive) actions. PMM is a sequential process, should the first two phases fail to collect and present accurate, reliable and up-to-date-data, decision-making will be trivial (Nudurupati et
al., 2011). This thesis will focus on each individual phase and be divided accordingly. Next, a
2.2.1. PMM hierarchy
Performance measurement should lead to performance management (Hervani et al., 2005): the multi-disciplinary process of assessing the differences between actual and desired outcomes, identifying critical differences, understanding why they have taken place and introduce corrective actions to close the performance gaps (Melnyk et al., 2014). PMM can be utilized on multiple hierarchical levels in a company (Maskell et al., 2011), as illustrated in the ‘performance pyramid’ of Lynch and Cross (1991) in figure 2.1. This is one of the many developed frameworks for PMM. The performance pyramid demonstrates how corporate vision manifests at lower levels in the company in the form of relevant performance measures. This ensures that senior management can embed its policies into every level of the company. This results in linked strategies, objectives and measures (Maskell et al., 2011).
Figure 2.1: Performance pyramid (Source: Lynch and Cross, 1991)
PMM should be balanced, implying the use of different measures (i.e. financial vs non-financial; quantitative vs qualitative; internal vs external) (Taticchi et al., 2015). This ensures that only data is collected that contributes to the overall strategy, providing a holistic view of company performance. Therefore, an important question to ask is whether companies have derived their shop floor measures from their corporate vision, thus adopted the ‘right’ measures. If this is not the case, PMM might not achieve expected results (Garengo et al., 2005), and will therefore be less effective.
overview of the most generic performance measures adopted. Neely et al. (1995, p. 21) have set basic guidelines to consider when setting measures:
➢ Defining goals for each department without creating inconsistencies
➢ Set measures across functional boundaries so more commitment is experienced ➢ Set both horizontal and vertical goals
➢ Make sure current measurement systems are understood, both formal and informal ➢ Consensus among management about the companies’ objectives and the means to
achieve those
Figure 2.2: Generic performance measures (Source: Hudson et al., 2001)
2.2.2. Shop floor measures
PMM has not been without critics. In an extensive literature review Neely et al. (1995) found that most companies still adopted mainly financial measures, which is considered a traditional approach (Gutierrez et al., 2015). This traditional accounting-based performance measurement method supposedly encouraged short-terminism, lacked strategic focus, encouraged local optimization, and encouraged minimization of variance instead of continuous improvement (Melnyk et al., 2014).
covering three additional aspects: 1) customers, 2) internal business processes, and 3) learning and growth. This enables companies to track financial results while simultaneously monitoring progress in building the capabilities and acquiring the intangible assets they would need for future growth; thereby supporting strategy implementation and adaptation (Kaplan and Norton, 1992).
An issue that has become prevalent since the development of the BSC is the quantity of measures that companies adopt. Radnor and Barnes (2007) encountered various cases where companies had hundreds of measures, leading to entire departments having to ‘feed the performance measurement and reporting beast’. If the amount of measures is not limited to a diverse, manageable set, companies will unavoidably drown in data. Although no perfect number exists, an excess of measures will render PMM less effective or even counter-productive (Radnor and Barnes, 2007), as a company will then ‘measure to measure’.
Next to the question if companies have set the ‘right’ measures, the quantity of measures is shown to be an equally important question for developing effective PMM activities. These questions will be addressed later this chapter. The next section discusses how IT can support PMM per phase and identifies critical issues when IT is absent.
2.2.3. Current IT-usage in PMM
The performance pyramid framework shown in figure 2.1 demonstrates PMM can (and should) be integrated on multiple hierarchical levels within the company. Each PMM phase can be linked to a corresponding level in the performance pyramid: shop floor measures (level 4) deal with performance measurement, whereas performance reporting and management deal with level 2, 3, and 4. This will be elaborated upon in the following paragraphs. Additionally, specific IT-systems that are suitable for each phase will be discussed.
Performance reporting (i.e. analyzing, presenting and interpreting data) is the process of transforming data into information and to assess the effectiveness and efficiency of an activity (Melnyk et al., 2014). Using paper-based performance data as input to manually generate information (e.g. graphs and figures) is the most basic method. Semi-manual methods include manually feeding a spreadsheet program which can automatically transform the data into graphs and figures. Preferably, reporting should be automated (Gutierrez et al., 2015) and presented in a comprehensible manner (Garengo et al., 2005). More efficient methods exist such as specially designed manufacturing execution systems (MESs) to transmit shop floor performance data to top-level systems (Saenz de Ugarte et al., 2009), as is depicted in figure 2.3. Thus, performance reporting comprises level 4 (obtaining raw data at shop floor level), level 3 (business operating system, i.e. MES) and level 2 (division-wide information). This triggers performance management, which will be discussed next.
