How can multinational
enterprises exploit
knowledge in
subsidiaries?
MASTER THESIS WITHIN: Business Administration NUMBER OF CREDITS: 30 ECTS
PROGRAMME OF STUDY: Global Management AUTHOR: Patrick Christof Raszka and Benjamin Rinkenauer JÖNKÖPING May 2020
Master Thesis in Business Administration
Title: How can multinational enterprises exploit knowledge in subsidiaries? – A middle manager perspective
Authors: Patrick Christof Raszka and Benjamin Rinkenauer Tutor: Michał Zawadzki
Date: 2020-05-18
Key terms: Knowledge Exploitation, Knowledge Management, Middle Manager, Multinational Enterprises, Subsidiary
Abstract
In this master thesis, the knowledge exploitation factors in multinational enterprises are studied. Hereby, the researchers analyse the perspective of subsidiary middle managers to understand their influence on knowledge exploitation and to understand the factors that influence middle managers in subsidiaries to share their knowledge. The aim is to provide practical implications for multinational enterprises on how to enhance knowledge exploitation from the subsidiaries and to contribute to the existing theory on knowledge exploitation factors in the light of middle managers. Thereby, the chosen qualitative research approach is abductive grounded theory, including interviews with middle managers from mostly European subsidiaries of multinational enterprises. The findings present which factors impact middle managers to share knowledge with the headquarters. Additionally, influencing factors of middle managers on knowledge exploitation are described. Hereby, the master thesis shows the interdependency between multiple factors of knowledge exploitation, resulting in a model to provide an overview. Furthermore, implications to the theory and practical implications are given on how multinational enterprises can enhance knowledge exploitation within the organisation.
Acknowledgement
We would like to take the opportunity to express our gratitude to several people that supported us in the process of writing our master thesis:
Firstly, we would like to thank our thesis supervisor Michał Zawadzki who advised us during the whole process of writing our master thesis.
Secondly, the seminars and feedback sessions with our fellow students gave us great opportunities to reflect and improve on our master thesis.
Thirdly, we would like to express our appreciation to the interviewed middle managers who openly shared their thoughts with us.
Fourthly, we want to thank our family and friends who supports us through the time of our studies and who proofread our master thesis.
Furthermore, we would like to thank the Jönköping University and their staff for providing guidance in all formalities constituting our master thesis.
Lastly, we would like to thank the city of Jönköping and Sweden for an unforgettable student life, times at the Lake Vättern, Swedish Fika breaks, great hiking tours in nature and to get in touch with the Swedish society and their culture.
Table of Contents
1.
Introduction ... 1
1.1 Background ... 1
1.2 Purpose ... 2
1.3 Research problem and question ... 3
1.4 Structure ... 3
2.
Literature Review ... 5
2.1 Knowledge management ... 5
2.2 Knowledge types ... 7
2.3 Knowledge transfer ... 8
2.3.1 Knowledge creation and conversion ... 9
2.3.2 Knowledge flow direction ... 10
2.4 The roles of middle management in knowledge transfers ... 11
2.5 Knowledge exploitation factors ... 13
3.
Methodology ... 16
3.1 Research philosophy ... 16 3.2 Research design ... 17 3.3 Theoretical sampling ... 18 3.4 Data collection ... 19 3.5 Data analysis ... 20 3.6 Research ethics ... 22 3.7 Research quality ... 234.
Empirical Findings ... 25
4.1 Process and structure ... 25
4.2 Time ... 26
4.3 Language and culture ... 27
4.4 Responsibilities and goals ... 29
4.5 Professionalism ... 30
4.6 Power ... 31
4.7 Value of information ... 33
4.8 Relationship ... 34
4.9 Trust ... 35
4.10 Operational knowledge distribution and transparency ... 36
4.11 Communication ... 37
4.12 Awareness and understanding ... 39
4.13 Human behaviour ... 40
5.
Analysis ... 42
5.1 Organisation ... 44
5.1.1 Influence from organisation on network ... 45
5.1.2 Influence from organisation on interpersonal relationship ... 46
5.1.3 Influence from organisation on individual MM ... 48
5.2 Network ... 49
5.2.1 Influence from network on interpersonal relationship ... 52
5.2.2 Influence from network on individual MM ... 53
5.3 Individual MM ... 53
5.3.1 Influence from individual MM on interpersonal relationship ... 55
5.3.2 Influence from individual MM on network ... 56
5.4.1 Influence from interpersonal relationship on network ... 59
6.
Conclusion ... 61
7.
Discussion ... 63
7.1 Implications ... 63 7.2 Limitations ... 65 7.3 Future research ... 658.
Reference List ... 68
9.
Appendices ... 76
Appendix A: GDPR thesis study consent form ... 76
Figures
Figure 1: SECI model ... 9
Figure 2: Coding example ... 21
Figure 3: Layers of influencing factors ... 42
Figure 4: Organisational influences on other layers ... 44
Figure 5: Network influences on other layers ... 49
Figure 6: Individual MM influences on other layers ... 54
Figure 7: Interpersonal Relationship influences on other layers ... 57
Figure 8: Impacts of the different layers on each other ... 61
Tables
Table 1: Comparison of knowledge process perspectives ... 6List of Abbreviations
HQ ... Headquarters ICT ... Information and Communication Technology KM ... Knowledge Management MM ... Middle Manager MNE ... Multinational Enterprise
1. Introduction
_____________________________________________________________________________________ The introduction represents our motivation and the importance to discuss our topic. The background provides a short overview of our research area, while the purpose and the research questions describe a specific research field and aim. Furthermore, the structure of the master thesis is presented.
______________________________________________________________________
1.1 Background
The perspective and understanding of knowledge management (KM) in organisations has been considerably changed over the last 60 years. The resource-based view initially promoted by Edith Penrose in 1959 defines knowledge as one of the resources which are essential for achieving a competitive advantage and therefore the success of an organisation (Penrose, 1995). While the resource-based view was further developed over the years, Peter Drucker was one of the first researchers who admitted that knowledge is the basic economic resource and others are less important to create a competitive advantage (Drucker, 1993). Additional research developed and discussed the now called knowledge-based theory (Kogut & Zander, 1992; Nonaka & Takeuchi, 1995) which perceives knowledge as the most import and strategic resource (Anh et al., 2006). However, besides the creation of competitive advantage, sharing knowledge enhances the organisational performance (Oh, 2019) underlining that internally managed knowledge flows increase the competitiveness of multinational enterprise (MNE) (Gupta & Govindarajan, 1991; Tseng, 2015). Hence, managing knowledge becomes a complex topic in a multinational setting (Massingham, 2010; Mom et al., 2007; Nishinaka et al., 2015) which is remarkable by the extensive interest in knowledge management, especially in the field of MNEs (Gupta & Govindarajan, 1991; Monteiro et al., 2008; Tippmann et al., 2014; Tseng, 2015). We acknowledge the value of the knowledge-based theory and its influence on the performance of an organisation and focus therefore on knowledge in MNEs.
