Internationalisation in the Cloud. An explorative case study into the foreign market expansion of SaaS providers

TABLE OF CONTENTS

Acknowledgements

Abstract

Table of Contents

List of Abbreviations

List of Figures and Tables

1 Introduction
1.1 Problem Definition and Objective
1.2 Course of Investigation

2 Literature Review
2.1 Conceptualisation of Cloud Technology
2.2 Theories of Internationalisation & The Uppsala Model
2.2.1 The Economic Perspective
2.2.2 The Behavioural Perspective
2.2.3 The Relational Perspective
2.2.4 The Entrepreneurial Perspective
2.2.5 The Uppsala Model: The Process Model of 1977
2.2.6 The Uppsala Model: The Network Model of 2009
2.3 Knowledge & Networks in Internationalisation Literature
2.3.1 The Process Model and Knowledge
2.3.2 The Network Model and Knowledge
2.3.3 Experiential Knowledge
2.3.4 Other Types of Knowledge
2.3.5 Individual Characteristics
2.3.6 Learning and Knowledge Absorption
2.3.7 Relationships and Network Positioning
2.4 Effects of Cloud Technology on Internationalisation

3 Case Company Overview

4 Methodology
4.1 Research Model
4.2 Data Collection
4.2.1 Qualitative Interview Structure
4.2.2 Participants
4.3 Data Analysis
4.4 Deductive Category Assignment
4.4.1 Definition of the Category System from Theory

5 Results
5.1 Market Knowledge / Knowledge Opportunities
5.1.1 Entrepreneurial Dispositions (H1+)
5.1.2 Network Opportunities (H8+)
5.2 Market Commitment / Network Position
5.2.1 Network Opaqueness (H3-)
5.2.2 Varied Portfolio (H6+)
5.2.3 Network Composition and Dependency (H7a-, H7b+)
5.3 Commitment Decisions / Relationship Commitment Decisions
5.3.1 International Employee (H9+)
5.4 Current Activities / Learning, Creating, Trust-Building
5.4.1 Learning (H2+)
5.4.2 Virtuality Trap (H11-)
5.5 Cloud-Specific Product Attributes
5.5.1 Relationship Service (H5+)
5.5.2 Information Good (H10+)
5.6 Cloud-Specific Market / Customer-Related Factors
5.6.1 Cloud Readiness (H4-)
5.6.2 Bandwagon Effect (H12-)

6 Limitations and Implications for Future Scientific Research

7 Conclusion

8 Bibliography

Appendix
Appendix 1: Interview Guidelines
Appendix 2: Category Assignment

ACKNOWLEDGEMENTS

Several people and institutions have contributed to realise this thesis whom I owe my sincerest gratitude.

First, I would like to thank my supervisor at Copenhagen Business School, Kristjan Jespersen, for his academic and professional support that enabled me to complete this journey.

A sincere thank you goes out to respondents in the field of the case company, NN. My deepest gratitude for going out of your way and taking the time and effort to share your opinions and experiences about your company’s internationalisation. It is highly appreciated and brought interesting insights to light.

Last, but not least, I would like to thank my loved ones. First, my mother Stephanie and my sister Cornelia, whose support through the entire process I will always remember. Your love and encouragement was incredible and thank you for always believing in me. Secondly, I would like to express my gratitude to my friends and boyfriend, in particular Biko Somuah, Umberto Moretti, Jonathan Bentsen, Yasmin Rahbari, Helen-Marie Gibson, Hengyi Xu, and Carla Hubbard; and finally Baran Baydar. All of your love and support has meant the world to me and has enriched this episode of my life in various ways.

ABSTRACT

This thesis investigates the internationalisation complexities of cloud technology vendors. To enhance the understanding of foreign market expansion of cloud providers, this study sought to answer the overarching research questions: how do cloud providers internationalise and is available literature to date still relevant for cloud technology vendors? The researchers utilised a single case study of NN, an expanding Danish cloud SME, to provide in-depth insights for the business literature realm which has so far attracted scarce research. Literature has focused almost exclusively on SMEs, MNCs and born globals, but cloud providers were mostly ignored. To date, little systematic investigations have considered that the dynamics and processes of internationalisation significantly change in a cloud context. The most important contribution is a research model encompassing twelve moderator variables impacting the performance of cloud firms. Their sug- gested influence was drawn from a thorough theoretical analysis of contemporary internationalisation literature from the 1950s to the present day. Thereby both versions of the Uppsala Model of Internationalisation by Johanson and Vahlne (1977, 2009) were selected as suitable to provide structure for the research mod- el, namely the process model and the network model. Complementing the four categories of the Uppsala Model, the research model was enhanced with two ad- ditional categories so as to examine cloud vendor internationalisation. Using a sample of five employees who are directly responsible for internationalisation ac- tivities of the case company NN, this thesis offers interesting insights into the complexities of cloud internationalisation. The findings illustrate that international- isation literature to date is still useful for examining internationalisation perfor- mance, and in particular the network model. However, since cloud vendors mainly operate virtually, this brings forth many advantages and disadvantages which have previously not been considered. Theoretical contributions and implications for future research are discussed. This thesis provides support for ten out of twelve items and offers various compelling areas for future research.

LIST OF ABBREVIATIONS

Abbildung in dieser Leseprobe nicht enthalten

LIST OF FIGURES AND TABLES

Table 1: List of Internationalisation Activity of NN

Table 2: List of participants

Figure 1: Cloud Models (Mell & Grance, 2011)

Figure 2: Cloud Computing Types - CC-BY-SA 3. Oby Sam Johnston

Figure 3: The Uppsala Model of Internationalisation (Johanson & Vahlne, 1977)

Figure 4: The Uppsala Model of Internationalisation (Johanson & Vahlne, 2009)

Figure 5: Research Model

Figure 6: Mayring’s Deductive Category Assignment Method (2015)

1 INTRODUCTION

1.1 PROBLEM DEFINITION AND OBJECTIVE

The past decade witnessed the rapid evolution of cloud computing, an information and communication technology (ICT) that rose in importance to the point that scholars refer to it as “potentially one of the major advances in the history of com- puting” (Marston et al., 2009). The ever evolving cloud paradigm promises same functionalities as existing ICT, enables additional capabilities, whilst drastically lowering upfront capital investment (Mell & Grance, 2011), thus tearing down bar- riers to market for many ICT vendors (Reuwer et al., 2013). Research institute Gartner found the cloud computing industry grew from $46 billion in 2008 to $150 billion by 2014 (Marston et al., 2009), with a strong forecast to maintain such astounding growth rates. As foretold, 2014 was the first year that the majority of workloads were processed in the cloud versus traditional ICT space, with 51% as an exact figure (Babcock, 2015). As of yet, 68 percent of all manufacturers have more than two applications in a private cloud, whilst 82 percent of firms worldwide are believed to have multi-cloud strategies (ibid). As this thesis will highlight, cloud technology allows firms highly increased business agility through attributes such as rapid deployment, parallel batch processing, big data and business ana- lytics, and interactive applications which respond to end users in real-time (Marston et al., 2011). This changes the paradigm for all SMEs, MNCs, and de- veloping country firms (DCFs) in various ways.