Figure 2.3: System interaction between hierarchical levels (Source: Saenz de Ugarte et al., 2009)
As PMM is a sequential process, performance management can only take place after the data is collected, reported and sent to a top-level system (Radnor and Barnes, 2007). This process involves initiating improvement plans, training, teamwork, or similar projects to effectively increase performance where necessary. As this process relies on performance measurement and reporting it is also concerned with level 2, 3, and 4 of the performance pyramid. Additionally, as the vision is the main input for the decisions that a company makes, level 1 directly relates to performance management. Well-known top-level systems are enterprise resource planning (ERP) systems (Saenz de Ugarte et al., 2009), such as those of SAP AG, Oracle, and Microsoft. ERPs present a holistic view of the company by integrating inventory-, financial-, sales- and human resource data, supporting decision-making and helping to achieve corporate vision, thereby allowing companies to identify where performance is not up to standard (Al-Mashari
inventory and/or financial data, etc.) they want to integrate in their ERP. Likewise, costs scale according to what the company requires. Despite this, off-the-shelf ERP systems tend to be expensive (Al-Mashari et al., 2003), which limits adoption among SMEs that lack such funds.
2.2.4. PMM evaluation
It is important to note that when IT is absent it causes various issues for all PMM phases. Efficiency issues arise due to significant time being spent on data collection, analysis and reporting (measurement and reporting phase). Furthermore, the effectiveness of PMM decreases as information is out-of-date and irrelevant and it is increasingly more difficult to manage measures on paper (reporting and management phase). Additional reasons which, in a general sense, limit efficiency and/or effectiveness of PMM include, but are not limited to (Nudurupati et al., 2011, p. 1):
➢ PMM being static and historic, instead of dynamic ➢ Lack of management support
➢ Employees not seeing added value as they do not understand potential benefits ➢ Using PMM as a command and control mechanism instead of supporting
Two questions have been raised that (partially) determine the effectiveness of PMM: whether companies have adopted the ‘right’ measures and what the quantity of measures is. These questions mainly relate to the measurement phase, and are part of a fundamentally important question: what exactly makes PMM efficient and effective? In terms of efficiency, PMM can be evaluated depending on the amount of resources that are used to achieve certain objectives (Otley, 1999). Hwang et al. (2017) remark that incorrect data is unusable, diminishing performance measurement effectiveness. Therefore, measurement effectiveness can be evaluated based on data accuracy. Gutierrez et al. (2015) note that the aggregation level of measures is important to consider, thus the level of detail (e.g. displaying OEE, or causes of shutdowns). Higher level of detail allows for better analysis, thereby increasing performance reporting effectiveness.
has achieved what it is capable of and therefore performance no longer improves but goals are still achieved (e.g. 95% OEE), PMM can still be considered effective. It is important to note that performance management is not associated with efficiency and can only be evaluated based on effectiveness (Radnor and Barnes, 2007). Table 2.1 presents evaluation criteria specifically per PMM phase which will be adopted in this research.
Table 2.1: Evaluation criteria per PMM phase
Since this thesis is divided per PMM phase it is of fundamental importance to understand what role IT can play in each phase. The next section will introduce the next generation of IT: smart manufacturing technologies. It will elaborate on how and why these technologies can be used within each PMM phase.
2.3. Smart manufacturing
No consensus for a definition of smart manufacturing currently exists in the literature (Kusiak, 2018). The term smart manufacturing has been adopted to refer to manufacturing systems that combine advanced manufacturing capabilities and digital technologies throughout the product lifecycle (Helu et al., 2016). Multiple definitions with significant overlap for the same concept can be identified, e.g. smart factory (Gabriel and Pessl, 2016), smart manufacturing (Helu et
al., 2016), ubiquitous manufacturing (Wang et al., 2018), and Industry 4.0 (Zhong et al., 2017).
Throughout this thesis none of these terms will be used interchangeably as it will be referred to as smart manufacturing. The definition that accurately describes smart manufacturing is given by NIST (National Institute of Standards and Technology) (Kang et al., 2016, p. 1; Kusiak, 2018, p. 2):
“Smart manufacturing is a fully integrated, collaborative manufacturing system that responds
Smart manufacturing aims to add intelligence to manufacturing systems by incorporating new information technologies (IT) (Zhong et al., 2017) such as: service-oriented computing (Kusiak, 2018), virtual reality (Moeuf et al., 2018), Internet-of-Things (IoT), cloud computing, cyber-physical systems (CPSs) (Kang et al., 2016; Kusiak, 2018; Tao et al., 2018), big data, and artificial intelligence (Tao et al., 2018). Although these concepts, and others, all enable smart manufacturing, this thesis will currently focus on three technologies which offer potential improvements for PMM efficiency (e.g. automation) and effectiveness (e.g. enhanced visualization): IoT, cloud computing and CPSs. These potential PMM improvements will be analyzed in detail in the following sections. IoT and cloud computing are relatively cheap (when compared to complex technologies as CPSs), resulting in a higher adoption rate among SMEs (Moeuf et al., 2018). Therefore, they will be included in this research. IoT and cloud computing can be considered the building blocks, or enablers, of CPSs (Ochoa et al., 2017). Although there are no reported cases of CPS use in SMEs (Moeuf et al., 2018), recent developments have led to higher availability and affordability of such systems (Lee et al., 2015). As companies continue to move towards implementing such high-tech systems, it is interesting to explore the potential of CPSs within SMEs. Due to time restrictions, other technologies are left out of the scope of this research.
The following sections will focus on these three fundamental technologies and how they can potentially benefit PMM.