One of the first ideas to manage knowledge in MNEs was by leading subsidiaries with an ethnocentric approach (Perlmutter & Heenan, 1977) where knowledge is generated, transferred and led by the headquarters (HQ) top-down to the subsidiaries. However, by using this top-down method of managing the MNE, the available knowledge from individuals in the subsidiaries might stay local and therefore unreleased for further areas of the organisation. Previous research addressed this problem and defined that knowledge is created and developed in all firm’s locations (Almeida et al., 2002). Therefore, MNEs should
elaborate on how knowledge of individuals in all firm’s locations can be used and transferred to increase the performance of the whole organisation.
We have recognised that actual research acknowledges the importance of bottom-up driven knowledge flows (Burgelman, 1983; Mintzberg & Waters, 1985; Mom et al., 2007; Tippmann et al., 2014). However, in comparison with the top-down approach, there is significantly less research on this process and its management. By studying knowledge flows we want to focus on the middle manager (MM), as we identify their importance for transferring and sustaining an effective knowledge flow from the individuals in the subsidiary to the HQ. Additionally, we are confident that the hierarchical level of MMs provides rich insights to understand how MNEs can benefit from knowledge of their subsidiaries. By doing so, this master thesis elaborates how knowledge can be understood in an MNE, how MMs influence knowledge exploitation and how organisations can encourage knowledge flows from the subsidiary to the HQ.
1.2 Purpose
The purpose of the study is driven by our professional experiences in subsidiaries of MNEs where the potential of using existing internal knowledge is greater and might result in an advantage for the organisation (Mom et al., 2007). We both experienced that subsidiaries create their own knowledge independently from the HQ. However, sometimes knowledge was shared and sometimes it stayed local. These differences led to misunderstandings and interrupted working processes. Therefore, we want to research knowledge flows in MNEs from the subsidiary to the HQ, and how they might increase knowledge sharing. Thereby, we see the important position of MMs in subsidiaries for knowledge exploitation in MNEs. By understanding the importance of knowledge exploitation and its execution, organisations are getting more effectively and obtain an advantage over their competitors (Kogut & Zander, 2003). However, the literature examined several factors which might hinder the knowledge flow from MMs bottom-up (Cohen & Levinthal, 1990; Mom et al., 2007; Oh, 2019).
Hence, this master thesis aims to assess influencing factors from a subsidiary MM perspective and elaborate on their role in MNE knowledge exploitation. By doing so, we want to reveal obstacles, create awareness and provide a meaningful insight for MMs and HQ’s executives, as well as deepen our own understanding of knowledge exploitation factors as perceived by the subsidiaries’ MMs. The master thesis should support MNEs to elaborate on their
elaboration of the MM role should be the starting point to dip deeper into the knowledge exploitation activities of the whole organisation. Furthermore, this master thesis aims to contribute to the actual research about knowledge management by connecting it to the hierarchical level of MMs and deepen the understanding of the researched phenomenon. We also aim to add to the existing theory on knowledge exploitation factors in MNEs and the MMs’ perspectives on them. By studying knowledge flows from different perspectives and various flow directions, further theoretical and practical insights can be created and used by MNEs.
1.3 Research problem and question
Research supports the essential role of MMs in communication and in connecting different hierarchical levels of the organisation (Huy & Guo, 2017; Nonaka, 1988; Tippmann et al., 2014). Furthermore, knowledge is seen as the most important resource to sustain the long term survival of an organisation (Anh et al., 2006; Gupta & Govindarajan, 1991; Tseng, 2015). By linking the increasing importance of knowledge and the inter-hierarchical connecting characteristics of MMs, we assume that organisations should understand the perspective of MMs more in-depth to take advantage of bottom-up knowledge flows. However, we identify the issue that there is little research about the MM and its role in exploiting knowledge from the subsidiary to the HQ. Previous research tends to focus on other connections to knowledge such as: the top management team (Athanassiou & Nigh, 2000), individuals of the work force (Holste & Fields, 2010; Levy et al., 2010) or the structure of the organisation (Massingham, 2010; Verbeke et al., 2013). To research the topic of exploitation of knowledge in MNEs with the perspective of the MM, we address following research questions:
• What influences MMs from subsidiaries to share knowledge with the HQ?
• How are MMs influencing the exploitation of knowledge to make it valuable for the whole organisation?
1.4 Structure
After our master thesis was introduced and motivated with the background, the research purpose and the research question in chapter 1, it continuous with the following structure. In chapter 2 the relevant literature for knowledge and middle management is reviewed. After creating the theoretical background, the methodology is described in chapter 3. Hereby, the
philosophical standpoint, and theresearch method and design are presented in detail. In chapter 4 the empirical findings from the interviews are displayed. Thereafter, in chapter 5 the empirical findings are analysed in the light of the reviewed literature from the second chapter to answer our research questions. The analysis leads to the conclusion in chapter 6. After presenting our conclusion the master thesis completes with our implications, limitations, and further research in chapter 7.
2. Literature Review
______________________________________________________________________ The purpose of this chapter is to provide the theoretical background of the topic. Hereby, the literature on knowledge management and middle management is reviewed to present an overview of the actual research. The section ends with the identification of factors which influence knowledge exploitation of subsidiaries. ______________________________________________________________________
2.1 Knowledge management
Research in KM has been a significant yet complex topic for the last decades (J. D. Collins & Hitt, 2006; Freeman et al., 2010; Holste & Fields, 2010; Liebowitz, 1999). The complexity of the field is already noticeable in the definition of knowledge. The subject is defined and interpreted in multiple ways. Liebowitz and Wilcox (1997) understand knowledge as the whole set of insights, experiences and procedures that are considered correct and true and therefore guide thoughts, behaviour, and communication of people. Another perspective recognises knowledge as information that has been organised and analysed to solve problems or to make decisions (Turban & Frenzel, 1992). In line with this perspective of Turban and Frenzel (1992) is the differentiation of data, information and knowledge from Roberts (2000). While data is raw material with no meaning, the arrangement of data is creating meaningful patterns which is seen as information. By applying information to the given understanding, experience, familiarities or learning, knowledge is created (Roberts, 2000). Further literature acknowledges this definition by describing knowledge as “deeper” and “richer”(Anh et al., 2006, p. 465) in comparison to information. Therefore, for the purpose of this master thesis the definition of Roberts (2000) is furthermore used.