On the other side, internationalisation, the process by which firms expand busi- ness activities abroad, has been extensively studied in business literature. Since the 1950s, researchers became increasingly interested in the processes and de- terminants of how and why firms internationalise. Many different narratives, theo- ries and models have been developed to date to explain such dynamics, which this thesis will examine. Contemporary inquiry encapsulates effects of globalisa- tion, but there remains a dearth of literature into business-related issues sur- rounding the cloud, especially considering that the computer-science academic realm produces “impressive amounts” (Marston et al., 2009). Scholars attempted to explain various implications resulting from how the ICT revolution affects inter- nationalisation and coined a term for explaining such phenomena: “internetalisa- tion” (Bell et al., 2001; Buttriss & Wilkinson, 2003; Mathews & Healy, 2008).

However, cloud providers have mostly been ignored and there remains a gap in business literature about how cloud providers expand to foreign markets. There- fore this paper endeavours to contribute to existing literature by exploring whether or not existing internationalisation theories to date apply to companies operating in a cloud context. Cloud firms often internationalise by default, meaning they have unprecedented reach into global markets with a fraction of traditional cost involved, and often without intending to do so (Yamin & Sinkovics, 2006). In addi- tion, they operate in mostly virtual environments and thus acquire and absorb in- formation relevant for market expansion in different ways than other companies. As evidenced by successful cloud providers such as Netflix, SalesForce, or Spoti- fy, cloud technology providers are able to expand in unparalleled speeds. Existing research identified that traditional barriers to entry are broken down enormously in a cloud context and providers benefit from massive economies of scale in their internationalisation activities. On the other hand, increased speed and intense global competition, as well as severe lack of resources, often render cloud firms in intricate positions. It is obvious that many circumstances change within a cloud context, which in turn could leave existing internationalisation theories less rele- vant. Thus this thesis will attempt to paint a picture of the processes and dynam- ics of a case company so as to assess and evaluate the expediency of such liter- ature and provide tangible evidence for this gap in the literature.

1.2 COURSE OF INVESTIGATION

Section 2.1 begins with a comprehensive conceptualisation of cloud technology, including cloud infrastructure, key elements and deployment modes so as to pro- vide the reader with a full grasp. Secondly, this paper will explain and discuss in- ternationalisation literature to date in section 2.2. Thereby the four main strands of inquiry will be explained and evaluated, ranging from the 1950s to today. The Uppsala Model of internationalisation (Johanson & Vahlne, 1977, 2009), the most cited model in contemporary study, was chosen to provide structure and inputs for this thesis’ purpose and shall be explained at the end of section 2.2. Since the Uppsala Model consists of two models and for the sake of simplicity for the read- er, following Hadley and Wilson's (2003) logic, this paper terms the original con- ceptualisation of 1977 the “process model”, whilst referring to the updated one of 2009 as “network model”. Next, based on the Uppsala Model’s structure, section 2.3 identifies areas in internationalisation literature that may be relevant for inter- nationalising cloud providers. For example, it has been postulated that knowledge acquisition and absorption, as a determinant of internationalisation success, is “one of the key assumptions in internationalisation models to date” (Brennan & Garvey, 2009, p. 120) as it represents the “centrepiece in theories” (Petersen et al., 2003, p. 2). In addition, firms’ networks and relationships are increasingly evi- denced to play a major role for internationalising companies. Moreover, section 2.3 serves to discuss such areas compared with existing research and assesses the Uppsala Model’s relevance. The literature review will conclude with section 2.4 which further discusses the role of knowledge and networks in internationali- sation in the light of the advancement of ICT, and more specifically the cloud. In particular, this section combines previous parts of the literature review and devis- es twelve hypotheses which constitute this thesis’ underlying research model. The hypotheses and their respective contributions to this paper’s research question will be explained.

Section 3 will briefly introduce the case company, NN, including relevant infor- mation and past and present processes surrounding expansion. Subsequently, the research model will be presented in section 4.1 based on all relevant factors identified in the literature review. The research model was developed on the grounds of literature gaps and the resulting hypotheses addressing factors influ- encing internationalisation effectiveness of cloud providers. Section 4.2 justifies and explains why a qualitative research approach of unstructured interviews was chosen and introduces the five interviewees as employees of NN who are directly involved in internationalisation matters. Next, section 4.3 clarifies the interview material was analysed via transcription and with the help of NVIVO. The method- ology will conclude with section 4.4 explaining Mayring’s (2015) deductive catego- ry application method and how a category system for evaluation of the material is derived. Consequentially, following the evaluation of the gathered data, section 5 elaborates on findings and results. Thereby ten hypotheses will be accepted and two rejected. Further, limitations and suggestions for future scientific inquiry shall be evaluated in section 6. Finally, this paper will present concluding statements in section 7.

2 LITERATURE REVIEW

2.1 CONCEPTUALISATION OF CLOUD TECHNOLOGY

This section will define and explain cloud computing, including brief discussions of cloud infrastructure, key elements and deployment modes. The purpose is to provide the reader with a conceptualisation of cloud computing.

