THE CHALLENGE OF NEAR-MISS IDENTIFICATION:
AN EXPLORATORY INVESTIGATION OF SUPPLY CHAIN NEAR-MISSES
Master thesis, MSc. Supply Chain Management
The University of Groningen, Faculty of Economics and Business 29. January 2017
LORENA POLLE Student number: 3183319 E-Mail: L.A.J.Polle@student.rug.nl
Supervisor/ university K. Scholten Co-assessor/ university
X. Tong
Acknowledgement
First, I would like to express my very profound gratitude to my supervisor Dr. Kirstin Scholten for always giving me constructive feedback, for her patience and directive guidance throughout the whole thesis process. The thesis could of course not have been realized without the support of the company Teijin Aramid. I would like to especially thank Mr. Boris Fenneman for answering all our questions and giving us advice for improvements. Next, a big thank you to all the interviewees for their time and effort participating in this study. Finally, I thank my family and friends for providing me with unfailing support and continuous encouragement.
Abstract
Purpose: This paper addresses the challenge of how to identify supply chain near-misses. Near- misses provide useful lessons to eliminate risks before they result in disruptions, however discovery is difficult because of their limited impact.
Method/Design: An exploratory interview approach in the chemical industry based on 12 interviews and 19 near-miss incidents has been chosen to examine how to identify near-misses.
By means of coding qualitative data different discovery strategies have been found.
Findings: This study finds that event characteristics such as the speed of risk and the probability of occurrence are triggering recognition of a near-miss incident. Five discovery strategies based on visibility make both factors apparent. However, risk management culture influences the identification indirectly.
Implications: This paper provides insights into the identification of near-misses in the supply chain setting. A series of propositions explains mechanisms of visibility and risk management culture on the near-miss identification.
Keywords: Near-miss identification, Supply chain near-misses, Discovery strategies CONTENT
1. INTRODUCTION ... 4
2. THEORETICAL BACKGROUND ... 6
2.1 The Near-Miss Identification... 6
2.2 Disruption Event Management ... 8
2.3 Discovery Concepts ... 9
2.4 Visibility ... 11
2.5 Risk Management Culture ... 12
3. CONCEPTUAL MODEL ... 15
4. METHODOLOGY ... 16
4.1 Research Design ... 16
4.2 Setting Selection ... 16
4.3 Case Selection ... 17
4.4 Data Collection ... 18
4.5 Data Analysis ... 20
5. RESULTS ... 25
5.1 Overview of Findings ... 25
5.2 Findings on Visibility ... 25
5.3 Findings on Risk Management Culture ... 30
6. DISCUSSION ... 33
6.1 The Influence of Visibility on Near-Miss Identification ... 33
6.2 The Influence of Risk Management Culture on Near-Miss Identification ... 36
7. CONCLUSION ... 38
7.1 Managerial Implications ... 38
7.2 Limitations ... 39
7.3 Future Research ... 39
8. REFERENCES ... 41
APPENDIX A ... 46
APPENDIX B ... 53
LIST OF TABLES Table 1: Near-Miss Definitions of Different Sectors ... 7
Table 2: Discovery Concepts in Literature ... 10
Table 3: Overview of Company and Interviewee Characeristics ... 17
Table 4: Overview of Case Characteristics ... 18
Table 5: Validity and Reliability of this Research ... 20
Table 6: Risk Management Culture and Support Systems of Case 1-19... 22
Table 7: Operationalization of Constructs ... 23
Table 8: Example Second-order-codes ... 24
Table 9: Example Third-order-codes (Discovery Strategies) ... 24
Table 10: Example Third-order-codes (Event Characteristics) ... 25
Table 11: Overview of Findings ... 26
Table 12: Overview of Discovery Strategies Based on Visibility, Support Systems, and Mechanisms in the Near-Miss Identification ... 30
Table 13: The Influence of Risk Management Culture on Visibility ... 21
LIST OF FIGURES Figure 1: Conceptual Model ... 15
Figure 2: The Influence of Visibility and Risk Management Culture on Near-Miss Identification ... 33
THE CHALLENGE OF NEAR-MISS IDENTIFICATION:
AN EXPLORATORY INVESTIGATION OF SUPPLY CHAIN NEAR-MISSES 1. INTRODUCTION
Supply chains seem to be more fragile than ever before, as supply chain disruptions become more frequent and severe (Wagner & Neshat, 2010). Therefore, in addition to highlighting the need of managing disruptive events, researchers and practitioners have stressed the need for dynamic and integrated tools to assess and manage risks (Handfield, Blackhurst, Elkins, &
Craighead, 2007). One tool that is frequently used in hazardous domains, such as the nuclear- or chemical industries to prevent accidents, is a near-miss management system (Gnoni & Saleh, 2017). Near-misses are hazardous situations that could have led to a catastrophic event, but ultimately did not (Dillon & Tinsley, 2008). Via observation of these situations, areas fraught with risk can be identified and eliminated (Dillon, Tinsley, Madsen, & Rogers, 2016). Doing so is especially significant, since learning from near-misses result in fewer costs than learning from supply chain disruptions (Gnoni & Saleh, 2017). However, near-misses often remain underreported as a result of the misrecognition of these events (Baron & Hershey, 1988 IN Dillon et al., 2016).