2.3.1. Internet-of-Things
IoT is one of the driving technologies behind smart manufacturing. It is regarded as an extension of the internet wherein the physical and digital world are merging (Jeschke et al., 2017). Cooperating robots, intelligent infrastructures and interconnected, autonomous cars are well-known examples that incorporate IoT-technologies (Jeschke et al., 2017). Figure 2.4 visualizes how a diverse range of objects (or ‘things’) can be connected to the internet, which are referred to as ‘smart’ objects. Subsequently, the IoT-paradigm can be defined as (Ochoa et
al., 2017, p. 1):
“A world-wide network of interconnected heterogeneous objects that are uniquely
Figure 2.4: Applications of IoT (Source: Gabriel and Pessl, 2016)
Key issues impairing the effectiveness of PMM are a lack of flexibility and not being able to respond to a quickly changing environment (Gutierrez et al., 2015). Bi et al. (2014) proposed the IoT as a solution to this problem as it enhances system flexibility and dynamicity through a range of functions. IoT-infrastructures can automate measuring, identifying (RFID), positioning, tracking, and monitoring functions (Bi et al., 2014). RFID technology has been used for identifying various objects in warehouses, production shop floors, logistics companies, distribution centers, retailers, and disposal/recycle stages (Zhong et al., 2017). Such smart objects have predefined interconnectivity with specific logics such as manufacturing procedures. This allows for immediate data collection and transmission to realize real-time data availability for production management; thus, reducing time lag between data collection and presentation.
increased data accuracy and up-to-date (real-time) data. Additionally, IoT replaces manual data collection methods which should enhance performance measurement efficiency.
Ochoa et al. (2017) argue that cooperation among smart objects (i.e. objects connected to the internet) allows loosely-coupled decentralized systems (or CPSs) to be built. Although Wang
et al. (2015) state that IoT is not required for CPSs, many other academics (Lu et al., 2016;
Monostori, 2014; Ochoa et al., 2017; Thoben et al., 2017) tend to agree that IoT is inherent to such systems. The next section describes another enabling smart technology: cloud computing.
2.3.2. Cloud computing
The body of literature on cloud computing has been rapidly increasing over the last few years. This resulted in many different definitions, however, the following will be adopted for this thesis as it covers all aspects (Alshamaila et al., 2013, p. 2):
“A style of computing where massively scalable IT-related capabilities are provided as a
service using internet technologies to multiple external customers”
According to NIST (National Institute of Standards and Technology) an ideal cloud comprises at least the characteristics, delivery models and deployment models as presented in figure 2.5 (Zhong et al., 2017).
Figure 2.5: Cloud computing aspects (Source: Bi et al., 2014)
pay-per-use delivery models: Hardware-as-a-Service (HaaS), Design-as-a-Service (DaaS), Machining-as-a-Service (MCaaS), and Infrastructure-as-a-Service (Wang et al., 2018). These models are supported by cloud computing, IoT, virtualization and service-oriented technologies which allow manufacturing assets to be shared and circulated (Zhong et al., 2017). Examples include, but are not limited to, design (CAD), computer-aided-engineering (CAE) and computer-aided-manufacturing (CAM). As companies need to be more adaptive to dynamic markets, web-based cloud manufacturing offers the potential to overcome limitations in current, static systems such as centralized resource utilization, unidirectional information flow and offline decision making (Wang, 2013). Information will be available in the cloud for authorized personnel to access it when necessary.
The approach proposed by Wang (2013) is a web-based interface for companies where decision-modules reside in the servers of the provider: the cloud. These decision-modules require up-to-date machine availability information (which can be provided if machines are connected via the web) which can then calculate the optimal process plan for the shop floor. The web-based interface provides a comprehensible, visual image of the shop floor availability and performance, thus enhances visibility. The company becomes more flexible, being able to quickly adapt to unexpected occurrences at the shop floor (e.g. machine failure, rush orders, delays) (Wang, 2013).
Cloud computing infrastructure can be provided as-a-service by third parties (Wang, 2013), lowering the threshold for SMEs to adopt such systems due to increased simplicity and allowing them to more easily manage the process. Furthermore, the scalability aspect leads to companies of all types and sizes to adopt cloud computing with a minimum budget and without investing in licensing new software, incorporating new infrastructure, or training new employees (Zhong et al., 2017). This results in less time and money being spent on maintaining the infrastructure as this is being outsourced to a third party. Additionally, this will remove the necessity for a multitude of separate systems; thereby increasing reporting efficiency. Availability of data and visualizationallow for more effective analysis of operators as they can quickly judge the situation; thus, supporting performance reporting effectiveness.
2.3.3. Cyber-physical systems
CPSs are one of the key technologies that enable smart manufacturing. CPSs can be defined as (Lee et al., 2015, p. 1):
“Transformative technologies for managing interconnected systems between its physical
assets and computational capabilities”
The embedded algorithms that control and monitor physical processes contain feedback loops between the physical assets and computational capabilities (Gabriel and Pessl, 2016). This process is visualized by Kusiak (2018), shown in figure 2.6.