Academics are yet to agree on a general definition of cloud technology, hence there are many definition proposals (Armbrust et al., 2010; Marston et al., 2009), some as long as 800 characters. For the purpose of this thesis, the official defini- tion of National Institute of Standards and Technology (NIST) will be used, since it encapsulates the most important aspects: “Cloud computing is a model for ena- bling ubiquitous, convenient, on-demand network access to a shared pool of con- figurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction” (Mell & Grance, 2011, p. 2). Put simply, the cloud centralises data storage, processing and bandwidth, and thereby delivers an on-demand service to customers over a network which is typically the internet. This service can be accessed without installation since computing resources, re- ferring to both hardware and software, derive from shared central remote servers that maintain data and applications (Bora & Ahmed, 2013). The cloud’s name originates from the abstraction of the internet visualised as clouds (Srinivasa et al., 2009). Historically large internet-based companies such as Amazon and Google realized their data storage capacities were immensely underused which led to renting out of remote server capacities, or “clouds” (ibid).

To grasp cloud infrastructure, one must differentiate between front end and back end (Srinivasa et al., 2009). Firstly, front end is the physical layer which includes hardware resources such as servers, storage units and network components (Mell & Grance, 2011). Secondly, back end describes the abstraction layer that encompasses the software enabling essential cloud characteristics (ibid). It is generally asserted that the cloud offers “infinite storage capacity for any software available in the market” (Srinivasa et al., 2009, p. 73). The cloud with its central- ised server administration system is able to balance client supply, adjust de- mands, monitor traffic and circumvent congestion - no technology has previously been able to deliver that. This controlling or balancing mechanism is referred to as “middleware” (Bora & Ahmed, 2013). All in all, the capabilities offered by cloud infrastructure are available over the network and are retrieved through standard processes which uphold utility “by heterogeneous thin or thick client platforms” (Mell & Grance, 2011, p. 2), for example smartphones, tablets or laptops. Cloud users typically access the cloud in a self-service fashion, independent of location and device with minimal need for interaction between the cloud provider and the end user. The necessary resources are “shared, dynamically scalable, rapidly provisioned, [and] virtualized” (Hamburg & Brien, 2014, p. 385).

As a so-called “product software”, the major difference to on-premises software is that no software needs to be installed on the customer’s hardware and that the end user is typically unaware of the location of physical resources and devices (Reuwer et al., 2013). Therefore, customers do not “own” the cloud, as they would for example with MS Office, an on-premises product, but instead pay for it through either a subscription based model or a pay-as-you-go model (ibid). Hence, there is no need for large up-front capital expenditures for the acquisition of software licences (Armbrust et al., 2010). Firms who use the cloud no longer incur capital expenses (CapEx), but now classify their cost as an operating expense (OpEx). This massive advantage is referred to as “cost associativity” (ibid, p. 52). Another aspect of the change in cost for the end user is the non-uniformity of time that the cloud enables. More specifically, the pay-as-you-go model means if a firm used 1,000 server-hours in one day and none the following day, they would still only pay for 1,000 server-hours in total. Enhanced competition within the networking industry further decreases prices for end users with this “usage-based pricing” (ibid, p. 53). It is asserted the cloud also changes stakeholder importance and dynamics. Cloud providers handle all pricing, sales, installation, licencing, consult- ing and general maintenance of the SaaS (Marston et al., 2011). Thus, there is a change in control and power dynamics. On one hand, end users now play a much more active role in ascertaining the SaaS actually delivers on their needs. Through this, they liaise with different industry groups and regulators (ibid). On the other hand, this enablement decreases stress on ICT departments of MNCs since they now spend less time on maintenance of systems and may thus deploy resources elsewhere. As previously mentioned, more and more firms adopt multi- cloud strategies, for example Capgemini or Vordel. These highlight the need for consistent organisational ICT policies. Hence, end users are ignorant about their own requirements and cloud providers fail to address those needs (Armbrust et al., 2010), regulators, such as sovereign government bodies or international enti- ties, are highly important now. They are expected to set up the playing field by passing laws and legislation (Marston et al., 2011).

SaaS started out as sales force automation and CRM, but has now entered multi- ple domains such as HRM, billing and invoicing, service desk management, sales pipeline management and many more (Conway, 2011). Conventional examples of cloud technology are various Google Apps which are used by billions of people around the world and which can be accessed mostly for free over a web browser (Srinivasa et al., 2009). However, one must differentiate between three different types of cloud models, as Figure 1 depicts: Software as a Service (SaaS), Plat- form as a Service (PaaS), and Infrastructure as a Service (IaaS). Beginning with SaaS, the solution is accessed via thin client interfaces, such as web browsers or program interfaces. Apart from limited user-specific application configuration set- tings, the end user does not control or manage this cloud infrastructure, which includes “network, servers, operating systems, storage, or even individual capa- bilities” (Mell & Grance, 2011, p. 2). Pioneering and well-known examples of SaaS are Salesforce.com (Bora & Ahmed, 2013) and Google web-based applica- tions (Srinivasa et al., 2009). The second type, PaaS, enables consumers to set up individually created or acquired applications via programming, libraries, or oth- er tools which are supported by the cloud provider (Mell & Grance, 2011). Again, the end consumer does not control any infrastructure, but has some control over applications, thus enabling higher-level programming and simplified maintenance and upgrading, e.g. Microsoft Azure Services and Google App Engine (Bora & Ahmed, 2013). PaaS is often deployed when multiple developers collaborating remotely. Finally, IaaS offers consumers the possibility to use own processing, storage, network and other fundamental resource capabilities. In some cases, this includes operating systems and applications run on arbitrary software (Mell & Grance, 2011). However, as with SaaS and PaaS, the consumer does not control any underlying infrastructure, but instead manages operating system and its cho- sen applications, as well as host networking components such as firewalls. A well-known example of IaaS is Amazon Web Services which offers storage, com- puting power and network communication (Bora & Ahmed, 2013). For the scope of this thesis SaaS, PaaS and IaaS will be referred to as cloud, since the “line be- tween low-level infrastructure and a higher-level platform is not crisp” (Armbrust et al., 2010, p. 50) and definitions still vary widely.

Abbildung in dieser Leseprobe nicht enthalten

Figure 1: Cloud Models (Mell & Grance, 2011)

In addition to the three modes, there are four deployment models of cloud tech- nology (Armbrust et al., 2010; Mell & Grance, 2011; Srinivasa et al., 2009): pri- vate cloud, community cloud, public cloud and hybrid cloud, as seen in Figure 2.

