A Systematic Approach to Analyze Industrial Clusters: A Case Study of The Iceland Ocean Cluster

(1)

IN

DEGREE PROJECT INDUSTRIAL MANAGEMENT, SECOND CYCLE, 15 CREDITS

STOCKHOLM SWEDEN 2020,

A Systematic Approach to Analyze Industrial Clusters

A Case Study of The Iceland Ocean Cluster

NIKKI DEN HOLLANDER

THORKELL V. THORSTEINSSON

KTH ROYAL INSTITUTE OF TECHNOLOGY

(2)

TRITA ITM-EX 2020:198

(3)

(4)

Master of Science Thesis TRITA-ITM-EX 2020:198

A Systematic Approach to Analyze Industrial Clusters: A Case Study of The Iceland

Ocean Cluster

Nikki den Hollander Thorkell Viktor Thorsteinsson

Approved

2020-06-12

Examiner

Terrence Brown

Supervisor

Kristina Nyström

Commissioner Contact person

Abstract

Regional industries see a trend of agglomeration of firms in concentrated economic areas.

Firms in these regions convene and form a cluster with a wide array of intensive collaborations between the individual actors. The result is the enhancement of a variety of aspects in the region such as innovation, and entrepreneurial activities and attitudes. Research on cluster theory has sho n an effecti e a of anal ing a cluster s strategy and current state by identifying the phase of a cluster in its Cluster Lifecycle. In order to locate a cluster in the Cluster Lifecycle, the characteristics of the cluster need to be mapped and analyzed. Here, a gap in the literature can be found as there is no consensus on a systematic method for the analysis of cluster characteristics. Hence, this paper derives a novel way of analyzing cluster characteristics using the Cluster Canvas. Building on these existing and newly derived methods, the paper answers the research question: What can we learn from this specific cluster by analyzing unique factors in its environment? To create the most complete representation of the phenomenon in its context, the research adopts a single case study method. In the specific case of this thesis a cluster, the Iceland Ocean Cluster (IOC), is used as the case study. Primary data is collected on multiple levels using a survey and semi-structured interviews which are subsequently analyzed qualitatively. All in all, leading to a Cluster Canvas populated with the most significant findings, creating an image of the cluster in its current state of sustainment. Significant conclusions about the state of the cluster relate to its diversity in terms of collaborations and portfolio of firms and how this relates to knowledge diffusion within and outside of the cluster. Finally, evidence shows that the ability of the cluster to adapt and renew strategies over the years has proven fruitful in sustaining its relevance in the industry.

(5)

Examensarbete TRITA-ITM-EX 2020:198

A Systematic Approach to Analyze Industrial Clusters: A Case Study of The Iceland