The KM process is defined by researchers by using different approaches. Biloslavo and Trnavčevič (2007) define it as a collection of processes which includes knowledge generation, storage, transfer, and usage with the aim to increase the organisational value for the stakeholders. Equivalent to this suggestion are the process steps of acquisition and learning, storage and maintenance, application and exploitation, dissemination and transfer, knowledge creation, and performance measurement (Burnett et al., 2004). The main discrepancy between these researches is the process step of knowledge creation which implies a “new way of doing things” (Burnett et al., 2004, p. 28). Therefore, their suggested framework functions as a closed loop where knowledge creation and performance measurement lead again to the first process step of acquisition and learning. Another research from van der Spek and Spijkervet (1997) divides knowledge into four processes, which are:
developing new knowledge, securing new and existing knowledge, distributing knowledge, and combining available knowledge. The main differentiation to the previous approaches is the process of securing new and existing knowledge. However, by studying the literature, conceptions are used which express related thoughts. While Biloslavo and Trnavčevič (2007) start the KM process with knowledge generation, other researches use the terms acquisition and learning (Burnett et al., 2004), or developing new knowledge (van der Spek & Spijkervet, 1997). The conception of transfer and usage (Biloslavo & Trnavčevič, 2007) is related to application and exploitation of knowledge (Burnett et al., 2004) and securing new and existing knowledge (van der Spek & Spijkervet, 1997). To compare the three different KM process perspectives and how they can be grouped in the concepts of adaptability and alignment we created a visualisation which is shown in Table 1.
Table 1: Comparison of knowledge process perspectives
Emerging concepts Biloslavo and Trnavčevič (2007) Burnett et al. (2004)
van der Spek & Spijkervet (1997) Adaptability (exploration) Knowledge generation Acquisition and learning
Developing new knowledge Alignment (exploitation) Transfer and usage Application and exploitation
Securing new and existing knowledge
The reviewed literature has two attributes of KM in common which are furthermore defined as adaptability (exploration) and alignment (exploitation). Adaptability is the “ability to move quickly towards new opportunities, to adjust to volatile markets and to avoid complacency” (Birkinshaw & Gibson, 2004, p. 47), while alignment is “a clear sense of how value is being created in the short term and how activities should be coordinated and streamlined to deliver that value” (Birkinshaw & Gibson, 2004, p. 47). For the purpose of this master thesis the beforehand activities for alignment (exploitation) include knowledge transfer, knowledge sharing and knowledge exchange. March (1991) defines adaptability and alignment as separate activities, while further research acknowledges that the activities are closely complementary (Almeida et al., 2002; Birkinshaw & Gibson, 2004). The interconnection of the two activities is defined as the concept of ambidexterity (Birkinshaw & Gibson, 2004; Birkinshaw & Gupta, 2013; Blindenbach-Driessen & van den Ende, 2014; Schnellbächer et al., 2019). Birkinshaw and Gibson (2004) express with the concept of contextual ambidexterity the choice of individual employees to align or adapt their activities in their
day-to-day work. Therefore, the concept of contextual ambidexterity acknowledges the ability of the individual to decide which activity is needed in any given situation.
2.2 Knowledge types
As seen in the previous chapter, knowledge can provide companies with an advantage and companies can manage activities to explore and exploit knowledge. However, there are different types of knowledge, which require specific kinds of activities and which have a different impact on a company’s success (Kogut & Zander, 2003; Li & Gao, 2003; Nonaka & Takeuchi, 1995).
In management research, the philosophical knowledge differentiation by Polanyi (1997) has been adapted, where one type of knowledge is called “explicit” and the other one “tacit” (Holste & Fields, 2010; Massingham, 2010; Montazemi et al., 2012; Scully et al., 2013). Explicit knowledge is the knowledge which can be codified, formulated and written down (H. M. Collins, 2010; Nonaka & Takeuchi, 1995). Examples of explicit knowledge are handbooks, written reports, patents, documents, codes, formulas, prototypes and products (Holste & Fields, 2010, p. 130). By taking this into consideration one example could be if a company uses a floor plan to show the location of rooms and therein located departments. On the contrary, tacit knowledge is personalised in the experiences, abilities and feelings of a person (H. M. Collins, 2010; Holste & Fields, 2010; Nonaka & Takeuchi, 1995; Polanyi, 1983). Hereby, an example of tacit knowledge could be the experience of an account manager in how to deal with stakeholders in specific situations. The individual has gained experiences, knows their context and has built relationships with his or her stakeholders.
This differentiation between tacit and explicit however is not the only knowledge differentiation used in management research. Whereas Nonaka and Takeuchi (1995) differentiate only between explicit and tacit knowledge, Kogut and Zander (2003) develop a scale of tacitness using the attributes codifiability, teachability and complexity. Furthermore, Li and Gao (2003) highlight the importance of differentiating Nonaka and Takeuchi’s tacit knowledge further into tacitness and implicitness of knowledge, with implicit knowledge being able to be shared and expressed but the person holding this knowledge is unwilling to do so due to various reasons. They conclude that implicit knowledge is included in Nonaka and Takeuchi’s term of tacit knowledge (Li & Gao, 2003).
Following Li and Gao’s (2003) approach, H. M. Collins’ (2010) term of tacit knowledge includes implicitness, too. He differentiates three categories of reasons why knowledge is not shared: relational tacit knowledge, somatic tacit knowledge, and collective tacit knowledge
and attributes these three knowledges with relational as weak, somatic as medium and collective as strong depending on how resistant they are to be made explicit. For H. M. Collins (2010) , relational tacit knowledge is weak tacit knowledge, because it is tacit out of “the contingencies of human relationships, history, tradition and logistics” (p.98) and can be made explicit when these contingencies change or disappear. Somatic tacit knowledge is tacit because of the limits and abilities of the brain and the body and is classified as medium. Collective tacit knowledge, being knowledge embodied in society, has the strongest resistance to be made explicit (H. M. Collins, 2010). For this master thesis we furthermore use the differentiation of Nonaka and Takeuchi (1995) while acknowledging that their term of tacit knowledge includes implicit knowledge (Li & Gao, 2003).
2.3 Knowledge transfer
Management research agrees on the importance of knowledge transfer (Almeida et al., 2002; Athanassiou & Nigh, 2000; Massingham, 2010; Montazemi et al., 2012) and according to Nonaka and Takeuchi (1995) knowledge sharing creates new knowledge. The terms of knowledge sharing, knowledge exchange and knowledge flows are sometimes interchangeably used to describe the transfer of knowledge (Mom et al., 2007). However, Mom et al. (2007) state that “whereas knowledge sharing and knowledge exchange imply a reciprocal relationship in terms of transfers of knowledge, the concept of knowledge flows allows greater precision about the directionality of the knowledge being transferred” (p.913). In general, a knowledge transfer is the dissemination of knowledge between a sender/donor and a receiver/recipient (Mom et al., 2007; Welch & Welch, 2008) and although knowledge can be transferred between units or groups, the interaction between individuals is the underlying process (Welch & Welch, 2008).
The distinction between explicit and tacit knowledge helps to understand further the process of transferring knowledge because knowledge has to be explicit or to be made explicit to be able to transfer it (Massingham, 2010; Welch & Welch, 2008). Simultaneously, the transfer of explicit knowledge needs tacit knowledge in order to understand the explicit knowledge (Welch & Welch, 2008). Therefore, both types are important for knowledge transfers. Furthermore, Kogut and Zander (2003) describe that companies “invest in ways to reduce the tacitness of technology by encoding its use and replications in rules and documentation” (p.524) in order to transfer knowledge within the own organisation faster than competitors are able to imitate.