Near misses are more difficult to detect than supply chain disruptions, because they are characterized by a limited or by no impact on the environment, processes or people. (Phimister, Oktem, Kleindorfer, & Kunreuther, 2003). However, since risk sources that can lead to a near- miss or a supply chain disruption are identical, identification might also occur in a similar manner (Wright & Van Der Schaaf, 2004). Therefore, this paper applies two concepts that were found to help in the discovery of disruptions, namely, visibility and risk management culture to the identification of near-misses. Visibility within a supply chain can disclose the location of resources and risks, as well as how disruptions are diffused throughout the network (Blackhurst, Dunn, & Craighead, 2011). A risk management culture, however, promotes the application of
proactive risk measures. (Lin & Wang, 2011; Schmitt,2011 IN Durach, Wieland, & Machuca, 2015). Companies with such cultures are more likely to take actions that track near-misses.
Researchers have been searching for information about factors that enhance recognition and attention of near-misses in safety management (Dillon et al., 2016). So far, factors and methods that improve near-miss recognition have been also insufficiently studied in supply chain literature. Although it is known that visibility and a risk management culture can help to discover disruptions, it is not clear if this also applies to near-misses and how this exactly functions. Based on this, the research question of this study is as follows:
“How to identify supply chain near-misses in the chemical process industry?”
An exploratory case study examining 19 near-miss incidents was undertaken to achieve the research objective of this thesis. Semi-structured interviews provide the data upon which the findings of this study are based. This paper contributes to theory and practice by determining two event characteristics: the probability of an event occurring and the speed of risk that influence the recognition of near-misses. Furthermore, based on the concept of visibility, this thesis lists five discovery strategies as a broad guideline for identifying near-misses and explains their underlying mechanisms. Additionally, it finds risk management culture to be an antecedent of visibility. This knowledge could enhance near-miss reporting, resulting in risky areas being revealed and eliminated before they can trigger a disruption.
This paper is structured as follows: Section 2 presents the theoretical background, which begins by describing the near-miss identification process. Next, the phases of a supply chain disruption are explained and concepts are presented, that help to detect near-misses or disruptions. The concepts serve as a starting point for answering the research question. Section 3 explains the conceptual model, while Section 4 describes the methodology and Section 5 the results. Section 6 elaborates on the results through a discussion, and Section 7 provides the conclusion.
2. THEORETICAL BACKGROUND 2.1 The Near-Miss Identification
A near-miss event must be identified, analyzed and actions have to be taken to prevent its reoccurrence (Phimister et al., 2003). The term “near- miss identification” can be defined as:
“The perception of a hazardous event that has the potential to result in a supply chain disruption” (Phimister et al., 2003; Tinsley, Dillon-Merrill, & Cronin, 2012; Tinsley, Dillon,
& Madsen, 2011). This paper uses the terms “discovery” and “identification” interchangeably.
One barrier in near-miss identification is that near-misses do not affect the performance of the company (Dillon & Tinsley, 2016; Phimister et al., 2003). In addition, in contrast to steep decline, a slow deterioration of conditions makes it difficult for near-misses to be recognized (R. L. Dillon, Rogers, Oberhettinger, & Tinsley, 2016). For this reason, organizations tend to overlook them. A third barrier to near miss identification were found to be supply chain complexity, non-transparent relationships and loosely coupled information (Skilton &
Robinson, 2009). Due to presence of these factors, Skilton and Robinson (2009) suggested in relatively tightly coupled networks to reduce complexity and allocate an enforcement role to central supply chain members to enable traceability of near-misses.
A fourth barrier in reporting near-misses is confusion about the concept and its definition. Such ambiguity must be resolved for identification to be possible (Phimister et al., 2003).
Kleindorfer, Oktem, Pariyani, and Seidner (2012) offered two recommendations aimed at achieving a more successful identification process: 1) agreeing on a clear definition of a near- miss and 2) using instruments to ensure that employees are aware of that definition at all times.
At present, there is no clear definition of a near-miss in supply chain literature. In line with Gnoni and Saleh (2017), who supported the idea of different definitions tailored of different sectors, this study employs its own definition for a supply chain near-miss. Therefore, near-
miss definitions from other sectors are compared to identify patterns. Table 1 provides an overview of common near-miss definitions across different sectors.
TABLE 1
Near-Miss Definitions of Different Sectors
Sector Definition Source
Process Industry
“A situation where the sequence of events could have caused an accident, if it had not been interrupted”
Brazier (1994: 244)
Management
“More precisely, a near-miss is an event that has some non-trivial expectation of ending in disaster, but because of luck did not.”