Figure 2.6: Data and information flows in CPSs (Source: Kusiak, 2018)
Figure 2.7: 5C architecture (Source: Lee et al., 2015)
The smart connection level is concerned with acquiring reliable and accurate data from machines, components or products (i.e. performance measurement) through smart sensors (IoT applications) and wireless communication (cloud computing applications) (Lee et al., 2015). At the conversion level information must be extracted from the gathered data (Qin et al., 2016). This allows for self-comparison by the machines, deficiency detection and prediction of future performance (Lee et al., 2015). The conversion- and cyber level mainly focus on data analysis and validation if performance measures are met (i.e. performance reporting). Subsequently, the cognition level is concerned with presenting the information automatically in a proper manner (e.g. graphs, tables and figures) such that expert users can instantly understand the current (and future) state of performance to identify where improvement is required (Lee et al., 2015); thus, supporting both performance reporting and management. Automation in this process increases reporting efficiency as less resources are required, whereas a CPSs ability to objectively present the data enhances reporting effectiveness.
These relations are of key importance to this research as they show how smart manufacturing can contribute to PMM. They are summarized in table 2.2.
Table 2.2: 5C architecture related to PMM (Source: adapted from Qin et al., 2016)
The body of literature has demonstrated a range of advantages of the key smart manufacturing technologies which can be directly linked to a PMM phase. These findings are summarized in table 2.3.
Table 2.3: Smart manufacturing linked to PMM
The next section will discuss why SMEs generally have issues implementing PMM and/or smart manufacturing technologies.
2.4. SME limitations
introduced not much over a decade ago. As it is relatively new, not all companies have started implementing these technologies into their processes and adoption rate among SMEs remains low (Moeuf et al., 2018). The literature review shows that there is consensus among researchers as to why this is happening. These main barriers impairing PMM and smart manufacturing technology adoption in SMEs are as follows (Ates et al., 2013; Garengo et al., 2005; Pekkola
et al., 2016):
➢ Lack of managerial expertise or organizational skills to successfully guide the process ➢ Lack of resources (time, financial or human) which results in day-to-day operational
issues getting prioritized over long-term strategic objectives
➢ Reactive approach to direct issues instead of focusing on long-term competitiveness ➢ Difficulties in translating values, missions and visions into hard measures, which is
inherent to PMM
➢ Rapid strategic changes, resulting in out-of-date performance measures before they are effectively implemented
These limitations must be considered as this research is aimed specifically at SMEs. The next section will discuss the research question and the proposed conceptual model.
2.5. Research question
The body of literature has presented various advantages of smart manufacturing, which can potentially be beneficial for PMM. As next to no empirical evidence confirms these proposed relations (Hwang et al., 2017; Moeuf et al., 2018), the explorative research question (RQ) that this thesis aims to answer is:
“How can smart manufacturing contribute to both effectiveness and efficiency of PMM for
SMEs?”
(Zhong et al., 2017) are the third-party infrastructure (removing system maintenance for the company) and eliminating redundant systems (thus requiring less resources). As these are characteristics of cloud computing, and not key advantages specifically for PMM, there is a direct relation between cloud computing and performance reporting efficiency. Additionally, PMM is a sequential process (Nudurupati et al., 2011), and measurement and reporting are inherent to effective management (Radnor and Barnes, 2007). Table 2.3 briefly presented these proposed relations between smart manufacturing and PMM, which forms the basis for the conceptual model. However, the conceptual model is more precise as each key advantage is directly linked to a PMM criterion. It is most likely not exhaustive, as empirical evidence can raise additional advantages or disadvantages of smart manufacturing when combined with PMM.
Figure 2.8: Conceptual model with proposed relationships
3. Methodology
This chapter justifies the way this research was conducted. First, arguments for the research design are provided. Then, the cases used in the research will be elaborated upon. Further, an explanation on what data was collected and how this was done is provided. Additionally, interviewee characteristics are described. Then, a detailed description on how the data was processed and analyzed is given. Last, arguments to guarantee the validity and reliability of the research are provided.
3.1. Research design
Although the body of knowledge on PMM is extensive, this is not the case for smart manufacturing as it is a relatively new concept. Especially interesting is the influence of smart manufacturing on PMM, as next to no (empirical) research had been conducted on this topic (Hwang et al., 2017). Therefore, this research took an explorative approach. This was reflected in the RQ, which aimed to answer how smart manufacturing can influence PMM. As a ‘how’ question is open to interpretation (Karlsson, 2016), multiple respondents were interviewed, thereby investigating different perspectives. Due to the explorative nature of this research, case research was particularly suitable (Pinsonneault and Kraemer, 1993; Yin, 1994). Case studies allow a phenomenon to be studied in its natural setting which allows meaningful, relevant theory to be generated through observation (Karlsson, 2016). A relatively full understanding of the potential of smart manufacturing for PMM resulted after applying a case study approach. Since this phenomenon was not at all understood beforehand, a case study was considered appropriate (Karlsson, 2016).
conducted which had two main advantages: augment external validity and guard against observer bias (Karlsson, 2016). The next section will describe the cases used in this research.