Ocean Cluster

Nikki den Hollander Thorkell Viktor Thorsteinsson

Godkänt

2020-06-12

Examinator

Terrence Brown

Handledare

Kristina Nyström

Uppdragsgivare Kontaktperson

Sammanfattning

Regionala industrier uppvisar en trend när det gäller koncentration av företag i ekonomiska områden. Företag inom dessa regioner centreras och bildar ett kluster med breda samarbeten företagen emellan. Resultatet blir en förbättring när det gäller flertalet aspekter inom regionen, såsom innovation och entreprenöriella aktiviteter och attityder. Forskning på klusterteori har visat på ett effektivt sätt att analysera ett klusters strategi och nuvarande tillstånd genom att identifiera dess nuvarande fas i klustrets livscykel. För att kunna lokalisera ett kluster i klustrets livscykel måste klustrets karaktäristika mappas och analyseras. Här finns ett gap i litteraturen då det inte råder någon konsensus över hur klusters karaktäristika kan analyseras på ett systematiskt sätt. Denna uppsats utvecklar ett nytt sätt analysera klusters karaktärsdrag genom att använda Cluster Canvas. Baserat på existerande och nyligen utvecklade metoder, svarar uppsatsen på forskningsfrågan: Vad kan vi lära oss från ett specifikt kluster genom att analysera unika faktorer i dess omgivning? För att skapa den mest kompletta representationen av fenomenet i dess kontext, analyseras ett specifikt case. Iceland Ocean Cluster (IOC) är det kluster som studeras i uppsatsen, Primärdata samlas in på flertalet nivåer med hjälp av enkätundersökning samt semi-strukturerade intervjuer, vilka senare analyseras kvalitativt. Vår analys resulterar i ett Cluster Canvas som innehåller de mest signifikanta resultaten och som skapar en bild av klustret i fokus för detta kluster i dess nuvarande form. Våra viktigaste slutsatser om klustrets status relaterar till dess mångfald i termer av samarbete och portföljen av företag, samt hur detta relaterar till kunskapsspridning inom och utanför klustret. Slutligen, visar resultaten att klustrets förmåga att anpassa och förnya strategier över tiden har varit betydelsefullt för att säkerställa dess långsiktiga relevans i branschen.

(6)

Table of Figures

Figure 1 The Cluster Lifecycle ... 6

Figure 2 The Cluster Canvas... 8

Figure 3 Populated Cluster Canvas ... 15

Figure 4 Start-ups created in the Ocean Cluster House ... 21

Figure 5 Composition of the Iceland Ocean Cluster ... 21

Figure 6 IOC's Position in The Cluster Lifecycle ... 23

(8)

1. Introduction

1.1. Background

Regional industries repeatedly see the convergence and agglomeration of firms in concentrated economic areas (Delgado et al., 2014). The seemingly magnetic force which pulls firms together in an assembly of interaction and cooperation with firms in related industries has settled on the term clustering. Therefore, clusters are defined here as the composition of organizations, institutions, and a central actor, belonging to related industries in concentrated economic areas. Although clustering occurs in concentrated areas, clusters are far from isolated economies. They are critical nodes in a global net ork of places (Martin & Mayer, 2008, p.

273). The opportunities which arise from clustering have enabled firms to enhance efforts in a variety of aspects such as innovation (Baptista & Swann, 1998), collaboration (Mitrofanova et al., 2019), and entrepreneurial activities and attitudes (Sternberg & Litzenberger, 2004, p.

787).

Jacobs & De Man (1996) suggest that one of the criteria needed for the creation of a cluster is a central actor which organizes its formation and manages the network of firms.

These central actors are not constrained to a certain model and can be of a wide variety such as for-profit or non-profit organizations or even institutions. There are numerous different opportunities for the central actors to support the clustered firms. Some for-profit central actors, such as the one focused on in the case study, offer services to the clustered firms throughout the cluster s lifec cle. Services include incubation of young companies, this includes providing firms with economies of scale, learning through e.g. seminars and lectures, and network effects by being part of the cluster s net ork (Bruneel et al., 2012).

In section 2, existing literature is reviewed and from it an analytical framework called the Cluster Canvas is introduced. Subsequently, section 3 describes the methodology of the research performed. This thesis utilizes the opportunity at hand to analyze a particular Icelandic cluster and its characteristics by gathering data from three different levels. Firstly, on the central actor s management le el. Secondl on the tenant firm le el, i.e. the firms within the shared office space of the cluster, these are both start-ups and small to medium sized companies.

Finally, on the peripheral firm level, i.e. firms outside the cluster s shared office space. Section 4 highlights the findings of the research through the Cluster Canvas and localization of the cluster in its Cluster Lifecycle. Finally, conclusions from the research are stated in section 5.