The knowledge conversion between tacit and explicit knowledge is discussed in the next chapter and the different directions of the knowledge flow in the following.
2.3.1 Knowledge creation and conversion
In chapter “2.2. Knowledge types” explicit and tacit knowledge are described and although they are different, they still interact with each other (H. M. Collins, 2010; Nonaka & Takeuchi, 1995; Polanyi, 1997). In their “four modes of knowledge conversion” model, Nonaka and Takeuchi (1995) use the distinction between tacit and explicit knowledge and illustrate four conversion types for knowledge transfer: tacit to tacit, tacit to explicit, explicit to explicit, and explicit to tacit, as seen in Figure 1.
Source: adapted from Nonaka and Takeuchi (1995)
The first conversion process of tacit to tacit is called socialisation. Through socialisation, “an individual can acquire tacit knowledge directly from others without using language” (Nonaka & Takeuchi, 1995, p. 62), for instance through observation and imitation. This process is on its own “a limited form of knowledge creation” (Nonaka & Takeuchi, 1995, p. 70), especially for leveraging knowledge in companies (Nonaka & Takeuchi, 1995).
The second conversion type (tacit to explicit) is externalisation and entails the articulation of tacit knowledge. To write down the own lessons learned or to speak to someone about own ideas are examples of externalisation. (Nonaka & Takeuchi, 1995)
The conversion from explicit to explicit knowledge is the third type of Nonaka and Takeuchi’s (1995) model and is called combination. In this process, explicit knowledge such as documents or computer databases are rearranged, combined, and added to create new explicit knowledge (Nonaka & Takeuchi, 1995).
The fourth conversion type is internalisation, in which explicit knowledge becomes tacit knowledge. This occurs when individuals read documents or listen to others’ stories and gain new knowledge out of these. (Nonaka & Takeuchi, 1995)
According to Nonaka and Takeuchi (1995), their model includes an iterative process, called the knowledge creation spiral, which shows that these four knowledge conversions are interconnected and can trigger each other.
In research, this knowledge conversion/creation model is used and referred to as the “SECI model” (Socialisation, Externalisation, Combination, and Internalisation) (Allal-Chérif & Makhlouf, 2016; Kahrens & Früauff, 2018; Richtnér et al., 2014; Scully et al., 2013). The SECI model has been criticised by several researchers for being too strongly inhibited in the Japanese culture, Japanese management and Japanese manufacturing context (Glisby & Holden, 2003; Kahrens & Früauff, 2018; Li & Gao, 2003). However, Kahrens and Früauff (2018) show that although the knowledge creation theory through the SECI model might need adaptation to the cultural context in which it is used, the SECI model and its knowledge creation theory still can be and are used in several industries and cultures.
2.3.2 Knowledge flow direction
The concepts of researchers concerning knowledge flow can be classified in the flow directions of horizontal and vertical (Mom et al., 2007; Tseng, 2015). While horizontal knowledge flows are taking place at the same hierarchical level for instance the flow between two subsidiary managers from different business units, vertical knowledge flows are proceeding between different hierarchical levels (Mom et al., 2007).
Vertical knowledge flows can be differentiated in top-down (deliberate) and bottom-up (emergent) (Mintzberg & Waters, 1985; Mom et al., 2007). Top-down knowledge flows which are intentional and top management-driven are a central part of the MNE strategy (Tippmann et al., 2014). In line with this declaration is the transfer of core competencies from the HQ to the subsidiaries (Massingham, 2010). The other vertical flow direction, bottom-up, is less researched yet, although the literature acknowledges its importance (Burgelman, 1983; Mom et al., 2007). In general, bottom-up knowledge flow means that knowledge is “coming from persons and units at lower hierarchical levels than the manager” (Mom et al., 2007, p. 915). Burgelman (1983) defines bottom-up flows as ad hoc, random, unpredictable and shared interaction between the donor and the receiver of knowledge. The successful occurrence of bottom-up knowledge flows depends on a high variety of factors which are elaborated after reviewing the considerable literature of the middle management.
However, in an MNE knowledge flows occur in the horizontal direction as well as in both vertical directions simultaneously. Therefore, Gupta and Govindarajan (1991) argue that the understanding of horizontal knowledge flows plays an important role in understanding the vertical knowledge flows and their influence on exploration and exploitation activities. Mom et al. (2007) elaborate on how the chosen approach influences exploration and exploitation activities of the manager. On the one side, the concentration of a top-down approach increases a manager’s exploitation activities, while this approach is not influencing its exploration activities. On the other side, exploration activities of managers increase by greater usage of bottom-up and horizontal knowledge inflows while the researchers cannot relate exploitation with bottom-up and horizontal knowledge flows (Mom et al., 2007).
2.4 The roles of middle management in knowledge transfers
In general, middle management is defined as being between senior/top management and the frontline/first supervising management (Harding et al., 2014; Huy, 2001; Tippmann et al., 2014), regardless of whether this position is located in the HQ or in the subsidiary, for instance as a subsidiary manager (Tippmann et al., 2014). Being in-between, the MM is in a dual position: on the one hand a subordinate, reporting to the senior management and obliged to implement their strategy. On the other hand, the MM is also a supervisor of the frontline managers and employees, and therefore receives reports and is responsible for the employees’ skills and resources (Harding et al., 2014). Hence, the MM is in the centre of the organisational hierarchy. Harding et al. (2014) describe this as “middle managers are both controlled and controllers, and resisted and resisters” and that they “constitute those [organisational] hierarchies” (p.1213).
MMs have been specifically selected and socialised for their role as agents who pursue the interest of the organisation (Fenton-O’Creevy, 1998). Due to the described characteristics of MMs, their actions become a “valuable asset” and are seen as the “central to pursuing key organisational outcomes” for organisations (Callari et al., 2019, p. 19). They have the possibility and capability to implement strategies in an informal way of communication, such as chats or personal networks (Callari et al., 2019). Hereby, research defines the MM as more effective as top managers in “bridging actions” due to more local technical knowledge and a greater local social network (Guo et al., 2017, p. 690). These networks provide the MM with formal and informal channels to gather information from peers, staff and field operators (Callari et al., 2019). Furthermore, these networks enable the MM to access and distribute knowledge faster, make sense of changes within the organisation and contribute to strategic
changes (Conway & Monks, 2011). However, MMs are not only acting according to the organisational goals, but also according to their own interests (Fenton-O’Creevy, 1998). Therefore, the importance of the role of the MM also provides possible threats for organisations. For instance, Lampaki and Papadakis (2018) describe the engagement of MMs in political power-plays and the resulting ignorance of given strategies to keep up their own personal goals. They further elaborate that a MM might distort relevant information of a given strategy and present it in its interest (Lampaki & Papadakis, 2018) or resist the implementation of a strategy because it could weaken the MM’s own job-security (Fenton-O’Creevy, 1998). Regarding knowledge flows, an example of this self-interest resulting in resistance are MMs in subsidiaries who defend a “gatekeeper” position, where they hinder a free knowledge flow by ensuring that any knowledge flow between HQ and subsidiary is controlled by themselves (Welch & Welch, 2008). In general, resistance from MMs or other employees due to self-interest can be elaborated more in-depth through the agency theory (Roscigno & Hodson, 2004; Vallas, 2006) with the senior management being the principals and the MMs their agents (Guggenberger & Rohlfing-Bastian, 2016). In this context, the social identity theory helps to understand that MMs are not only their given organisational roles but like any other individual in an organisation, their self-concept is formed by their self-identity and by the groups’ identities they belong to (Reiche et al., 2015). In the case of MNEs, the managers from a subsidiary can feel that they belong to a different social group than the managers in the HQ, for example due to a different culture or language, creating the feeling of insiders and outsiders even inside the same organisation (Hogg et al., 2005; Lauring, 2008).