Dillon-Merrill, Tinsley,
& Cronin (2012: 1) Chemical Industry “Industrial near-misses are defined as events
with minimal consequences.”
Kleindorfer et al.
(2012: 2) Healthcare
“[W]hen an event could have happened (for example, because of hazardous conditions) but did not.”
Dillon & Tinsley (2008: 2)
Process Industry
“[W]here an accident could have happened had there been no timely and effective recovery”
Van Der Schaaf &
Kanse (2004: 69)
Automotive
“[A] hazardous situation where the event sequence could have led to an accident if it had not been interrupted by a planned intervention or by a random event”
Andriulo & Gnoni (2014: 154)
Healthcare
“[E]vents that could have harmed patients, but did not due to a timely intervention or a convenient evolution of the circumstances”
Cure, Zayas-Castro, &
Fabri (2011: 738)
A comparison of these definitions in terms of similarities revealed that two characteristics seemed to be highly common: 1) an emphasis of a near-miss being an event or situation and 2) the avoidance of failure through either human action or chance.
Within the supply chain context, near-misses could have led to a supply chain disruption instead of an accident. Since supply chain disruptions are characterized by a break in the flow of goods and materials resulting in non-delivery for the customer (Blackhurst, Craighead, Rungtusanatham, & Handfield, 2007), this thesis proposes the following definition of a supply chain near-miss:
A supply chain near-miss is a hazardous event or situation that could have disrupted the flow of goods and materials, resulting in non-delivery for the customer, but which was interrupted by a random event or planned intervention from either the company or a supply chain partner.
Supply chain literature does sufficiently specify, how to identify near-misses in the supply chain. However, Wright and van der Schaaf (2004) found that the causal pathways of near-miss incidents and major accidents are similar, which might also apply to supply chain near-misses and disruptions. Accordingly, disruption discovery might yield more insights into how to identify near-misses. The following section presents disruption event management, and especially discusses its identification.
2.2 Disruption Event Management
In contrast to near-misses, supply chain disruptions can result in immediate or delayed negative consequences regarding the operations and performance of a company (Blackhurst et al., 2011;
Sheffi & Rice, 2005). Blackhurst, Craighead, Elkins, and Handfield (2005) separated the management of disruptions into three phases, namely, 1) disruption discovery, 2) disruption recovery and 3) supply chain redesign. “Discovery” is defined as the point in time, at which people become alerted to a disruptive event in the supply chain (Macdonald & Corsi, 2013).
Manuj and Mentzer (2008) argued that the speed of risk is one factor that determines how fast a disruption is discovered. The term “speed of risk” refers to the rate at which loss occurs. The more rapidly a supply chain disruption is detected and communicated, the more time is available to take preventive measures to reduce its severity (Blackhurst et al., 2007). Accordingly, a timely discovery might determine whether an event results in a near-miss or an actual disruption. A study conducted by Macdonald and Corsi (2013), found three ways of discovering a disruption, namely, the conventional mode, the adventitious mode and the proactive mode.
Although Mcdonald and Corsi (2013) provided examples for each mode, they did not further elaborate on them. Therefore, this paper creates its own definitions based on their work where
the conventional method is defined as “the perception of a hazardous event by receiving information from the supplier or customer,” the adventitious mode is defined as “the perception of a hazardous event by chance” and the proactive mode is defined as “the perception of a hazardous event by a pre-established identification process”. The three discovery methods of Macdonald & Corsi (2013) might also apply to near-miss identification. The following section, introduces two concepts, that help support the three discovery methods.
2.3 Discovery Concepts
This study found that two main elements, visibility and a risk management culture, enhance detection abilities of companies and thereby support the three discovery modes. An overview of both concepts, outlined by supply chain literature, is presented in Table 2. The most frequently mentioned antecedents of these concepts were categorized and further investigated.
Nonetheless, the literature named other factors that facilitate disruption and near-miss discovery. For example, collaboration has been found to have a positive effect on supply chain visibility through information sharing, the mutual creation of knowledge and joint relationship efforts (Scholten & Schilder, 2015). However, since the effects of collaboration on discovery are less direct than those of visibility and risk management culture, this study focused only on these latter two factors.
TABLE 2
Discovery Concepts in Literature
Discovery concepts Operationalization Sheffi & Rice (2005)
Blackhurst et al. (2005)
Christopher
& Peck (2004)
Jüttner &
Maklan (2011)
Macdonald &
Corsi (2013)
Brandon et al.
(2014)
Chowdhurry
& Quaddus (2016)
Handfield et al.
(2007)
Visibility
Information sharing Information sharing
Information sharing
Information
sharing Information
sharing
Control systems
Predictive analysis (risk management
tools)
Discovery method
Connectivity (resources &
infrastructure)
Visibility
systems
Warning capabilities
Capacity (knowledge
about bottlenecks)
Risk management culture
management Knowledge
Risk awareness Risk sharing
effort
Situational
awareness
Risk assessment in
decision making
Decision to
communicate
Risk consideration
in decision- making
Leadership
commitment Leadership Leadership
Risk management teams
Empowering front-line employees
Risk management
team
SC- continuity
team
Note: Studies of Sheffi and Rice (2005), Blackhurst et al. (2005), Handfield et al. (2007) and Macdonald and Corsi (2013) contain a strong focus on the discovery of supply chain disruptions.