3.2. Case description and selection
This research was done in cooperation with the Lean-QRM center of the Hogeschool van Arnhem en Nijmegen (HAN), which provided the cases. The SMEs were involved with the HAN in a project closely related to the subject of this thesis, which made them suitable cases. The research was conducted at six different, diverse companies. Of these companies three were manufacturing SMEs, two were (smart) IT-consultancies and one was a service-oriented SME. As the SMEs had yet to start adopting (or investigating) smart manufacturing technologies, consultancy companies were approached to provide their vision on the potential these technologies have for PMM. Although this is an indirect approach, it provides valuable insights as no research had been done on the link between smart manufacturing and PMM (Hwang et
al., 2017; Moeuf et al., 2018). The manufacturing and service-oriented SMEs provided insight
in their PMM activities. Each company had a different approach to PMM, and each approach came with its unique issues. PMM activities and issues were presented to the (smart) IT-consultancies which shared their vision on how smart manufacturing technologies can support PMM. The next section will discuss the exact process of collecting this data.
3.3. Data collection
This section justifies the way in which data was collected and outlines how the interviews were conducted.
3.3.1. Protocol
interviewee without any inconsistencies, redundancies or mistakes, which increased reliability and ensured that only relevant data was collected to allow for valid conclusions. The main advantage of a semi-structured interview as opposed to a structured interview is that the interviewee had the opportunity to share additional information that (s)he considered to be important and relevant. The researcher could ask detailed questions on subjects of interest raised by the interviewee. This reduced the risk of overlooking any information of significant importance. Triangulation is the act of bringing more than one source of data to back up a conclusion (Marshall and Rossman, 2014). This was partially realized by conducting interviews at six companies. Additionally, secondary sources for information were also used. This decreased the probability of misinterpretation of the data. Secondary data ranged from reports on historical and actual performance, and reports containing analyses of failures.
3.3.2. Interviews
Table 3.1: Interviewee summary
Next, the analysis of the interviews will be discussed in detail.
3.4. Data analysis
Table 3.2: Coding tree
3.5. Reliability and validity
4. Results
This section will present the results based on the data collection. First, a description of PMM activities per manufacturing SME case will be provided. Second, a cross-case analysis is given to highlight differences and similarities between the cases. Third, the main experienced PMM issues are highlighted. Fourth, the potential that interviewees see for smart manufacturing are presented per technology. Last, arguments as to why companies have not yet adopted smart manufacturing are provided.
4.1. Case findings
All interviewees agreed that a company should not be managed based on feeling. Generally, the companies had several ideas as to where performance could be improved. They acknowledged that proper data collection (performance measurement) and analysis/presentation of data (performance reporting) are inherent to acting and decision-making (performance management). Interviewee A2 states:
“We started PMM activities to remove management based on gut feeling. This is essential if
you strive for improvement.”
The PMM approaches found within the cases are, however, quite diverse. The next sections will describe per case how they have established PMM in their company and what their current PMM activities are.
4.1.1. Company A
absent no goal is given. In that case, they first measure what current performance is for a certain period and identify potential improvements afterwards.
Data collection is performed in different ways. Machine hours are registered manually, paper-based. Employees can register their working hours (semi-)automatically, they simply must select their current task and clock it. Other shop floor measures are generally done on paper or an employee manually inserts it in a computer. Afterwards, the production leader collects this data and inserts it into an ERP. All interviewees confirmed this raises questions regarding the accuracy of the data, as it is sensitive to errors. Data emerges from multiple sources and combining this to construct the data set is proven to be difficult. There are approximately 20 - 25 systems company wide. Management must manually extract this data and insert it into a spreadsheet. This results in significant time spent on combining, analyzing and reporting data to create the BSC. Afterwards, the reports are printed and displayed on performance boards throughout the company. This process is visualized in figure 4.1. Interviewee A1 argues:
“I have to spend at least 1 hour per week on this process due to the number of data-sources.
This could be done way more efficiently.”
Figure 4.1: PMM processes at company A
It is apparent to this company that some form of overarching system should be adopted to reduce time spent on this process. However, according to all six company interviewees, the management phase offers the greatest source of improvement. Interviewees from company management expressed a desire for showing performance prediction instead of historic data. Interviewee A3 elaborates:
“We can tell if the results of our measures are up to our standard or not. However, we are
On average, results are discussed two times per week at shop floor level. Out of the three shop floor employees (team leaders), two experience peer support for PMM. The other experiences very little support, as his/her peers feel that the discussed measures are irrelevant to them as they cannot directly influence them (e.g. turnover). Additionally, some people have been working at the plant for 20+ years, resulting in an interesting view on PMM as interviewee A5 describes:
“Those employees have always done it like this. They do not feel the need to change, they say
they are doing fine.”
4.1.2. Company B
For this company measures are partially derived from vision/strategy, which focusses on quality. However, low-margin products result in OEE being the dominant measure, to actively detect and eliminate waste. Interviewee B4 explains:
“Our vision is concerned with delivering high-quality products. However, at the end of the
day we need to pay our bills, so we actively monitor and manage our margins.”
This company has several plants, which make closely related products. Despite this, data collection occurs in various, non-similar methods at each plant. An example includes the measurement of machine shutdowns, which is paper-based in one plant but semi-automatically performed by a sensor at another. Interviewee B2 mentioned that the paper-based method is affected by both interpretational issues (not knowing what an employee means or how long a shutdown took) and time-lag issues (if the employee is asked what happened after a few days they generally do not remember the cause). The sensor increases accuracy of the data, as it can more precisely track shutdowns in terms of time.