The formerly described dual position, simultaneously upwards- and downwards-looking (Huy & Guo, 2017), can also be found in the role of middle management concerning knowledge flow and transfer (Tippmann et al., 2014). There, the middle management’s role is to combine the vision, theoretical and strategic knowledge from the top management with the frontline management/employees’ operational, practical, and daily business knowledge (Nonaka, 1988) and therein spanning the boundary between senior management and operational employees. Hereby, the strategic role of the MM is detecting, handling, and filtering information between the mentioned different organisational layers (Callari et al., 2019) of top management and frontline management.
Furthermore, besides being able to receive and combine knowledge from senior and frontline management, the MM needs to be able to actively search for new knowledge, inside and
Tippmann et al., 2014). Additionally, the MM has to be able to absorb and exploit the received internal and external knowledge (Cohen & Levinthal, 1990; Mom et al., 2007; Rafique et al., 2018). To summarise, the MM takes a central position in intra-organisational knowledge flows and this position is influenced by the organisational role and the MM’s own interest. Furthermore, several factors which influence the knowledge exploitation will be described more in detail later on.
2.5 Knowledge exploitation factors
The literature explores a variety of factors that influence the knowledge exploitation in vertical directions, both top-down and bottom-up. Furthermore, the most frequent factors which are involving the MM in knowledge exploitation are discussed.
The factor relational capital/relationship is broadly reviewed by the research community (Almeida et al., 2002; Colli et al., 2013; Freeman et al., 2010; Holste & Fields, 2010). A personal relationship is seen as a requirement for sharing knowledge (Holste & Fields, 2010) with co-workers as well as with the managers. Freeman et al. (2010) move a step further and characterise relationships as the “best and fastest” channel to exchange knowledge (p.77). The view is shared by J. D. Collins and Hitt (2006) who describe the likelihood that relational capital is increasing knowledge exchange and exploitation. The building of a reliable relationship sets the base to elaborate on further factors of knowledge exploitation.
The next factor is trust which can be seen as a dimension of the relational capital (Inkpen & Tsang, 2005). Trust can be defined as the “expectation that arises within a community of regular, honest, and cooperative behaviour, based on commonly shared norms, on the part of other members of that community” (Fukuyama, 1996, p. 26). By sustaining effective social interactions between the knowledge exchange partners, for instance the subsidiary MM with the co-workers, trust is developed (J. D. Collins & Hitt, 2006). The development of trust is necessary as trust takes a key role in the willingness of the knowledge owner to share knowledge (Freeman et al., 2010; Holste & Fields, 2010; Inkpen & Tsang, 2005). Therefore, Holst and Fields (2010) summarise that leaders should invest in the development of relationship and trust among co-workers. However, the international setting of MNEs might involve difficulties with effective relationships and trust. The development and maintenance of trust is more complicated in an international context compared to a domestic (J. D. Collins & Hitt, 2006) due to the strong effect of various country cultures on the development of trust (Doney et al., 1998). As trust is based on commonly shared values (Fukuyama, 1996) the development of trust between the MM and the workforce or the MM and the HQ might
be more difficult. Thus, Bhagat et al. (2002) highlight culture as a barrier of knowledge exploitation which can be overcome by the development of a strong relationship. Further research agrees with this view and highlights the importance of a pre-existent relationship to establish trust and therefore knowledge exchange (Hutchings & Weir, 2006; Levy et al., 2010).
By building and sustaining trustful relationships within all parties of the organisation, social capital is built that leads to higher exchange, respectively exploitation, of knowledge (J. D. Collins & Hitt, 2006). Hereby, J. D. Collins and Hitt (2006) define social capital as “a firm’s relationship and the resources available to the firm as a result of its relationships with other companies” (p.155). This social capital is needed at all levels of the organisation to transfer, adapt and apply context specific knowledge successfully (Freeman et al., 2010; Kogut & Zander, 1992; Montazemi et al., 2012).
Besides the broadly discussed factors of relationships, trust, and the resulting social capital, individual-related factors of the MM are identified. Hereby, sharing of knowledge might be a threat for the MM. Especially, the sharing of tacit knowledge includes the risk to lose their own advantages over the peers (Holste & Fields, 2010). This might lead to resistance against the HQ or against emerging knowledge from the peers to safeguard their own position and to secure their own job (Fenton-O’Creevy, 1998). As mentioned in the elaboration of the MM, this threat could lead to a “gatekeeper” position and a weaker flow of knowledge (Welch & Welch, 2008). Another form of resistance of the MM is rooted in the emotional bonding with the employees (Huy, 2011). Hereby, the MM wants to guard its employees from a threat outside of the subsidiary.
Lastly, on a broader organisational level, communication is identified as a knowledge transfer factor. After this factor is discussed, the organisational structure and especially the concept of absorptive capacity which takes a major role in the literature is reviewed.
Within the factor communication, the ability to articulate knowledge might be a factor which hinders the exploitation of knowledge (J. D. Collins & Hitt, 2006). Due to the international setting of an MNE, the transfer of knowledge happens across various times, spaces and languages (Schlegelmilch & Chini, 2003). People, including the MMs, might not realise which aspects of their knowledge should be shared (Levy et al., 2010). Therefore, no knowledge, insufficient knowledge, or not understandable knowledge is created. The organisational distance increases the ambiguity of knowledge and can decrease the effectiveness of knowledge transfer (Schlegelmilch & Chini, 2003).
The last identified factor is Cohen and Levinthal’s (1990) concept of absorptive capacity. This concept received a high acknowledgement within the research community and often builds the base to further discuss on knowledge transfer factors (Anh et al., 2006; J. D. Collins & Hitt, 2006; Freeman et al., 2010; Massingham, 2010). Absorptive capacity is “the ability to recognize the value of new, external information, assimilate it and apply it to commercial ends” (Cohen & Levinthal, 1990, p. 128). Although the concept implies external information instead of internal, Almeida et al. (2002) state that the ability of the subsidiary to engage in absorptive capacity also increases knowledge exploitation from the subsidiary back to the HQ. Therefore, the likelihood that bottom-up knowledge exploitation takes place correlates with the ability to engage in absorptive activities.