Studies of Jüttner and Maklan (2011), Chowdhurry and Quaddus (2016), Christopher and Peck (2004) and Brandon et al. (2014) use these concepts, however, their main focus is not on discovery.
2.4 Visibility
Since speed is an important factor in the discovery process (Macdonald & Corsi, 2013), a high level of information visibility reduces the time between the occurrence of a hazardous event and its discovery (Handfield et al., 2007). Consequently, visibility supports both, the conventional mode and the proactive mode of discovery. “Visibility” is defined as “the extent to which actors within the supply chain have access to or share timely information about supply chain operations, other actors and management which they consider as being key or useful to their operations” (Jüttner & Maklan, 2011: 251). It refers to the ability to review the entire supply chain including locations of resources, risks and the way in which disruptions are diffused throughout the network (Blackhurst et al., 2011; Christopher & Peck, 2004). This study operationalized visibility into information sharing and warning capabilities. This operationalization is in line with Brandon-Jones, Squire, Autry, and Petersen (2014), who described visibility as an outcome of information sharing and information technological (IT) infrastructure.
“Information sharing” refers to “the extent to which a firm shares a variety of relevant, accurate, complete and confidential ideas, plans, and procedures with its supply chain partners in a timely manner” (Scholten & Schilder, 2015: 251). Consequently, it can support the conventional discovery mode of disruptions and near-misses (Macdonald & Corsi, 2013).Van der Vorst and Beulen (2002) argued that uncertainty emerges from a lack of accurate and up- to-date information. The exchange of information between supply chain members can reduce this uncertainty (Christopher & Peck, 2004). Such interactions consequently lead to a more accurate perception of a situation, and therefore, a more rapid discovery of near-misses or disruptions.
However, one common barrier to information sharing is a lack of internal structures within a company (Christopher & Peck, 2004). In order to share information, reduce uncertainty, and in
turn, increase visibility, organizations must create linkages within the supply chain (Mabert &
Venkataramanan, 1998). Warning capabilities can enable the development of these connections. This paper defines warning capabilities in a similar way as the previous work of Brandon-Jones et al. (2014) and Blackhurst et al. (2007) as “an organization’s technological resources, processes, and competencies to gather, share and analyze risk-related information.”
A visibility system can constitute such a resource, as it enables quick detection of a disruption or near-miss (Sheffi & Rice, 2005). Through tracking and tracing inventory, demand, or capacity levels, a company can acquire and share real-time information, which results in awareness when a problem emerges (Handfield et al., 2007). However, modern visibility systems not only detect disruptions or near-misses reactively, but also feature predictive analysis tools that proactively forecast them (Handfield et al., 2007). Warning capabilities thus allow a company to discern an impending or already realized hazardous event and to convey the information to relevant entities within the supply chain (Blackhurst et al., 2007).
Consequently, warning capabilities can support both discovery modes, the conventional mode and the proactive mode of discovery (Macdonald & Corsi, 2013).
2.5 Risk Management Culture
Although the aforementioned warning capabilities present an important factor in the proactive and conventional discovery, Fawcett et al. (2007) emphasized the importance of company culture, which influences an organization’s willingness to share information. In particular, identifying of a hazardous event requires a supportive culture that is sensitive to risk-related information (Sheffi & Rice, 2005). The term “culture” indicates a system within an organization of shared norms and values, as well as a number of familiar practices (Reichers &
Schneider,1990 IN van Dyck, Frese, Baer, & Sonnentag, 2005). A “risk management culture”
is defined in this work based on work of Chowdhurry & Quaddus (2016) as “an environment that helps to facilitate the implementation of risk management measures into the organizational
practices of a company“. In line with prior research by Christopher & Peck (2004), this study operationalized a risk management culture as risk management teams, leadership representation and risk awareness in decision-making.
Blackhurst et al. (2005) particularly stressed the importance of risk awareness in the daily operations of a company. “Risk awareness” refers to the inclusion of risk assessment as a formal part of the decision-making process at every level of an organization (Christopher & Peck, 2004). Employees evaluate how several aspects of their decision-making process could lead to a disruption or a near-miss (Dillon et al., 2016) and in doing so they are encouraged to take actions aimed at managing risk. Additionally, Macdonald & Corsi (2013) found that one factor that influences the discovery of a disruption is the initial decision to communicate the event internally. An employee who takes risk-related information more seriously will be more likely to share such information. Such communication can support the conventional mode, the adventitious mode and the proactive mode.