1.2. Purpose

The methodological purpose of this research is classified as analytical. In other words, the phenomenon that is investigated is analyzed. Additionally, an explanation is given about why and how the phenomenon is happening by measuring causal relations among cluster characteristics. In the specific case of this thesis a cluster, the Iceland Ocean Cluster (IOC), is measured by its characteristics and the collected data is analyzed to formulate lessons learned explaining the causal relations between the characteristics. By doing so, the thesis dives into finding a solution for a specific research question making this applied research. (Collis &

Hussey, 2014)

Besides the methodological purpose, the purpose of this paper in context of industry is to create a novel analytical framework in which the characteristics of clusters can be

(9)

systematically analyzed. Also, by providing a systematic method for analyzing clusters this paper aims to improve accuracy of triangulation of a cluster s position in its lifec cle.

1.3. Research Question

Of the research already done surrounding clusters, literature has seen a trend of measuring the effectiveness of clustering, for example the effect on regional industrial performance (Delgado et al., 2014), direct and indirect economic contribution (Sigfusson et al., 2013) and increases in regional growth rates (Porter, 1998b, The Adam Smith Address). At the same time, another part of the literature is focused on the anatomy of clusters, i.e. cluster characteristics and the different phases in the lifecycle of a cluster (Enright, 2003a; Van Klink & De Langen, 2001).

However, there is a gap in the literature surrounding the analytical mapping of these characteristics of clusters in a systematic way. At the same time, the central actor s involvement in cluster activities could influence the set of characteristics present within a cluster. Analyzing cluster characteristics and the role of the central actor can give insight into the current state of the cluster. Therefore, the following research question is asked:

What can we learn from this specific cluster by analyzing unique factors in its environment?

1.4. Case

The research in this paper is conducted by doing a single case study on a particular industrial cluster in the Icelandic blue economy, i.e. ocean related industries. Ne t to some of Iceland s big fisheries in a port in Reykjavik, the Ocean Cluster House (OC House) is situated. The OC House is a shared office space for a plethora of firms and home of the Iceland Ocean Cluster (IOC). The IOC is the company that acts as central actor in the collection of interrelated firms and is the owner of the OC House. The IOC provides office space, incubation, and acceleration services to firms within the cluster to support them in their venture processes (Sigfusson, 2020).

Since its inception, the core of the operation has been to facilitate connections. With its portfolio of services revolving around the clustered firms, the goal of the IOC is to create value and discover new opportunities by connecting entrepreneurs, businesses and knowledge in the marine industries. (IOC, n.d.).

Fishing is an important part of Iceland s heritage and Icelanders ha e depended on the ocean s resources for centuries. The seafood industry is one of the pillars of the nation s economy. The fishing and fish processing industr s part of the GDP has fluctuated since the turn of the century. In 2001 it accounted for 12% of the GDP and has fluctuated since then to just over 5% in 2007, 10% in 2011, they year which the IOC was founded and 6% in 2019 (Radarinn, n.d.). Ocean related innovations in the country have primarily focused on marine technology such as navigation and ocean monitoring as well as process innovations in fishing.

The main purpose for these innovations is to be able to export world-class produce.

(Íslandsstofa, n.d.)

In the past, utilization of the natural resource had been found to have significant waste rates. The estimation was that 8% of the fish was being discarded and that had accumulated to a 35% waste rate at the consumer level (WorldFishing, 2019). Previously, the only utilized resources were the edible parts of the fish, discarding everything else. However, regulation changes in recent years have given the fish processors the opportunity to utilize 100% of the harvest including fish by-products which enabled innovation and optimization regarding the

(10)

complete use of the natural resource (WorldFishing, 2020). The full utilization of the resource has enabled some firms to create novel products at various steps in the value chain. The founder of the IOC, Thor Sigfusson, conservatively estimates that full utilization of e.g. cod, can double the resource s monetar alue compared to lea ing out the previously discarded parts (Sigfusson, 2020, p. 114).

Clustering seems to have sparked the interest of some of the larger ocean related companies in Iceland to collaborate on an entrepreneurial level. Therefore, the interest to perform this research comes from the curiosity of what characteristics are involved in this particular cluster and what we can learn from the different entities within it.