3. Methodology
_____________________________________________________________________________________ The chapter of methodology describes our qualitative research approach. Here, we clarify our research philosophy, argue for our research design, and define our chosen sample, the collection, and analysis of data. The chapter concludes with research ethics and research quality.
______________________________________________________________________ According to Easterby-Smith et al. (2018), the selection of a methodology and research design is connected to the ontological and epistemological viewpoint the researchers have. Therefore, the viewpoint that we take will be explained before elaborating on the chosen methodology and the justification for its selection.
3.1 Research philosophy
There are different ontological viewpoints researchers can take on a scale from realism on the one end and nominalism on the other end (Easterby-Smith et al., 2018). On this scale close to nominalism, a relativist ontology applies the most to us as we acknowledge that there are several perspectives which can be taken to research a phenomenon. This relativist viewpoint influenced our master thesis by choosing one of several perspectives, in our case the MM perspective, to assess knowledge flows in MNEs. We admit that other perspectives exist, besides the MM perspective, and that our chosen perspective is only able to reflect the MMs’ truths which are also influenced by our own interpretation and understanding. Our belief that several truths exist, dependant on the context and the perspectives, excludes other ontologies such as realism and internal realism, which assume that there is only one truth to be assessed directly (realism) or indirectly (internal realism) (Easterby-Smith et al., 2018). In line with our relativist ontology is our constructionist epistemology as we acknowledge that there are several perspectives for a researched phenomenon, for instance an MM perspective. We admit that it is necessary to research a topic not only from several group perspectives but also “to collect the views and experiences of diverse individuals and observers” (Easterby-Smith et al., 2018, p. 148). Additionally, as our work in MNEs influenced the topic of this master thesis, we think that human interests are important drivers for research and that we as researchers are not independent from the investigated phenomenon (Easterby-Smith et al., 2018). This is in contrast to positivism, where the phenomenon is assessed through objective measurement instead of subjective perceptions and where the researchers should be independent and not driven by human interests
(Easterby-Smith et al., 2018). However, as this master thesis aims to get insights into the MMs’ perceptions, it follows a constructionist epistemology and relativist ontology.
3.2 Research design
Following a relativist ontology and a constructionist epistemology, we have decided to choose a qualitative research design in accordance with Easterby-Smith et al. (2018). The qualitative research design suits this master thesis as it allows to deepen the understanding regarding a person’s or group’s perspective, embedded in their contexts. There are several different qualitative research designs, for example action research, ethnography, case study, and grounded theory (Easterby-Smith et al., 2018). We have chosen grounded theory, because it allows us to look “at the same event or process in different settings or situations” (Easterby-Smith et al., 2018, p. 228) which is aligned to our constructionist epistemology. Hereby, different approaches of grounded theory can be used. Easterby-Smith et al. (2018) present the differentiation of grounded theory of Glaser and Strauss, which both take an inductive approach. In an inductive approach, theory is generated out of the observations and solely out of the observations without building up on already existing theory (Timmermans & Tavory, 2012). However, we chose abductive grounded theory as described by Charmaz (2006) and Czarniawska (2014) as we see a pre-understanding of the research field and topic as useful to answer our research questions. In the abductive approach, the collected data is compared constantly with each other and with the already existing theory to find new insights which add to the already existing theory (Timmermans & Tavory, 2012). Abductive grounded theory therefore also supports our research questions and aim as we thrive to contribute to the existing theory about the MM’s roles in knowledge transfer and its factors, to add to the actual research and to provide insights to MNEs. Although grounded theory in general is rather a way of analysing qualitative data (Turner, 1983), it also has its implications on how research is conducted (Czarniawska, 2014; Easterby-Smith et al., 2018). To answer our research questions, we conducted interviews with subsidiary MMs to gain access to their perspectives, insights and realities. For grounded theory, these interviews should be comparable but without beforehand defined categories (Easterby-Smith et al., 2018). To ensure both in our master thesis, our interviews are semi-structured insofar that some questions, for example questions regarding the MM and subsidiary context, are the same for every interviewee whereas other questions arise individually throughout the interview. Although no predefined categories were used, we acknowledge the already existing research on knowledge flows which influenced our interviews in accordance with the
abductive grounded theory approach described by Czarniawska (2014). In grounded theory, insights from the first interviews are used to adapt the next interviews to the already analysed data as data analysis and collection are partially conducted simultaneously (Czarniawska, 2014).
However, grounded theory also has some limitations and weaknesses which can influence the research. Firstly, as data collection and data analysis are partially conducted simultaneously and should influence each other, it is difficult to define in detail all research steps before conducting the research (Czarniawska, 2014). Secondly, during the research, some compromises have to be taken due to time or interviewee constraints which might alter the initially planned research process and pose limitations towards the research results (Easterby-Smith et al., 2018). Additionally, Easterby-Smith et al. (2018) describe that students are tempted to use grounded theory without acknowledging that there are different approaches in grounded theory.
3.3 Theoretical sampling
Easterby-Smith et al. (2018) describe the importance of the researcher to explain how access to data was gained, how the data was created and recorded, which organisations were chosen and its process of selection. Therefore, the process of our theoretical sampling is described more in detail.
In line with our constructionist epistemology and the standpoint of several existing perspectives of a researched phenomenon we agreed that the sample should be MMs from various countries, companies, professions, and industries. The requirements for participants were that they must be in a MM position in a subsidiary of an MNE with an interconnection to the HQ. By choosing an open approach to allow MMs from different occupations in different countries, we see the possibility to research the phenomenon of knowledge exploitation in a broader way and to understand how different people perceive and explain the topic and its influence on the organisation. As every company is influenced by the topic of knowledge exploitation, we are confident that the diversity of our interviewees will increase the quality of our empirical study and therefore enrich our findings and contribution to the actual research.
The rather strict timeline and difficulties to reach out for MMs who are willing to share their thoughts on the topic, brought us to the decision to use an ad-hoc sampling. In this sampling strategy the cases are selected based on their availability and the ease to access them (Easterby-Smith et al., 2018). Additionally, the speed of this sampling is seen as an advantage
which had a high priority for us (Easterby-Smith et al., 2018). By doing so, our personal and professional network supported us to connect to MMs from different companies and countries. After identifying the potential interviewees, they were chosen if they fulfilled the requirements of the empirical study. In total we identified eight interviewees who agreed to conduct an interview with us. The interviewees were holding positions as a MM in the industries of medical devices, logistics, retail, software, drinks and food, and machine manufacturing. While the smallest company had approximately 600 employees the biggest sample counts almost 500,000. Furthermore, our sample consists out of two women and six men. However, in the following we will use pseudonyms which are independent of the gender of the interviewee in order to ensure confidentiality. In Table 2 we introduce these pseudonyms and the connection between the HQ and the subsidiary to which the MM belongs to. In the following chapter of data collection, the interviews and its process are described more in detail.