Multiple researchers have emphasized leadership as an antecedent for implementing a risk management culture (Christopher & Peck, 2004) since the strategic initiatives of top management are the base for the execution of common goals (Speier et al., 2011 IN Durach et al., 2015). A top manager can consequently support the implementation of the aforementioned warning capabilities. Doing so results in more attention paid to vulnerabilities in the supply chain, which in turn might increase the likelihood of a disruption or near- miss being discovered both conventionally and proactively. For this reason, Christopher and Peck (2004) recommended having a supply chain member on a company’s board of directors to increase the focus on supply chain risks.
Several authors have stressed the need for cross-functional risk management teams that audit risks using the aforementioned tools and systems (Chowdhury & Quaddus, 2016; Christopher
& Peck, 2004). Firms with special teams that frequently monitor and assess their environments will recognize an hazardous event earlier approaching than firms, waiting to observe a significant performance drop (Bode & Macdonald, 2017). Since cross-functional teams create formal linkages between departments, they allow team members to more easily share information (Mcdermott, 1999; Pagell, 2004), which will consequently improve the visibility within a company. Accordingly, cross functional teams might improve a proactive discovery mode.
3. CONCEPTUAL MODEL
Although several authors stressed the value of managing near-misses (Peck, 2005; Sheffi &
Rice, 2005), there are insufficient studies on near-miss identification in supply chain management literature. This study will seek to gain knowledge about this process by using two concepts that enhance the discovery of disruptions or near-misses, namely visibility and risk management culture as a starting point and applying it on the three discovery modes. The conceptual model in figure 1 displays assumed relationships between the variables.
FIGURE 1 Conceptual Model
4. METHODOLOGY 4.1 Research Design
To investigate how near-misses can be identified, an exploratory case study design was selected. The process of near-miss identification has hardly been investigated in the past.
Therefore, the use of a case study is highly relevant for this research as it stands out due to its ability to explore unfamiliar situations and develop new theories (Meredith, 1998). Since risk management culture and visibility vary between different environments, case study is the most appropriate approach for this research. The case study approach enables observations of different surroundings and their underlying mechanisms (McCutcheon & Meredith, 1993). This study strives to gain new insights regarding near-miss identification that can be generalized to help managers report near-miss incidents. To that end, multiple interviews were conducted.
Utilizing multiple interviews creates more robust theories and results. Moreover, they are more generalizable due to possibility to compare across interviews (Eisenhardt & Graebner, 2007).
Based on the research objective, near-miss incidents in the supply chain were determined as the unit of analysis.
4.2 Setting Selection
The chosen setting for this study incorporated medium and large companies, working in the chemical process industry. This setting is especially appropriate because work in this sector often involves operations with highly sensitive materials in terms of security, which according to Giunipero and Eltantawy (2004) requires more extensive risk management. Hence, these companies might have a culture of risk management or may invest in technological resources to gather risk-related information. Additionally, in hazardous domains with stable and predictable operations (e.g. chemical process plants), an enormous emphasis is placed on process control (Reason, 2004), which facilitates the recognition of deviations in the operations.
Since Thia, Chain, Bauly, and Xin (2005) found that the adoption of quality control systems
partially depends on the culture of a company, firms with different risk management cultures provide their staff with different warning capabilities to identify near-misses. It is for this reason that companies with a variety of risk management cultures were selected. To derive the level of risk management culture, interviewees have been asked via questionnaires about the presence of 1) a risk management focus in the board of directors 2) cross-functional risk management teams and 3) risk awareness in the decision making in their companies. In case all three characteristics are present, a company is labeled with a “high” risk management culture. Table 3 shows an overview of the characteristics of the selected companies and interviewees.
TABLE 3
Overview of Company and Interviewee Characteristics
Company Sector Employees Operating area
Risk management
culture
Interview length (H: M: S)
A Consumer chemicals 22000 Global Medium 00:33:58
B Machinery for Chemistry 2000 Global High 00:52:02
C Machinery for Chemistry 75.000 Global High 01:19:45
D Consumer chemicals 1.314 Global Low 00:35:17
E Chemistry 4.000 Global High 00:58:45
F Chemistry 3.800 Global Medium 01:08:45
G Machinery and Packing
for chemicals 20.000 Global High 01:02:20
H Machinery and Packing
for chemicals 20.000 Global Low 01:31:19
I Chemistry 56.000 Global Medium 01:02:20
J Chemistry 1.400 Global Low 00:56: 00
K Chemistry 8.300 Global High 00:56:29
L Chemistry 20.786 Global High 00:55:44
4.3 Case Selection
Nineteen near-miss incidents were selected based on theoretical replication. Accordingly, cases must be heterogeneous in terms of the characteristics, they exhibit. Cases were selected regarding the tools (warning capabilities), the company uses in order to identify the near-miss.
Whereas Sheffi and Rice (2005) stress the usage of visibility systems, Scholten and Schilder
(2015) mention telephone usage as a system that can help to discover disruptions. It is expected that different systems or the absence of systems reveal the near-miss in different ways. Cases 1-19 and their support systems are displayed in Table 4.