Abbildung in dieser Leseprobe nicht enthalten

Figure 2: Cloud Computing Types - CC-BY-SA 3. Oby Sam Johnston

Firstly, the private cloud is used by a single organisation with different business units and is controlled and managed by the firm, a third party or both (Mell & Grance, 2011). Furthermore private clouds can be located on or off premises. The private cloud works well for industries with privacy concerns, for example in- tellectual property-intensive industries or ecommerce-sites that accept credit cards (Babcock, 2015). The second deployment model is the community cloud. It may exist on or off premises (Mell & Grance, 2011) and is utilised by a group of organisations with shared interests, for example the U.S. Federal Government as using Terremark’s Enterprise Cloud platform with domains such as forms.gov, and cars.gov, which are all on one official portal, namely USA.gov (Marston et al., 2011). Thirdly, the public cloud exists on the premises of the cloud provider, is open to the public, and is managed and owned by an organisation, a firm or a combination of both (Mell & Grance, 2011). Public clouds are great for firms in industries without major data privacy concerns and with large amounts of data (Babcock, 2015). For example, the New York Times utilised Amazon EC2 public cloud service to archive 5.5TB of S3 storage to create PDFs of more than 11 mil- lion articles whilst paying a fraction of the cost (Srinivasa et al., 2009). Hence, public clouds allow firms to deliver highly scalable and reliable applications rapidly at a fraction of initial costs, whereby a major drawback is security. Lastly, the hybrid cloud represents a combination of private, community or public cloud infrastructures. Each one remains a unique entity, while they are connected via standardised or proprietary technology enabling data and application portability (Mell & Grance, 2011), so called middleware as previously mentioned. In a nutshell, a public cloud is available to the public mostly in a pay-as-you-go manner whilst a private cloud refers more to internal data centres of a specific firm or organisation which is not available to the public (Armbrust et al., 2010).

2.2 THEORIES OF INTERNATIONALISATION & THE UPPSALA MODEL

This section will firstly highlight main internationalisation theories to date, subsequently explain both Uppsala Models of internationalisation, and finally rationalise why it was chosen within the scope of this thesis.

Since the advent of globalisation and the ICT revolution has opened up possibili- ties for firms to seek business opportunities abroad, academic literature has wit- nessed an explosion of theories and empirical evidence as to how and why firms internationalise. Internationalisation refers to the process whereby firms adapt their “operations [...] to international environments” (Calof & Beamish, 1995, p. 116). It is viewed as a dynamic concept, meaning the process of increased in- volvement in international activities should be conceptualised as both inward and outward (Welch & Luostarinen, 1988). Therefore, a more comprehensive defini- tion by Beamish suffices for the scope of this thesis: “the process by which firms both increase their awareness of the direct and indirect influences of international transactions on their future, and establish and conduct transactions with other countries” (Beamish, 1990, p. 77). However, there is a distinction between tradi- tional internationalisation and online internationalisation (OI) of firms. As will be discussed in subsequent chapters, many aspects change in a cloud context. Therefore, OI refers to the “performance of business exchange across national boundaries, where national boundaries occur in the virtual digitalised domain, thereby creating an international online setting for independent online vendors” (Safari et al., 2013, p. 237).

Overall, many strands of research in internationalisation exist and to date there is no one universally accepted theory (Reuwer et al., 2013). Thus, it makes sense to view four different lines of enquiry that have developed over the past decade, in chronological order: (1) the economic perspective, (2) the behavioural perspective, (3) the relational perspective, and finally (4) the entrepreneurial perspective. Generally, different streams of literature draw upon various theoretical perspectives, thus should be considered as complementary (Coviello & McAuley, 1999; Forsgren, 2002; Johanson & Vahlne, 2003).

2.2.1 THE ECONOMIC PERSPECTIVE

This strand has its origins in economic trade theory. According to Mtigwe (2006), it is founded on eminent and influential works such as Adam Smith’s absolute ad- vantage theory in 1776, David Ricardo’s comparative advantage theory of 1817, and Heckscher and Ohlin’s theory on factor proportions in 1933. Those theories focused on the study of production factors of MNCs and foreign direct investment (FDI), as well as market entry strategies, as a determinant of transaction cost in- volved (Mtigwe, 2006). Economic theories revolve around how firms “choose their optimal structure for each stage of production by evaluating the cost of economic transaction” (Coviello & McAuley, 1999, p. 225), in order to minimise it. However, those theories of the 1950s focused on nations as unit of analysis, thus ignored individual differences in behaviour among MNCs. The 1970s saw an increase in economic theories on the MNC itself, such as Hymer (1960) and Vernom (1966). During this time academics became interested in studying entry mode choices and FDI, leading to Dunning’s influential OLI framework (1989) and internalization theory (Buckley & Casson, 1976; Hennart, 1982; Rugman, 1981). The former combines microeconomic trade theory and microeconomic theories of firms and provides a framework by which firms determine how to best deploy FDI based on an overview of their advantages, split into ownership (O), location (L), and inter- nalization (I) (Dunning, 1989, 2001). Another highly influential theory emerged in the 1980s as Porter’s theory of competitive advantages (Porter, 1980) which emphasized firms’ competitive advantages at an industry level of analysis. Porter’s generic strategies majorly influenced strategic management literature and pioneered the outside-in perspective of firms whereby market structure determines strategic directions and performance of firms. The inside-out view also originated from economic theories of firm internationalisation, with the first and most cited model, the resource-based view (RBV) (Barney, 1991; Penrose, 1959). This perspective asserts firms’ strategic decisions are determined by an inside view into a company’s unique resources and capabilities.

To summarise, all mentioned theories have been criticised for being too static (Newbert, 2007) and thus less useful for examining processes of firms’ interna- tionalisation. Furthermore, they apply to MNCs and are thus less applicable to SMEs and start-ups (Kogut & Singh, 1988). Finally, it follows that the study of FDI is less relevant for SMEs, because FDI is a resource-intense entry mode and ra- ther applies to MNCs. Therefore economic theories are excluded as potential models of analysis to examine SME cloud providers’ internationalisation.