Table 2: Sample of MMs
Pseudonyms Nationality Age group HQ Subsidiary
Anna French 30 - 40 Germany France
Brad Finnish 50 – 60 Germany Finland
Catharina Finnish 40 - 50 Finland Italy
David German 40 - 50 Finland Germany
Eileen Russian 40 - 50 France Russia
Franziska German 50 - 60 France Germany
Gerrit German 30 - 40 USA Germany
Howard Swiss 50 - 60 Germany Switzerland
3.4 Data collection
After receiving contact details of a potential MM from our network, we contacted the person and introduced ourselves and the topic briefly. The first contact was focused on creating trust between the interviewee and us. As we acknowledge the research topic as sensible, we wanted to create an environment where the MM felt confident and safe to share his or her own thoughts on the topic of knowledge exploitation. Hereby, the GDPR consent form (Appendix A) was used to convey our ethical behaviour in research, the rights of the interviewee as well as a short summary of the research topic to call the attention and interest
of the interviewee for our research field. After the first contact was made and the interviewee agreed on taking part in our research, we scheduled the interview on one of the common video conference platforms depending on which was the most convenient for the interviewee. Face-to-face interviews were not possible due to the wide spread of location of the interviewees and the travel restrictions in Europe in the spring of 2020. However, the execution of the interviews via a video conference platform was sufficient to access the thoughts of the participants.
In total we conducted eight interviews which took place between the 25th of March and the
10th of April 2020. In general, each interview was scheduled for one hour and was conducted
by both of us together with the following agenda (Appendix B). Firstly, the interviewers introduced themselves and referred to the rights of the interviewee. Before the interviewee introduced him or herself, we asked for permission to record the interview for writing a transcript afterwards. All interviewees agreed on the recording and introduced themselves and the context of the company. Secondly, after introducing each other, the research topic, aim, and purpose were presented again as a starting point to get to the main interview. The semi-structured interview started always with the same question and continued in different directions depending on the content the interviewee shared. Hereby, the first question was to elaborate on an example of a knowledge flow the interviewee has experienced. This approach is called the critical incident technique which allowed us “to go straight to the heart” of our research topic (Easterby-Smith et al., 2018, p. 283). This is in line with our grounded theory approach as the interview was built on the incident the interviewee elaborated on. After finishing each interview, it was transcribed and sent to the participant to give them the possibility to make use of their rights to change, erase, comment or add on their responds. Due to the international environment seven interviewees shared their thoughts in English while one interview was held in German. As we are both native German speakers, we were able to translate the content of the interview to English afterwards.
3.5 Data analysis
Our data analysis was conducted in accordance to abductive grounded analysis as this analysis approach is closely linked to grounded theory (Easterby-Smith et al., 2018). In abductive grounded analysis, the first transcripts are analysed simultaneously with the collection of new data (Czarniawska, 2014). However, due to time restrictions and our approach to send the transcripts to the participants for reviewing the interview, we needed to make a compromise (Easterby-Smith et al., 2018). Although we could not analyse the transcripts while conducting
new interviews because the participants had not returned the reviewed transcripts at the time, we still used topics which came up in previous interviews for questions in the next interviews to keep the idea of abductive grounded analysis as described by Czarniawska (2014). Furthermore, according to Turner (1983), two sets of files should be used: firstly, our transcribed interviews in chronological order and secondly a list of categories which arise out of our data and evolves during the analysis of the transcribed interviews. Therefore, we used the interview transcripts in chronological order and an Excel file for the themes, categories, and codes. We first both coded the interviews individually and then compared our codes for each interview to avoid personal misunderstandings of the data. After this comparison, the final codes were then transferred into our analysis Excel file, with the themes and categories on the left and the pseudonyms of the interviewees on the top as illustrated in Figure 2.
Figure 2: Coding example
During the coding new categories and themes emerged, categories and themes were renamed, and codes and categories were shifted to other themes and categories. Hereby, especially the comparison of codes between different interviews helped us to recognise if we still had the same understanding of what comprises a category or a theme. In total, we identified 13 themes which are divided in 36 categories and contained 617 codes. According to these categories and themes, the empirical findings are presented in chapter 4. However, during the further analysis of categories and themes, we realised how closely interconnected most of the themes were to answer our research questions. Therefore, we put the themes and the more important categories into a diagram and arranged them according to their interconnections to support our interpreting process. Afterwards, we used the already existing theory on our topic to analyse further these categories and their interconnections in accordance with the abductive grounded analysis approach (Timmermans & Tavory, 2012). This resulted in the four meta-themes which we use in chapter 5 to present the analysis results.
3.6 Research ethics
While conducting research it is expected from the researchers that their study will not harm any individual or organisation that participate in the study (Easterby-Smith et al., 2018). However, the research of students in business schools had been criticised by lacking ethical practices of their research (Bell & Bryman, 2007) which might lead to a harming of the participants of a study. To overcome this issue, Bell and Bryman (2007) identified eleven categories of ethical principles which are applicable for management research in business schools. Along these principles our study was designed and ensured that a constantly fulfilment of ethical behaviour took place. Furthermore, our ethical behaviour is described along the identified categories and principles of Bell and Bryman (2007).
The first six principles are defined by Easterby-Smith et al. (2018) as “protecting the interest of the research subjects of informants” (p.251). Firstly, our research design ensured that no harm happens to the participants and the dignity was respected at all times of the research process. Hereby, the GDPR consent form was used to inform the participant about their rights, for example the withdrawal of answers, the erasure of certain sentences or the withdrawal of the whole interview. Furthermore, Bell and Bryman (2007) describe that the participants’ privacy protection has to be fulfilled. Our research complied with this principle as we used pseudonyms for all interviewees with the gender randomly chosen. By using this method, we could also protect the anonymity which is part of the principles. Additionally, the described context of the companies does not lead to an identification of the organisations the participants work for. Therefore, our research follows the principle of confidentiality not to relate to individuals, groups or organisations. Furthermore, the gathered data was recorded and stored in folders which are only accessible by us. Therefore, the confidentiality of the research data is and was ensured, all time.
Moreover, the other principles are described as “intended to protect the integrity of the research community”(Easterby-Smith et al., 2018, p. 251). Firstly, Bell and Bryman (2007) described the potential for deception through the research process. Hereby, we controlled each other not to lie or to take any behaviour which is misleading (Bell & Bryman, 2007). Furthermore, the principle of affiliation had no influence on the research as any funding, sponsoring or personal conflicts of interest do not apply to this master thesis. However, the further principle of honesty and transparency was fulfilled as an open and honest communication took place to all involved parties of the research process. Lastly, we avoided
at all time that misunderstanding, misrepresenting or false reporting of the research findings happened.