TABLE 4
Overview of Case Characteristics
Cases Company Support
Systems
Case 1 A ERP-system
Case 2 A Email
Case 3 B ERP-system
Case 4 C Telephone
Case 5 C Internet
Case 6 D Telephone
Case 7 D ERP-system
Case 8 E Machine alert function
Case 9 F Telephone
Case 10 F Telephone
Case 11 G Newscast
Case 12 H -
Case 13 I -
Case 14 I ERP-system
Case 15 J Email
Case 16 K Email
Case 17 K Email
Case 18 L -
Case 19 L Email
4.4 Data Collection
The data collection took place between November and December 2017 by conducting 12 semi- structured interviews in 12 different companies. The companies were approached via telephone, e-mail or the social network LinkedIn. The participants were selected based on criteria such as work experience, job types and the willingness to participate in this research. To ensure their experience with various types of near-miss incidents, participants needed at least two years of working experience to take part in this study. Since this study focuses on supply chain near- misses, participants needed to work in the fields of purchasing, supply chain management or
logistics. Based on the prior literature review, an interview protocol was developed providing the structure for the data collection process. After the confirmation of the interviewees to participate in this research, the interview protocol and the questions for the interview were sent to the respondents via e-mail five days before the interviews were scheduled to take place.
Before the start of the interviews, the research topic was clarified to the participants, along with a short introduction. Permission to record the interviews was requested, as well as details and guarantees about the confidentiality and privacy for the interviews. Two interviewers raised alternatingly open questions in face-to-face interviews and one phone interview. The respondents had to recall two near-miss examples to describe the discovery process, the communication process and their assessment of the situation from a retrospective. The description was followed by general questions about the structures, processes, and technology of the company and general information about the supply chain characteristics. Before carrying out the interviews, a pilot interview was conducted to assess the validity of the interview questions. Construct validity was additionally assured by reviewing the study and its construction from one company. The use of an interview protocol, recording device, and multiple interviewers aimed to reduce any observer bias. Table 5, based on Voss et al. (2002) and Riege (2003) represents all the steps undertaken to guarantee the validity and reliability of this research.
TABLE 5
Validity and Reliability of this Research Type of validity
or reliability Procedures
Construct validity
• Clear procedure of data collection
• Use of academic papers (Source of evidence)
• Pilot study (Source of evidence)
• Questionnaires to measure risk management culture (Source of evidence)
• Observations of processes of one company and conversations with their employees (Source of evidence)
• Semi-structured interviews (Source of evidence)
• Review of study from company through presentation Internal
validity
• Cross-case analysis
• Clear explanation of results External
validity
• Replication logic: multiple cases (Generalizability)
• Theoretical ground of research Reliability • Use of case study protocol
• Interviewing and data analysis carried out by multiple researchers
• Audio record of data and transcriptions of interviews
Data was collected until theoretical saturation was reached regarding the effect of visibility and culture on near-miss identification. A total of 19 near-miss incidents complies with qualitative research propositions stating that between four and ten cases can provide theoretical saturation (Eisenhardt, 1989). Afterward, interviews were transcribed and sent back to the interviewees to receive their final approval of the content used for this study. All interviews were recorded and transcribed according to the 24-hour rule of Eisenhardt (1989). One company (C15) was studied more in depth, by talking to employees and observation of the company’s processes. This information was afterward taken into account regarding the data analysis.
4.5 Data Analysis
After transcribing the interviews, the researcher analyzed all of them by using Microsoft Excel.
Of the original 24 near-miss incidents, five had to be excluded. Since these events had a major effect on the customer side, they stood in conflict with the near-miss definition of this paper.
Before starting with the analysis, it was verified whether the risk management culture affected
the systems used for the near-miss identification. Table 6 depicts the level of risk management culture, derived from the questionnaire and the near-miss identification system across the 19 cases. Though companies with highly developed risk management culture did not necessarily use different systems, they tend to have a wider variety of systems to identify near-misses.
Despite this observation, this study found more cases in which the risk management culture of the company was highly developed. Therefore, this conclusion can neither be supported nor rejected.
TABLE 6
Risk Management Culture and Support Systems of Case 1-19
Risk management culture Support system
High
(C3, C4, C5, C8, C15, C16, C17, C18, C19)
Telephone, ERP-system, email, machine alert function, newscast, internet
Medium
(C1, C2, C9, C10, C12) Telephone, ERP-system, email
Low
(C6, C7, C11, C13, C14) Telephone, ERP-system, email
The interview analysis was based on the three steps proposed by Strauss & Corbin (1990): open, axial and selective coding. Firstly, the data were reduced to quotes, sentences, and/or paragraphs that were relevant for answering the research question (first-order-codes). Secondly, the first-order codes were either linked to the operationalized constructs of risk management culture or visibility. Table 7 shows an overview of the investigated variables, their operationalization, and definitions.