Once all information is gathered it is sent to all stakeholders after approximately one week, after which it is printed and physically displayed where necessary. This process is visualized in figure 4.2. On shop floor level, OEE results are discussed daily before each shift. Subsequently, improvement actions/projects can be initiated where deemed appropriate. Multiple interviewees addressed examples in which an employee was assigned such an improvement project, which highly motivated them to succeed.
Figure 4.2: PMM processes at company B
4.1.3. Company C
Customer satisfaction is at the core of the company vision. All interviewees acknowledged that their main purpose is to deliver high quality products on time to their customers. However, the measures are not directly derived from this company vision. Each department has set measures according to their own desires, which results in a lack of uniformity. Additionally, PMM is not transparent as departments are not fully aware of each other’s goals, potentially leading to conflicting measures. Interviewee C5 explains:
“Operations, production, engineering, R&D and sales, they all have created their own
measures. I feel that PMM should be more in context of our vision.”
Subsequently, the data is used to generate performance reports semi-automatically in a spreadsheet, after which it will be printed and displayed on a performance board at the shop floor. The PMM process is depicted in figure 4.3. Interviewee C1 explains a crucial problem in this process:
“As each department has defined their own measure, our data is scattered. We have a lot of
data but are unable to analyze it.”
Figure 4.3: PMM processes at company C
Although a significant amount of data is available which is discussed on a weekly basis, concerns were raised about the fact that actions remain limited. All interviewees acknowledged the importance of PMM but admitted that not enough time is spent on analyzing data. They tend to focus on daily business, as stated by interviewee C2:
“Every week we print our results and display them on the shop floor for everybody to see.
Generally, that is it, no analysis follows.”
4.1.4. Cross-case analysis
An interesting result is that no company mentioned that their measures are changing over time. Once they have implemented a measure it generally will not be replaced by another. On the contrary, they all reported that their measures have been stable for a few years on average. At most, the measure goal changes (i.e. adjusting 20% increase in efficiency to 25%).
A common complaint in each company is the redundancy within their PMM processes. First, information must pass multiple employees before it can be processed into reports, as can be seen in figure 4.1, 4.2 and 4.3. Although some interviewees remarked that this keeps employees aware of their performance, others were worried about the potential errors that could be made by employees. The latter group would rather eliminate the human aspect from the data collection. Second, the data generally emerges from multiple sources and flows through multiple systems in each company: spreadsheets, databases, and ERPs. Company A has approximately 20 – 25 systems, whereas company C uses multiple spreadsheets in addition to their ERP system. Company B experiences a similar issue according to interviewee B3:
“Our ERP is not suitable for visualizing data. We use a spreadsheet for visualization, while
the other plant uses their own database. If we want to exchange data, we often must manually transmit it into our own spreadsheet. This is inefficient.”
Furthermore, interviewees unanimously agreed that simply being aware of performance is vital to improvement. Generally, employees tend to think they are doing fine, and do not necessarily feel the need to measure performance. However, interviewee A5 confirms the importance of measuring:
“We used to spend around 25% of our time on non-value adding tasks. After we were aware of this as a result of measuring performance, we initiated improvement actions and now we
are down to 11%. We even exceeded our initial goal of 15%.”
This reality check is necessary to start investigating where potential improvements can be made to performance.
4.2. Experienced issues
experience when adopting and/or performing PMM activities, limiting both PMM effectiveness and efficiency. This research reveals the eight most prevalent issues. Smart manufacturing might be especially interesting in (partially) solving these issues, as will be addressed in the next chapter.
4.2.1. Loss of focus
One of the main concerns regarding PMM is loss of focus. Generally, companies have a meeting once every one to two weeks in which business performance is discussed. At such meetings the most important measures are discussed after which improvement plans are initiated, if necessary. Although in theory this works well, execution can still be improved. Interviewee C5 elaborates:
“Once we discuss a measure during a meeting everyone agrees on its importance. After the meeting, they tend to be forgotten and business as usual continues.”
Thus, employees tend to focus on their core tasks and focus less on (performance) numbers. This is exemplified in figure 4.4, which shows data on the amount of ‘proposals to change a unit’. These proposals have been submitted by employees. Interviewee C1 stated that employees acknowledge it is important to process these changes, but it tends to be forgotten due to daily business which resulted in the increase of changes not handled in time.
4.2.2. No added value
It is of fundamental importance to have shop floor employee support for PMM. They tend to be a great source of data (input for PMM); if they do not support PMM for any reason it will significantly lose impact in terms of effectiveness. One of the major causes of lacking support is the idea that it does not add any value. Interviewee C1 explains:
“My peers complained that they have performed very well at their job for over 20 years and they are doing fine.”
To which interviewee A5 adds:
“The time I have to waste on registering data I could also spend on producing.”
All case companies experienced this in the early stages of implementing PMM. Generally, after approximately 6 – 12 months most employees have adapted to the new way of working (including data collection). Then, only a last few non-cooperating employees remain, usually those who have been employed for decades.