2.2.2 THE BEHAVIOURAL PERSPECTIVE

This second strand of literature viewed internationalisation as an incremental and staged process (Cavusgil, 1984; Johanson & Vahlne, 1977; Mtigwe, 2006). It is based on notions of learning and commitment of resources and directly opposes static economic models. Renowned works in this field are Penrose’s theory of firm growth (1959) and Cyert and March’s behavioural theory (1963). The Uppsala Model of internationalisation (1977) also developed from this school of thought. Other influential process models include those conceptualised by Bilkey and Tesar (1977), Cavusgil (1980), Reid (1981) and Czinkota (1982), as well as a Fin- ish one by Welch and Luostarinen (1988). It mentions both import and export in- ternationalisation activities. The common notion of these theories is that firms in- ternationalise incrementally.

In general, process models have received wide-spread empirical and theoretical support. Process models greatly influenced SME research, as exemplified by studies of U.S. companies with under 500 employees (Leonidou & Katsikeas, 1996) and SMEs in other countries (Hamill & Gregory, 1997; Korhonen et al., 1995). However, most research still examines MNCs (Coviello & McAuley, 1999; Coviello & Munro, 1997; Oviatt & McDougall, 1994). Overall, SME research on process models saw mixed reactions from academics. As the emergence of ICT changed the business environment firms operate in drastically, it is questionable whether traditional process models still apply. Furthermore emerging markets and DCFs are mostly left out of these theories (Meyer & Gelbuda, 2006).

2.2.3 THE RELATIONAL PERSPECTIVE

The relational, or network, perspective on internationalisation views it as a pro- cess of initiating, developing, and sustaining relationships in order to establish a position in a foreign market network (Johanson & Mattson, 1988). The 1980s wit- nessed large amounts of network theory literature regarding the firm as a nexus of relationships and networks (Coviello & McAuley, 1999). These theories emerged as academics realised the influence of direct buyer-seller relationships and indirect buyer-supplier and supplier-seller relationships (Jansson, 2007). The economic perspective, and in particular transaction-cost theory (TCE) (Williamson, 1979), proved to be insufficient since it referred mainly to Western European imperfect markets at that time (Jansson, 2007). Network theory drew from two sources: firstly, behavioural theory which evolved from further research in Uppsala, Sweden; and secondly, inter-organisational theory in the realm of so- ciology (Johanson & Kao, 2010; Zain & Ng, 2006). Business networks are defined as two or more firms connected via relationships (Emerson, 1981). Thereby rela- tionships are “continually established, maintained, developed, broken and dis- solved in order to achieve the objectives of the firm” (Johanson & Mattson, 1988, p. 306). The relational perspective branched out into three views: links between networks and relationships, networks as structures, and networks as processes (Axelsson & Johanson, 1992). The first view focuses on the relationship itself, with regards to “what they look like, how they are established or whether they are direct or indirect” (Jansson, 2007, p. 34). The second view examines structures inherent to networks, i.e. the “number of links and the degree to which the organi- sations are linked to each other” (Lagrosen & Svensson, 2006, p. 377). The third view studies relationship processes as sub-processes and stages (ibid). Pioneer- ing research models within the relational perspective were both the network mod- el (Johanson & Mattson, 1988) and the activities-resources-actors (ARA) model by Håkansson and Johanson (1992). In the network model, past conditions are seen as essential since they determine present and future network relationships (path dependency).

Major advantages of the relational perspective are that it accounts for the net- work’s context (Forsgren, 2002) and that it analyses the firm’s current positioning and future opportunities. It is the most dynamic school of thought of internationali- sation to date and received widespread academic support (Chetty & Blankenburg Holm, 2000; Gilmore et al., 2006; Ojala, 2009). The relational perspective is seen as highly applicable for examining SME internationalisation (Chetty & Blankenburg Holm, 2000; Coviello & McAuley, 1999; Coviello & Munro, 1997) be- cause SMEs typically use network relationships to overcome size-related barriers to internationalisation. This effect is amplified when investigating DCFs (Johanson & Kao, 2010; Meyer & Gelbuda, 2006).

2.2.4 THE ENTREPRENEURIAL PERSPECTIVE

The last strand of research, international entrepreneurship theory (IET), originated in the 1990s as certain firms deviated from beaten paths of incremental interna- tionalisation. The companies in focus are SMEs, often high-tech, whose interna- tionalisation processes diverged because they were “international from inception” (Svensson & Payan, 2009) by extracting more than 25 percent of revenues from export activities within three years. Such firms are referred to as born globals (Chetty & Campbell-Hunt, 2004; Madsen & Servais, 1997). Existing research in this realm stems from two paths: the academic field of entrepreneurship and in- ternational business literature (McDougall & Oviatt, 2000). The sudden emer- gence of born globals is assumed to have resulted firstly from the ICT revolution (Bell & Loane, 2010), and secondly from globalisation through the liberalisation of trade (Chetty & Campbell-Hunt, 2004). SMEs typically face a lack of resources as a barrier to market, which leads them to seek alternative ways to the incremental approach to overcome resource constraints (Gabrielsson et al., 2008). For exam- ple, often discussed topics of IET circle around SMEs’ superior capabilities to in- ternationalise by drawing knowledge from prior experience of the founder or from network actors, rather than gradual experiential knowledge accumulation. Further, IET views characteristics of individual entrepreneurs as the principal driver for internationalisation and sustained competitive advantage (Reuwer et al., 2013). Such characteristics are socially complex (ibid), such as entrepreneurial skills and experience, as well as risk-taking behaviour and innovativeness (Alvarez, 2001). Since those are specific to the individual they are inimitable and become competitive advantages (Madsen & Servais, 1997).

Generally, IET received strong empirical support and contributed to the explana- tion of deviating behaviour of born globals. On the other hand, it is criticised for its too holistic, inconsistent nature (Mtigwe, 2006), and for neglecting innovation po- tentials of MNCs (Coviello & McAuley, 1999). In addition, it is suggested that IET place more importance on the relational perspective, since combining the two fields could provide useful insights into different actors within a network creating opportunities.