TABLE 7
Operationalization of Constructs Variable Operationalization Definition
Near-miss identification
Conventional mode
Recognition of a hazardous event by receiving information from a supplier/customer. (adapted from Macdonald & Corsi (2013), Phimister et al.(2003))
Adventitious mode
Recognition of a hazardous event by chance.
(adapted from Macdonald & Corsi (2013), Phimister et al. (2003))
Proactive mode
Recognition of a hazardous event by a pre- established process. (adapted from Macdonald
& Corsi (2013), Phimister et al. (2003))
Visibility
Information sharing
"The extent to which a firm shares a variety of relevant, accurate, complete and confidential ideas, plans, and procedures with its supply chain partners in a timely manner" (Scholten &
Schilder, 2015: 251)
Warning capabilities
Warning capabilities refers to an organization’s technological resources, processes, and
competencies to gather, share and analyze risk- related information (adapted from Brandon et al. (2014), Fawcett et al. (2007), Blackhurst et al. (2007))
Risk management
culture
Risk awareness
Inclusion of supply chain risk assessment as a formal part of the decision-making process at every level (adapted from Christopher & Peck (2004))
Leadership representation
The representation of a risk management focus in the board of directors (adapted from
Christopher & Peck (2004)) Risk management teams
Existence of cross-functional teams that audit risk using frameworks and tools (adapted from Christopher& Peck (2004))
Thirdly, first-order codes were developed into descriptive second-order codes such as
“displaying trends”, “displaying deviations” or “frequency of monitoring”, as it is shown in Table 8. These codes gave already two indications: 1) event characteristics that played a role in the identification of near-misses and 2) discovery strategies that can be derived from them.
TABLE 8
Example of Second-order-codes
Data reduction First-order-code Second-order-code Link to visibility
“In the systems, they can see stock levels over time and they see that they were really going into the negative when we did nothing”- (C2).
Displaying trends Warning capabilities
“The correction had been done one and a half weeks earlier. Normally, you only check the materials you have to order from suppliers once a week. So, suddenly the problem occurred. “(C7)
Frequency of monitoring Warning capabilities
In the case of a direct relationship, second-order codes that were previously connected to risk management culture and visibility were linked to the different modes of near-miss identification. This revealed that codes related to risk awareness and leadership such as
“requesting risk management actions” or “initiative to communicate externally” had only an indirect influence on the identification of near-misses. Furthermore, no relationship has been found between risk management teams and near-miss identification. Both, first- and second- order-codes were developed by focusing on each case individually. On one hand, the focus on individual cases helps to cope with the amount of data, and on the other hand it aids the understanding of the dynamics within different settings (Eisenhardt, 1989). In a subsequent step, the near-miss incidents were analyzed across cases. First, cases were compared within one company and afterward across all 19 cases. Two different dimensions of third-order-codes were established. One pattern found was discovery strategies based on visibility. Five discovery strategies were derived such as “monitoring machinery” and “monitoring business resources.”
Table 9 shows an example of how third-order-codes were developed.
TABLE 9
Example of Third-order-codes (Discovery Strategies)
Data reduction First-order-code Second-order- code
Third-order- code
Link to near- miss identification
Link to visibility
“What happened was that we could not produce it [machine failure]. We only noticed it, at the moment it happened. So, we were trying to produce it [product], but the production line was not able to get the right specifications of the product”- (C13)
Timely information
Monitoring machinery
Adventitious
discovery Information sharing
“That is mostly how it is done. We have a stop on the line, when we have an incident or a disturbance in the production line” - (C18)
Timely information
Monitoring machinery
Adventitious discovery
Information sharing
The comparison across cases enabled the identification of another pattern, that two event characteristics, “speed of risk” and “probability of an event occurring” influence the recognition of near-miss incidents. This is presented in Table 10.
TABLE 10
Example of Third-order-codes (Event Characteristics)
Data reduction First-order-code Second- order-code
Third- order-code
Link to near- miss identification
Link to visibility
Link to risk manage-
ment culture
“Sometimes, you also recognize it earlier, but, yes, …when we are missing a lot of supply, we can foresee that earlier”- (C2)
Displaying deviations
Speed of risk
Proactive mode
Warning capabilities
“For example, if I see now, that I am not able to supply a certain customer, I will not inform them today. But I will first evaluate:” Ok are there any actions that I can take? Can I change my production?
Are there any other customers that can postpone their orders?” Because, that always happens of course. And, as I really do not see any other solution, I will inform them [the customer].”- (C13)
Initiative to communicate risk externally
Probability Conventional Mode
Risk awareness
5. RESULTS 5.1 Overview of Findings
The central discovery of this study is that two event characteristics lead to the recognition of a hazardous event: the speed of risk and the probability of an event occurring. To make both factors apparent, this study found five discovery strategies that companies have derived based on visibility. Moreover, four factors that influence the near-miss identification indirectly result from leadership representation and risk awareness. Table 11 depicts an overview of these findings.