1.5. Sustainability

As a species, our relationship with the ocean has been articulated into the term blue economy in the context of global environmental governance (Silver et al., 2015). One of the key concepts which the term encompasses is the sustainable practices of businesses in ocean related industries. Martin & Mayer (2008) describe how clusters face three main challenges regarding sustainability. Namely, these are globalization, social equity and climate change. Two of these relate to our thesis.

Clusters have proven to be accelerators of economic activity in their concentrated regions. However, it can be questionable whether this progress causes economic prosperity or greater equity in the region. We believe a cluster should focus on increasing economic prosperity with an eye for social sustainability, making sure everyone in the region benefits. In order to contribute to social sustainability, Martin & Mayer (2008, p.274) describe the importance of researching clustering and its effect on economic prosperity of corresponding regions. Subsequently contributing to the body of knowledge surrounding social sustainability.

Entrepreneurial activities are vital for creating new technologies which aim to mitigate and adapt to climate change. As agents of innovation, clusters can foster entrepreneurs in the fight against climate change. This research aims to enrich the body of knowledge surrounding clusters while contributing by giving real world examples of how characteristics of clusters affect the United Nations sustainable development goals. Specifically ho the IOC s characteristics affect goal 14 relating to Life Below Water (UN, n.d.). Sub-goal 14.A, describes the necessity to research, develop and transfer of marine technology. The characteristic of knowledge spillovers within clusters (Baptista & Swann, 1998) support sub- goal 14.A. Therefore, the research aims to include this sustainability aspect when investigating the Iceland Ocean Cluster.

The topic of this thesis relates to strategic management within industrial management by providing a tool for the situational analysis of industrial clusters. Furthermore, the systematic analysis of a cluster helps management steer the direction of the cluster s trajectory in the Cluster Lifecycle.

(11)

2. Literature Review

2.1. Agglomerated Businesses and Central Actors

Clusters are defined by many factors in the literature which together provide insight into their function. However, two aspects are always present in regions where one can speak of clustering.

First, one can speak of a cluster where there is a strong agglomeration of businesses from the same- or closely related industries. These companies collaborate on different levels, because they are normally in different steps of the value chain. For instance, a company specialized in extracting raw materials could be situated at the beginning of the value chain and collaborating with a start-up specialized in processing this specific raw material. Second, wherever agglomerations of these companies form there is usually an institution, company or government acting as a central actor with the main purpose of connecting the ties between the different companies in the geographical region (Porter, 1998b, The Adam Smith Address). In a nutshell, this is how clusters and their networks function in their most basic form. However, there is much more to it, as the literature defines clusters and its related terminology extensively. Hence, the literature review will discuss the relevant terminology, elaborate on relevant theories and combine found methods into an analytical framework in the form of a Cluster Canvas.

2.2. Terminology

Clusters are extensively defined throughout the literature. Between the numerous definitions there is much overlap. As Baptista & Swann (1998, p. 525) describe it, a cluster is an ensemble of agglomerated businesses in related industries within a relatively limited geographical area.

This is a fairly broad definition since a relatively limited geographical area leaves room for interpretation and there is no mention of a central actor within a cluster. Similar to this definition, Porter s (1998a, Clusters and New Economics of Competition, p.78) definition says:

Clusters are geographic concentrations of interconnected companies and institutions in a particular field . At first glance, both definitions look very similar. However, the definition by Porter is more elaborate in the sense that it includes institutions. Porter describes institutions as actors that provide specialized training, education and information. These functions are similar to those of a central actor as described in (Rinallo & Golfetto, 2011) therefore we assume an institution as defined by Porter can also be interpreted as a central actor.

Apart from the established definitions by Baptiste, Swann and Porter there is a more indefinite approach to defining clusters. Namely, the following approach of defining according to cluster dimensions proposed by Jacobs and De Man (1996):

1. The existence of regionally concentrated forms of economic activity within related sectors, usually connected to the central actor in the cluster.

2. Narrowly defined sectors where the different stages in the production process are mostly sequential, also called vertical production chains, e.g. the chain of supplier-assembler-distributor-customer.