2.2.5 THE UPPSALA MODEL: THE PROCESS MODEL OF 1977

The original Uppsala Model of internationalisation, a result of the behavioural per- spective, was developed by Uppsala scholars Jan Johanson and Jan-Erik Vahlne in 1977. To date, it is cited as the first and most influential process model (Coviello & McAuley, 1999). The Uppsala Model was founded on results of an earlier study of Swedish firms (Johanson & Wiedersheim-Paul, 1975) which had grown from SMEs to MNCs during the early 1970s. The two starting points were firstly Penrose’s (1959) assumption that companies grow by combining and de- veloping resources, and secondly on the fact that firms’ decision-making process in internationalisation was limited and thus affected by bounded rationality (Cyert & March, 1963). The authors define the internationalisation process as a “gradual acquisition, integration, and use of knowledge about foreign markets and opera- tions, and on its successively increasing commitment to foreign markets” (Johanson & Vahlne, 1977, p. 23).

Abbildung in dieser Leseprobe nicht enthalten

Figure 3: The Uppsala Model of Internationalisation (Johanson & Vahlne, 1977)

The State variables market knowledge and market commitment demonstrate the firm’s current state of internationalisation (see Figure 3). The Change aspects commitment decisions and current activities show what has changed that state.

Starting with market commitment, it is defined as the amount and degree of re- sources committed to internationalisation (Johanson & Vahlne, 1977). For exam- ple, deploying physical assets in a foreign market, such as a wholly owned sub- sidiary, would count as high commitment compared to using agents which require less commitment. The process model’s line of thought is that expansion takes place sequentially because knowledge develops incrementally and takes time (Johanson & Wiedersheim-Paul, 1975). This is based on an understanding that knowledge is ascertained from gradual internationalisation experience. Therefore, both lack of market information and uncertainty about markets lead firms to com- mit resources incrementally (Hörnell et al., 1972; Johanson & Vahlne, 1977, 1990). Therefore, market knowledge, the other State aspect, is seen as the main driver and determinant of internationalisation. Johanson and Vahlne distinguish between four types of knowledge, of which the most significant type is experiential knowledge, since it is gained through increasing levels of experience in foreign markets. This reveals the model’s connected nature: experience is attained through small incremental steps into foreign markets; thus knowledge increases 15 which in turn leads to intensified commitment. The Change variable current activi- ties may hereby be understood as the time lag that occurs between activities car- ried out and their consequences. For example, knowledge absorption, the learn- ing organisation and knowledge codification represent important issues but are elaborated on in section 2.3.6. Therefore, the Change variable commitment deci- sions consists of decisions based on opportunities and threats. Those decisions are made by individuals in the HQ as they gained experiential knowledge, or may be taken by people outside the firm who are qualified to propose decisions (Johanson & Vahlne, 1977).

Uppsala scholars believe firms internationalise incrementally “largely due to this lack of market information and the uncertainty occasioned thereby” (Buckley, 2002, p. 43). Equally importantly, they believe knowledge gaps hindering fast in- ternationalisation relate to differences between countries (Johanson & Vahlne, 1977). These differences are referred to as psychic distance (PD), defined as “the sum of factors preventing the flow of information from and to the market” (1977, p. 24). It concerns perceived differences in, for instance, language, business prac- tices, education systems, culture and industrial development (Chetty & Campbell- Hunt, 2004). PD justifies that firms internationalise to psychically close countries first, as was found in empirical observations of the Swedish firms in picture. Firms that possess little or no knowledge about factors that create PD likely suffer con- sequences of “liability of foreignness” (1977). Johanson and Vahlne postulate that timing of such sequences was important, for example increasing intensity of entry mode choices: “generally the development of the firm seems to be in accordance with the incremental internationalisation view” (Johanson & Wiedersheim-Paul, 1975, p. 25). For example, choice of entry mode is reflected on such instances when firms begin the process with less intense commitment such as franchises or dealerships and eventually choose to increase the intensity by engaging in alli- ances or FDI (Welch & Luostarinen, 1988). This occurs because of less perceived risk as a result of improved market knowledge, a view widely supported by re- searchers (Brewer, 2007; Ellis, 2008; Nordstrom & Vahlne, 1992; Sharma & Blomstermo, 2003; Sousa & Bradley, 2006).

The process model is celebrated for its dynamic nature and the first to mention experience as a factor. It aided understanding of foreign investment behaviour and stood out during a time when other theories were primarily based on economic models of transaction cost considerations. It leaves room for differences in time and entry mode choice of firms as it does not propose any deterministic degree of commitment. In addition, the logical sequencing of the internationalisation process was met with tremendous extensive backing of empirical evidence.

However, major critiques have been voiced by several scholars. Its simplistic na- ture ignores factors such as firm size or domestic country characteristics, and thus “hampers developing other explanations of internationalisation” (Sullivan & Bauerschmidt, 1990, p. 28). Furthermore, Andersen called its “intuitive logic” and “theoretical parsimony “almost tautological” (1993, p. 216). Most importantly, the process model was criticised for ignoring differences of both individuals and firms (Petersen et al., 2003). For example, willingness to take risks does vary greatly in reality supported by empirical evidence from various strands of inquiry. More re- cently, the emergence of born globals such as Facebook and Google disrupted global business environments and brought the process model’s idea of incremen- tal sequences onto the hot seat (Barkema, Bell, & Pennings, 1996; Bell et al., 2001; Loane et al., 2009). These SMEs leapfrog through stages and empirical evidence does no longer support the process model for those cases (Coviello & McAuley, 1999). In addition, Andersen (1993) appraises the it for being too weak- ly underpinned theoretically and explains that it does not apply to firms with a great amount of resources available in stable market conditions. Overall, the in- cremental nature cannot be dismissed as of yet since even those rapidly interna- tionalising firms have been shown to go through such stages (Madsen & Servais, 1997).

2.2.6 THE UPPSALA MODEL: THE NETWORK MODEL OF 2009

As a consequence of the relational influence on modern internationalisation litera- ture, Johanson and Vahlne revisited the variables of the process model and came up with the network model (Johanson & Mattson, 1988; Johanson & Vahlne, 2009). The authors responded to critique by stating the business world had changed since the 1970s and that “events move more quickly and assume somewhat different forms. Nonetheless, one constant in coping with uncertainty remains: firms need to learn, and to create or strengthen relationships […] to exploit opportunities” (Johanson & Vahlne, 2009, p. 1423). Thus they adjusted their viewpoint on internationalisation so that it now concerns outcomes of activities which fortify network positions. Thereby firms choose entry modes and strategies according to existing networks, especially since such networks are borderless (ibid) now rendering traditional entry barriers less important. In other words, whereas the process model placed emphasis on PD between home and host as determinants for internationalisation, the network model believes that opportunity exploitation and identification result from cumulative and dynamic processes of learning, trust and commitment in networks (2009).