TABLE 11 Overview of Findings
Discovery
concepts Operationalization Discovery strategies and influencing factors How? Near-miss identification
Visibility
Information sharing
Risk communication (C4, C6, C9, C10, C19, C15, C16, C17)
Direct visits (C12)
Making event characteristics more apparent:
- Probability - Speed of risk
Direct influence:
- Adventitious - Conventional - Proactive Warning
capabilities
Monitoring news (C5, C11)
Monitoring machinery (C8, C13, C18)
Monitoring business resources (C1, C2, C3, C7, C14)
Risk management
culture
Leadership representation
Providing resources (C17, C16, C14)
Requesting risk management actions
(C13, C14, C15)
Increasing visibility by:
- Increasing frequency of tool use - Established risk management processes &
providing tool
Indirect influence:
- Information sharing - Warning capabilities Risk awareness
Initiative to communicate risk externally (C7, C1,
C3, C8, C11, C15)
Initiative to communicate risk internally (C1-19)
Risk management
teams - - -
5.2 Findings on Visibility
This study found that event characteristics such as the speed of risk and the probability of an event occurring determine whether or not a hazardous event is recognized. To assess these characteristics, companies have derived different near-miss discovery strategies based on visibility. Table 12 presents an overview of the discovery strategies, support systems, and underlying mechanisms to make the probability and the speed of risk apparent. In the following, this study will explain, how the aforementioned discovery strategies reveal near-misses.
TABLE 12
Overview of Discovery Strategies Based on Visibility, Support Systems and Mechanisms in the Near-Miss Identification
Discovery strategies
Supporting systems
Mechanisms to make speed of risk and/or probability apparent
Event
characteristics that trigger near-miss identification
Near-miss discovery method Risk
communication (C4, C6, C9, C10, C19, C15,
C16, C17)
Email, Telephone
• Timely information (C9, C10, C19, C4, C6)
• Precise information (C9, C4)
Probability Conventional mode
Direct visits
(C12) No system • Progress control (C12)
Probability Proactive mode Monitoring
news (C5, C11)
Newscast, Internet
• Recent information
(C11) Probability
Proactive mode
• Displaying deviations (C5)
Probability and speed of risk
Monitoring of business resources (C1, C2, C3,
C7, C14)
ERP- system
• Displaying deviations (C7, C3, C14)
Probability and speed of risk
Proactive mode
• Recent information
(C7) Probability
• Displaying trends
(C1, C2) Speed of risk
• Complete information (C14)
Probability and speed of risk Monitoring of
Machinery (C8, C13, C18)
Alert function, no system
• Displaying deviations (C8)
Probability and
speed of risk Proactive
• Machine failure (C13,
C18) Probability Adventitious
In nine cases, risk communication was used to recognize a near-miss. This means a supplier or customer was sharing risk-related information about a delay (C6, C9, C10, C15, C16, C17, and C19) or financial problems (C4) with the company. This discovery strategy was supported by systems such as telephone and email, to increase the speed of information sharing (see Table 12). Although in all cases, the near-miss was identified before it caused any damage, identification took place after the hazardous event had already occurred. Recognition therefore only occurred when an event had a 100% probability of taking place. Therefore, in five cases, interviewees specified that timely information is an important mechanism in order to discover a near-miss: “The sooner we get the information about potential issues, the better it is “- (C9).
Timely information enables a recognition before an event reaches a 100 % probability. The case
9 interviewee additionally mentioned precise information as an important mechanism in order to determine probability “He told us exactly this material and that material comes late. This information was then used by the planning department, they know which product they will run for the production and which packing is needed for this”- (C9). Risk communication leads to a conventional identification of a near-miss incident.
This study found one case, in which the near-miss was identified by a direct visit to the supplier’s site (C12). The implementation of this discovery method did not require any support systems or tools (see Table 12). As the following quotation reveals, such visits allow an order’s progress to be controlled: “I think one of our guys went to the supplier, thinking the product would be 50% ready, which it was not” - (C12). Directs visits enable the awareness of the probability of an event occurring, before it reaches 100% certainty. Since, the direct visit can be seen as a preestablished measure to identify risk, this discovery method is proactive.
This study found five cases, in which monitoring business resources revealed a near-miss (C1, C2, C3, C7, and C14). In all four cases of monitoring business resources, an ERP-system supported the discovery method (see Table 12). Case 1 and 2 identified the near-miss through a negative trend of the stock levels/ quality of products: “In the systems they can see stock levels over time and they see that they were really going negative when we did nothing”- (C2).
Identification took place before the event reached a 100 % probability of occurrence. However, in cases 3, 7 and 14 the near-miss was identified via deviations in the performance such as stock levels that were too high or too low. “They found out really late and then: Oh [no]we don’t have the stock of the 50.000 products, but we only have 20.000 in our warehouse.”- (C7).
Consequently, recognition took place after the event already occurred. However, in all cases (C1, C2, C3, C7, C14), the near-miss was identified before it caused any damage. Interviewees from case 1,3 and 7 highlighted recent information about both, the probability and speed of risk