3.7 Research quality
To achieve a good research quality of a qualitative study, several aspects are important to be taken into account, namely reflexivity, transparency, a systematic and coherent research, data quality, and the reader. These aspects have been identified by Easterby-Smith et al. (2018) as the most important ones for a qualitative study, although they describe that there are more possible aspects that can be taken into account and various checklists available. Therefore, the before mentioned aspects guided us in the conducting and writing of our master thesis to ensure a good quality. Additionally, the aspect of ethics has been particularly important to us, thus we included this as a guidance for a good quality, too. These aspects guided us in the following way:
a) Reflexivity
During the research process, we reflected upon our approaches in two ways. Firstly, we weighed the strengths and the weaknesses of the planned research process, not only to choose the best fitting research approach but also to be aware of the limitations of the planned process and the results it can generate. Secondly, we stopped at several points throughout the research process to critically question if our actual actions are still in line with the planned process and with our research questions and aim.
b) Transparency
We paid attention to document in detail how we conducted our research. Therefore, we elaborated fully our planned and actual research process and explained when they differed due to compromises.
c) Systematic and coherent research
We followed one research approach in consistency with our methodology without mixing different actions which do not supply our research questions and aim. To ensure this systematic and coherent research, our work-in-progress was reviewed by other students and our supervisor several times and we adjusted our processes according to their remarks.
d) Data quality
According to Charmaz (2006), the ability of data to be put into valuable categories, data suitability, and data sufficiency are the three main criteria for good data quality. Therefore, we ensured that our sampling suits our research, and that we collect enough data, both to answer our research questions and to put the interviewees’ answers into context. Additionally, in line with the grounded theory approach, emerging topics were taken to the next interview to enrich our data on it.
e) The reader
During the writing of the master thesis, we paid attention to choose and describe the topic in a way to call the reader’s interest. Furthermore, we strived to keep a good readability and to enrich the reader’s understanding instead of creating confusion. We ensured this by letting several students and our supervisor read through our master thesis during the work-in-progress and comment on these issues.
f) Ethics
How we ensured that research ethics guided our research process is explained in detail in chapter 3.6 Research Ethics.
4. Empirical Findings
_____________________________________________________________________________________ In the empirical findings section, we summarise the shared opinions and thoughts from the interviewed MMs. These findings are structured according to the themes which evolved through the first steps of our analysis. Hereby, thirteen themes were identified and are presented descriptively in the following. The analysis of these empirical findings follows in chapter 5.
______________________________________________________________________
4.1 Process and structure
In the theme of process and structure, the interviewed MMs described different backgrounds in their companies. Anna said: “No, we do not have a specific process. This is more like logical and in order to have a good interaction between France and Germany”. She has her contacts in the HQ and whenever she wants to share her knowledge, she contacts them by email or via a call. For all other interviewed MMs, there are both informal and formal processes for knowledge sharing in place. For example, as Catharina stated: “now we are growing very fast, so this kind of practices are yet quite unformalized”. However, they have created a central team where every market has one representative participating in the central team to exchange knowledge and best practices between the markets. Additionally, they have regular online meetings with all operational country directors. Catharina stated that this process is “brutally simple”. For her, it is important that the process must be adapted to the purpose and that it is useful to have the regular meetings as a formality, but that for “operative actual things” the processes must be more agile. Similar processes are in place in other interviewees’ companies and are complemented by the informal way where the MM can contact the HQ besides the formalised processes. Brad for example, described that due to the small company size and because of the CEO being the company owner, he is frequently in contact with the CEO to exchange knowledge. Gerrit agreed: “So, I do think that size matters in terms of what you can really do. You cannot have this kind of open communication [in a team] on a European-wide level”. Despite the size, also process changes due to standardisation and centralisation were mentioned during the interviews, where standardisation was described as a benefit for the whole company even if some local processes had to be changed for this reason. Furthermore, two MMs explained that they do not want to have the knowledge flow centralised via themselves. Instead they want to encourage their employees to get in direct contact with the HQ or other countries, for example through information and communication technology (ICT) tools like an internal
social media platform. Howard elaborated more in depth about the ideal future process of knowledge transfer:
“Although from all kinds of psychological studies we always know that even though it is better to act as a group, there will be always one individual inside the group who is smarter than the collective. I mean it is a small drawback. So, at the end it is necessary to find that individual, to find that individual to make that individual speak and share the opinion and share the knowledge and then to have a process inside the community who kind of agree that this is the best idea. But I believe even if you sometimes do not decide for the best idea, I think it is important to have a common agreement that everybody is aligned behind whatever has been decided. I think that is important”.
Additionally, for him, knowledge should be shared in a group immediately. He furthermore stated:
“Then in a good functioning learning organisation, you would then always have along the path something which I might call an after-action review”.
4.2 Time
Time is a topic that was mentioned by all interviewed MMs. Most of them said that they or the employees need to invest time to share knowledge. Anna for example said that the interactions with the HQ represent 10 % of her time. Most of the MMs have regular check-up calls, monthly online meetings, or weekly reports to collect and compile information. Some explained that the high workload and therefore the lack of time in the HQ leads to slower knowledge spreading and idea implementation. David described that also the employees in the subsidiaries sometimes do not share their new ideas and solutions due to a lack of time. Apart from taking time to share knowledge, Eileen for example explained that information processing takes time, too, and that sometimes she got too many emails to process. Nevertheless, she admitted that she also replied on emails although she knew the other person had no time to read it:
“This is a vicious circle. When I see someone is sending me a Covid letter from Brazil, I feel obliged to send back a Covid letter from Russia. […] Although I understand that they do not have time to read it. This is how it works”.
Most MMs said they understand that the HQ has not much time and needs to prioritise. In this regard, Catharina said:
“It needs to be balanced that you understand: okay he got the point, but he does not have the time now. So, then you will bring it on the table another time. So, you do not let it be forgotten. But at the same time, you do not take it personally that it is not taken on the table immediately”.
However, Franziska explained that some employees might stop sharing their ideas when they do not see them implemented in an appropriate time, for example less than a year. Apart from investing time for the knowledge flow, Catharina, Gerrit and Howard mentioned that also the timing of the knowledge transfer is important. Howard further explained, the question here is:
“Shall I wait a while to see whether a storm is settling down and then bring the information into the game or should I just spit it out and maybe increase or create some turmoil by saying it”.
4.3 Language and culture
All interviewees agreed that language is a factor which hinders knowledge exploitation. Anna spoke about language difficulties, Catharina said that there is no way to survive without the local language which is in her case Italian. Gerrit added on this point: “It helps very much when you can talk in your native language”. Furthermore, David explained that language can increase the difficulty and Eileen described that communication is much better with the local Russian language compared to English. In Franziska’s company they have issues to find English speaking staff for their projects, so she described the factor of language as a barrier of knowledge exploitation. Another viewpoint on language is given by David’s statement that “we need to talk about the same things”. Gerrit said that they have quite different languages in different departments. For instance, special words, acronyms, or abbreviations led to misunderstandings and a loss of knowledge as he mentioned:
“So, on the one hand you have the company language, you definitely have. We have our own language and it takes you some time to talk and understand it and it really takes time which obviously causes loss of information for people that do not understand it. Because we have so many abbreviations for stuff and so many specific words for things. And then you have at the department, for example for me