Within this context, a network is defined as “the relationship between a firm’s management team and employees with customers, suppliers competitors, gov- ernment, distributors, bankers, families, friends, or any other party that enables it to internationalise its business activities” (Zain & Ng, 2006, p. 184). Both advanc- es in marketing research and business network literature led the Uppsala theo- rists to assert that networks are “linked in various, complex and, to a considerable extent, invisible patters” (Johanson & Vahlne, 2009, p. 1411). Since they are bor- derless, internationalisation success mostly depends on having “insidership” with- in the right networks. To elaborate, the network model claims that a lack of insid- ership, i.e. losses made by not being an insider in a network, leads to less effec- tive internationalisation, referred to as the liability of outsidership (Johanson & Vahlne, 2009). This liability of outsidership does not explicitly refer to countries, but should rather be conceptualised on a firm-level or wider region (ibid).

Abbildung in dieser Leseprobe nicht enthalten

Figure 4: The Uppsala Model of Internationalisation (Johanson & Vahlne, 2009)

Referring to Figure 4, the first State variable knowledge opportunities had oppor- tunities added because those are seen as a subset of knowledge (Johanson & Vahlne, 2009) and its renaming has occurred due to the authors’ intention to em- phasise the uppermost importance of opportunities within a firm’s knowledge base. Other crucial components of knowledge opportunities include “needs, ca- pabilities, strategies, and networks of directly or indirectly related firms in their in- stitutional contexts” (ibid). The recognition of opportunities stems from a sharp increase in opportunity literature. They base their arguments on two sources: first- ly, Kirzner (1973), who claims the identification of opportunities is inherently en- trepreneurial and due to the individual’s preparedness for surprises, thus leaving it to luck, alertness and flexibility. The second influence by Shane (2000) builds on Kirzner’s contention and adjoins that prior knowledge is equally, if not more, important than personal characteristics. Literature has not explained adequately why firms discover opportunities so there is no one agreed upon view. However, there is some general understanding that Kirzner exaggerated the role of seren- dipity (Gruen, 2005; Shane, 2000). All in all, the network model asserts that firms in networks do face privileged access to information, both through business part- ners building knowledge together, and through an entrepreneurial awareness of how they could match those sources of knowledge (Johanson & Vahlne, 2009). Thus, similarly to the RBV (Barney, 1991), firms should internally be aware of how to combine knowledge sources and sustain competitive advantage in possible opportunity encounters. Thereby opportunity identification may be unilateral, bi- lateral or multilateral (Johanson & Vahlne, 2009). Unilateral opportunities occur when one firm identifies an opportunity by learning about the others’ needs, ca- pabilities and markets. Bilateral opportunities are created when two firms identify an opportunity together, whereas multilateral refers to the process where other players within a given network of two firms are also involved through trust. In sum, Johanson and Vahlne comment: "opportunity development is an interactive pro- cess characterised by gradually and sequentially increasing recognition (learning) and exploitation (commitment) of an opportunity, with trust being an important lub- ricant” (2009, p. 1420). The second State variable is network position and relates to the fact that relationships among international networks are determinants for internationalisation performance. Those relationships possess varying levels of knowledge, trust and commitment and differ in how they promote successful in- ternationalisation (Johanson & Vahlne, 2009). The strength and position of the network provides actors with resources which they in turn deploy (Håkanson & Johanson, 1992). Thus, commitment decisions are based on both players’ wider network and their willingness and awareness to combine and share (Birkinshaw et al., 2000; Håkanson & Johanson, 1992; Johanson & Vahlne, 2009). The Change variable relationships commitment decisions describes the extent to which a firm is willing to commit resources within a network. The process model viewed commitment in a more one-sided firm manner, whereas the network mod- el acknowledges that both sides must commit through relationships. Finally, the last variable is learning, creating, trust-building. The basic premise is that it is beneficial for firms to build and maintain relationships since they lead to accumu- lation of knowledge and building of trust which in turn develops into greater com- mitment and thus increases joint productivity (Johanson & Mattson, 1988; Johanson & Vahlne, 2003, 2009).

Generally, the network model has received strong academic support (Barkema et al., 1996; Bonaccorsi, 1992; Delios & Beamish, 1999; Ellis, 2000; Li, 1995; Luo & Peng, 1999; Sharma & Johansson, 1987; Zain & Ng, 2006). Relationships for in- ternationalisation success matter in terms of entry mode choice and selection of market, for rapidly internationalising firms (Coviello & Munro, 1997; Oviatt & McDougall, 1994) and particularly for SMEs (Chetty & Blankenburg Holm, 2000; Coviello & Munro, 1997). Nonetheless, scholars still call for further research within the field (Coviello & Munro, 1997; Ellis, 2000; Ojala, 2009; Zain & Ng, 2006) in particular regarding a need for the examination of other industries than the tech- nology industry. Further, it is still unclear exactly what kinds of relationships affect entry mode and choice of market. Johanson and Vahlne (2009) suggested com- bining the network model with other approaches. For example, they hypothesise that a combination of internalisation theory (Buckley & Casson, 1976; Rugman, 1981) or the eclectic paradigm (Dunning, 1989), i.e. the economic perspective, could work well in conjunction.

2.3 KNOWLEDGE & NETWORKS IN INTERNATIONALISATION LITERATURE

This section focuses on the roles and importance of knowledge and networks within the discourse of internationalisation literature. Firstly it turns to the original Uppsala Model of 1977, secondly follows discourse on knowledge in academia towards the final updated Uppsala Model of 2009, and subsequently discusses other potentially important aspects.