4.2.3. Data interpretation
Another major disruptor of PMM that came to light is concerned with the reporting phase. Presentation of the data is of key importance for people to understand their performance. If they do not comprehend what is presented to them, discussions arise on how to interpret the data. This can lead to different conclusions and subsequently problems in decision-making, as illustrated by interviewee C1:
“Demand for product X dropped abruptly to 0. This lowered our turnover by 10%*. We saw a relation between turnover and workforce, therefore we cut our workforce by a similar amount. Turned out, approximately 1%* of our workforce was responsible for that 10%*
turnover, so our decision was far from optimal. We had the right data, but we did not comprehend it fully.” (*Changed due to confidentiality)
4.2.4. Measure influence
Another aspect that heavily contributes to shop floor employee support is the type of measures that are presented to these employees. If they are not relevant to them, employees generally do not care and will not pay attention to the information, as interviewee A6 describes:
“We are shown the material throughput time every week. We cannot influence that in any way, so why should we care?”
This is also influenced by the depth of the information. If the shown measures are company- or departmental wide, the individual contribution of the employee will be limited, resulting in limited commitment to the measure. However, when measures are shown that are relevant and can be influenced by their individual performance it tends to motivate employees and increase commitment to PMM. Interviewee C3 explains:
“We have some employees that turn their performance into a game. Each day they try to get faster and better than each other”
4.2.5. Management support
In accordance with literature, management support proves to be of key importance to successful PMM. Interviewee B3 provides an example:
“We emphasize the importance of quality. However, our company focused too much on producing as much as possible. I started measuring quality and doing test to improve it. This
process must be continued to ensure constant high-quality products. However, after I left, management stopped focusing on this. Such a shame, because it took a lot of effort to get it up
to this level.”
Interviewee C5 adds another example:
“If it concerns a family-owned company with an owner-director who thinks he can manage the entire process based on feeling, how are we supposed to change that?”
4.2.6. Lack of resources
General concerns regarding time, people or money are prevalent in the case companies. This can range from focusing on short-term, daily business (time), not having skilled employees (people) or simply lacking financial resources (money). Interviewee B1 elaborates:
“I am responsible for continuous improvement, but also managing operations. This makes
time an issue, next to money which is always a problem.”
4.2.7. Lack of feedback
Communication in a company is important. Management expects data collection from shop floor employees, but in turn they should be provided with feedback. In the literature this is mentioned as a crucial enabler, else PMM will fail. This is confirmed by all interviewees and exemplified by interviewee C2:
“We asked employees to fill out forms when they had to spend additional time on a product. We also wanted them to write down all improvement ideas they had, errors they saw. We did utilize that data, but more on the background. Then people stopped filling out those forms as
they said: ‘we keep filling out a lot of forms, but never hear anything’.”
Interviewee B4 adds:
“Last week I heard that employees stopped collecting data on machine shutdowns. They did not hear or see anything about it. Weeks later it turned out we needed that information.”
4.2.8. Time lag
Time between data collection and presentation is on average approximately one week at the companies. Some issues were raised, as this time lag is a potential source for errors. Generally, paper-based data collection was handed to a production leader at the end of a shift. Whenever there is a gap or an error in the data, the employee needs to justify this. This can be problematic, as illustrated by interviewee B2:
“When we are unsure what the operator means we will need a clarification. However, they are going home immediately after handing in the data. If we then ask them after a week what
This corrupts the data which limits its use for analyzing. Desire for more up-to-date data was expressed by most interviewees. The results are summarized in table 4.1, together with how many interviewees mentioned the issues.
Table 4.1: Mentions of main PMM issues
The next section will discuss how smart manufacturing can potentially contribute to PMM.
4.3. Smart manufacturing and PMM
It became apparent that smart manufacturing technology adoption is limited. Two out of four SMEs (including the service-oriented SME) have adopted cloud computing. IoT and CPSs have not yet been implemented. This section will present the potential impact IoT and cloud computing can have based on interviewee perspectives, including arguments as to why some will not embrace cloud computing. Last, the reasons for absence of CPSs will be outlined.
4.3.1. Internet-of-Things
Out of 18 interviewees, 14 expressed concern regarding the accuracy of data collection. Some with regards to accidental mistakes like registering their hours (clocking) which is prone to errors due to employees forgetting to register. Other concerns were raised because of intentional mistakes potentially being made. Interviewee B3 elaborates:
data from sources that cannot be influenced, like sensors. This way of thinking results partially from trust in machines and partially from distrusting people.”
This captures the essence of what IoT can do: collect objective, trustworthy data. Interviewee D1, who has had significant experience with IoT, confirms:
“If you want to have mistakes with IoT, you would have to actively sabotage the system.”
As objectivity of data is no longer questioned, accuracy will be ensured, hence decision-making will be based on correct data. As mentioned in the previous section, time lag is a serious concern to 10 out of 18 interviewees. This was exemplified by interviewee B2, who stated that operators sometimes had to justify shutdowns a week later, after which they generally had forgotten what happened, which poses problems for data analysis. IoT (e.g. RFID chips and smart sensors) can automatically transmit data instantly via the internet. Thus, time between non-performance and performance reporting is reduced significantly. This increases the odds of e.g. preventive maintenance, avoiding serious breakdowns which would significantly reduce performance. Interviewee E1 explains:
“Once you have implemented smart sensors you can link them to your measures. These sensors will then provide real-time data, allowing you to live-track your performance. They
are programmed to speak the same language, allowing interchangeability of data.”
percentage was idle time or breakdown time, but not what caused non-performance. Interviewee E1 states:
“IoT removes a significant part of the programming and mechanical installation, while allowing you to extract more information from your machines.”