2.3.1 THE PROCESS MODEL AND KNOWLEDGE

The process model conceptualised knowledge within the State variable market knowledge, defined as “information about markets, and operations in those mar- kets, which is somehow stored and reasonably retrievable - in the minds of indi- viduals, in computer memories, and in written reports“ (Johanson & Vahlne, 1977, p. 26). It explains what existing knowledge the firm possesses at the pre-entry level of the internationalisation process. The authors distinguish between four types of knowledge which have different impacts: objective, experiential, market- specific and general (ibid, 1977, 2009). Those types have been assessed, empiri- cally tested and extended and shall be discussed subsequently. In line with other process models, and the RBV (Barney, 2001), knowledge was conceptualised as a resource within the company. The process model’s assumptions about knowledge were based on Swedish case studies, so it could be argued that it only applies within a Western European context. For example, practices of knowledge exchange or knowledge acquisition may be more similar within similar cultural contexts. However numerous international non-Western European studies sup- ported the process model’s conceptualisation of knowledge, among them: a U.S. study (Denis & Depelteau, 1985), a Hawaiian study (Hook & Czinkota, 1988), a study of Japanese export firms (Johansson & Nonaka, 1983), a study of Turkish exports (Karafakioglu, 1986) and a study of Australian firms (Bartlett, 1986). Since its inception, the role of knowledge as stated in the process model has been chal- lenged and is now recognised as “far more complex and intriguing than originally acclaimed by Johanson and Vahlne” (Petersen et al., 2003, p. 5). It must, howev- er, be stated that the process model did recognise the importance of some as- pects of knowledge and, as the most cited internationalisation model, the Uppsala scholars have pioneered and sparked discussion about the role of knowledge among academics.

2.3.2 THE NETWORK MODEL AND KNOWLEDGE

In the light of the aforementioned critique, scholars have set out to reconceptual- ise the role of knowledge in internationalisation. As explained above, network theorists see a firm’s internationalisation as a natural development from network relationships with foreign individuals and firms (Johanson & Mattson, 1988). The variable knowledge opportunities is therefore no longer recognised as a gradual and incremental process attached to one’s own cultural factors, but is rather viewed as a bundle of network knowledge opportunities arising during the interna- tionalisation process. More specifically, knowledge development underpins the foundation of the framework, in which its accumulation allows companies to col- lect opportunities emerging from their internationalisation activities (Johanson & Vahlne, 2006). Compared with the rather rigid view of knowledge types and ac- quisition, the network model accounts for the state of both the firm and the mar- ket, and thus makes room for multilateral influences on decision making. It ena- bles firms to better understand possible opportunities and constraints by paying attention to the roles and strengths of network actors (Axelsson & Johanson, 1992; McKiernan, 1992). These types of activities are classified under the term “relationship-specific knowledge” (Blankenburg Holm et al., 1999; Hoang & Rothaermel, 2005) that is developed through interaction between two partners and includes knowledge about each other’s heterogeneous resources and capa- bilities. It may provide a firm with competitive advantages that are inimitable (Burt, 1997). This is firstly because the information provided by foreign networks, which is leveraged by the internationalising firm, is not available for third parties. Sec- ondly, the likelihood of referrals and further business opportunities increases due to the network (ibid). Thirdly, networks open up a wider number and variety of knowledge sources which lead to an increased likelihood of vicarious learning (Inkpen, 1996; Osland & Yaprak, 1995). Various researchers support the notion that awareness and utilisation of knowledge through networks leads firms to enter international markets (Ellis, 2000; Loane et al., 2009; Ojala, 2009; Zain & Ng, 2006). Further, numerous studies support the notion that relationships have am- plifying effects on both learning and knowledge acquisition which in turn reduces the risk involved in internationalisation (Brennan & Garvey, 2009; Chetty & Patterson, 2002; Gabrielsson et al., 2008; Gilmore et al., 2006; Sharma & Blomstermo, 2003; Turnbull et al., 1996).

2.3.3 EXPERIENTIAL KNOWLEDGE

The process model proposes a distinction between objective knowledge and ex- periential knowledge, based on Penrose's (1959) assertion that objective/market knowledge is something that can be acquired by teaching, whereas experiential knowledge may only be obtained by personal experience (Johanson & Vahlne, 1977, 1990). Therefore market knowledge is presumed to be attained through experience from current business activities (ibid). The process model claims the “critical kind of knowledge” in internationalisation is experiential knowledge (ibid) since “experience itself can never be transmitted” (Penrose, 1959, p. 53). Objec- tive knowledge concerns issues such as market methods, statistical tools and in- sights that can be generalised and replicated by other firms (Badaracco, 1991; Petersen et al., 2003). Experiential knowledge, in contrast, is much more endemic (Petersen et al., 2003) as it is acquired through learning by doing and relates to matters such as national culture, distributive structures and customer characteris- tics (Inkpen, 1996; Nonaka, 1994). The process model has been criticised for its narrow view on knowledge and many types of knowledge have been proposed and studied since. Researchers mostly agree that in reality there is no clear cut between objective and experiential knowledge as the process model asserts. For instance, Andersen (1993) concedes that international activities require both general and experiential knowledge. Petersen et al. (2003) further elaborate that experiential knowledge naturally develops into objective knowledge thus opposing the distinction (see also Nonaka, 1994).

Overall, there has been a lack of empirical evidence on experiential knowledge, possibly due to the struggle with operationalisation of the process and network models’ constructs (Hadley & Wilson, 2003). The assumption that experiences obtained from current foreign market activities translate into valuable objective knowledge has been supported by the literature. For example, Brockmann and Anthony find that firms use this mechanism for “direct knowing, immediate under- standing, learning without conscious use of reasoning, or making a choice without formal analysis” (1998, p. 455). This is why firms take small, incremental steps towards foreign market expansion according to the process model (Johanson & Vahlne, 1977). Recently, researchers have criticised the Process Model for over- emphasizing the importance of experiential knowledge (Andersen, 1993; Eriksson et al., 2000; Petersen et al., 2003).

In addition, the process model may be too simplistic to assume that a firm’s re- source commitment proportionately increases with experiential knowledge. In fact, a U-shaped relationship between experiential knowledge and a firm’s propensity to invest resources, i.e. entry modes varying in degrees of control, has been pro- posed (Erramilli, 1991). Krishna Erramilli’s logic is that at first firms are over- optimistic about foreign market opportunities due to a lack of experiential knowledge, hence resource commitment is high. Once the firm has gathered knowledge through experience, it is likely to reduce its willingness to commit due to the awakening to the “hard facts” (ibid). Finally, as it makes sense of the knowledge and confidence increases, it finds itself more willing to commit high- scale investments and high-commitment operation methods (ibid). This model found widespread support (Ekeledo & Sivakumar, 2004; Petersen et al., 2003).

[...]