Now, companies must investigate what happened when they see performance drop (e.g. OEE), which can be a time-consuming and inefficient process. With IoT, the root-cause can easily be identified as e.g. smart sensors can verify what happened. Examples include change in temperature, vibrations, pressure or part failure. Thus, the level of detail of data (transparency) increases, allowing management to quickly verify if machines are still performing as desired, thereby eliminating the necessity of investigating. If, for example, they monitor that a part is overheating, it allows for preventive maintenance. This will reduce the likelihood of machine failure, thereby keeping performance up to standard. Table 4.2 summarizes these advantages and how many interviewees saw the same potential.
Table 4.2: Mentions of IoT advantages
4.3.2. Cloud computing
Out of four SMEs (including the service-oriented SME), two have adopted cloud computing. The main argument for not adopting cloud computing is the lack of necessity. When confronted with the possibilities, most responses were in line with that of interviewee D1:
“Most of my clients have internal servers. Whether you save your data in those servers or
Interviewee C5 adds:
“Hard to comprehend. All our data in the cloud? Seems vulnerable. Now I know our data
resides in our servers which no one else can access, I can lock the room with a key. I can believe the cloud offers opportunities, but I do not think we will follow that trend.”
On the other hand, those who adopted it did not mention any downsides. Both companies stated cybersecurity is important but had no negative feeling towards the cloud. In fact, interviewee F1 states:
“We started adopting the cloud to increase availability of our data. We have around 40
employees on a different site who can still access our own system through the cloud, which is convenient.”
This saves a considerable amount of time, as employees first had to print reports and physically bring them to the other site, which was inefficient. This sometimes resulted in forgetting data, which would then limit their ability to fully comprehend performance data. Interviewee B4 elaborates:
“Because we have access to the data of the other plant, we can easily benchmark
performance. We can see the issues (e.g. machine breakdown) they dealt with. If we experience something similar, we can see how they solved it and apply that solution too.”
4.3.3. CPSs
The results for IoT are positive, whereas cloud computing still receives mixed responses. As for CPSs, the same cannot be said. What is apparent is that the companies do not see any use for this complex set of technologies. Interviewee B3 states:
“SMEs simply do not have the financial resources for CPSs.”
Interviewee D1 summarizes the main problem with CPSs for SMEs:
“I have not yet seen any examples of this. I think large companies must invest in this first, incorporate CPSs in their processes. The technology will eventually become cheaper, which
will allow smaller companies to start looking into the possibilities.”
It appears that CPSs are not viable for SMEs in the next few years. Interviewee E1 adds:
“I think systems will not take over decision-making any time soon.”
4.4. Smart manufacturing adoption in SMEs
As far as CPSs concerned, it is apparent why SMEs have not started to adopt this technology: because of financial constraints they have not explored the potential. Currently, no interviewee thought that CPSs will be useful in their processes. Cloud computing adoption partially depends on cybersecurity and absence of necessity. But why have companies not yet started to implement IoT if they see the potential it offers? The answer to that question is two-sided: focus on internal processes and lack of knowledge.
4.4.1. Focus on internal processes
Automation of data collection is generally the first advantage that comes to mind when discussing IoT. Currently, SMEs are still struggling with the PMM issues discussed earlier this chapter. Before automating this, they tend to focus on optimizing their PMM processes. Interviewee E1 explains:
“We need to thoroughly understand our processes. The past 70 years we focused on craftmanship. We will need to analyze our processes, but then we need to know what to
Out of 18 interviewees, 14 agreed that they first need to fully comprehend their processes. Thorough understanding of what to measure, how to analyze and how to utilize this data is key before automation. Interviewee C1 states:
“We could automate the process of data collection with such technologies. But if I am collecting the wrong data, why should I want to automate this? Then I just end up with more
useless data.”
These are valid questions, which will be addressed in the discussion. First, another reason for limited IoT adoption will be addressed.
4.4.2. Lack of knowledge
Most interviewees could only describe the potential they see for IoT after elaboration on the subject. Out of 16 SME interviewees, five were familiar with IoT. Unfamiliarity with IoT became more apparent after consulting interviewee D1:
“We serve more than 400 companies. As ambassadors of smart manufacturing, we attend related conferences. We present some of the few examples that exist, after which many people come to us and ask: ‘can we use that technology too?’. Most of the companies are completely
unaware of the existence and/or possibilities.”
Unawareness plays a significant role in the limited adoption of IoT. However, it also brings with it another issue. To most, IoT remains a futuristic concept; they presume it to be costly. This common misconception is refuted by interviewee D1:
“I have seen just a few SMEs with IoT (e.g. RFID). Cost did not play a significant role; the technology has become a lot cheaper over the last few years. Especially when you compare it
to failure cost: IoT is a very efficient way to collect data and significantly reduces failures.”