3. Sectors which are defined at a high level of agglomeration, e.g. the fishing industry.

On the one hand, the approach by Jacobs and De Man is less concise. On the other hand, it can be seen as more practical, due to the fact that it is easy to either comply or not comply with the dimensions and is therefore more measurable than the other definitions.

(12)

In addition to the definitions, Porter & Stern (2001) also describe how the geological closeness between the firms in the value chain has made firms realize the increasing opportunities to innovate. Cluster characteristics foster a few very important factors which have increased firms inno ati e potential, such as knowledge spillovers (Porter, 2011, Competitive Advantage of Nations, p. 138). The phenomenon occurs when knowledge flows between firms as a side-effect of geological closeness during e.g. research and development. These knowledge spillovers are suggested by Baptista and Swann (1998, p. 526) to be the main source of a cluster s success.

While clusters are closely linked to innovation and entrepreneurial activities, there are a few concepts which are increasingly interesting to address in the cluster context. Institutional entrepreneurial activities encompass an array of different methods for harboring innovation.

One of which, acceleration, describes the different methods for aiding the new venture process (Cohen, 2013, p. 19). Furthermore, the usual method for this concept is to provide startups with seed capital, office space and to foster them for a limited amount of time to build their business.

Incubation is another type of entrepreneurial aid which aims to help firms by supporting the new venture process by providing them with opportunities through economies of scale, learning and network effects (Bruneel et al. 2012, p. 110). The European commission has defined incubation holistically as:

A business incubator is an organization that accelerates and systematizes the process of creating successful enterprises by providing them with a comprehensive and integrated range of support, including: Incubator space, business support services and clustering and networking opportunities.

(European Commission, 2002)

In contrast to acceleration programs, new ventures that use incubation services are encouraged to stick around as long as possible instead of being part of a cohort of new ventures with a fixed end-date which is common in acceleration (Cohen, 2013). In a way, incubation is more about providing new ventures with a safe place to grow, providing them with a strategic value-adding intervention system of monitoring and business assistance (Hacket, S.M. and Dilts, 2004). In return for these services, the new ventures pay monthly program fees or membership dues. Hence, the revenue model of an incubator is very different from an accelerator which normally provides the venture with seed capital with the aim to get a profitable return on the investment.

All in all, the revenue model and paradigm of an accelerator and incubator are very different, but they share the similarity of working towards the same goal. The goals of both incubators and accelerators are as Smilor and Gill (1986) formulate, to link talent, technology, capital, and know-how in order to leverage entrepreneurial talent and to accelerate the development of new companies.

2.3. The Cluster Lifecycle

Throughout the literature extensive research has been done on Cluster Lifecycles (e.g. Menzel

& Fornahl, 2009; Shin & Hassink, 2011; Van Klink & De Langen, 2001). The Cluster Lifecycle puts the concept of clustering in an evolutionary perspective. This makes it possible to analyze the different phases which industrial clusters evolve through, while taking their dynamic and

(13)

diverse nature into consideration (Van Klink & De Langen, 2001). Menzel & Fornahl (2009) describe the four phases which the development of a cluster can be characterized in the lifecycle model. The first phase is emergence, in this phase the cluster has a few members, mostly small companies. The growth stage is characterized by the growing number of employees in the cluster. Sustainment is the third phase, here the cluster maintains its relatively high number of employees stable. The fourth phase is decline, which occurs when a cluster diminishes.

In order to come to a conclusion in which phase a cluster is situated, general characteristics of a cluster need to be identified and described. Examples of these characteristics are economic interaction in a value chain, strategic inter-firm relations, and cooperative competition (Menzel & Fornahl, 2009). However, a thorough analysis and explanation of characteristics will be done in section 2.4 Creating an Analytical Framework.

Measuring and mapping these characteristics gives the indication of which phase of the lifecycle a cluster is currently located.

An illustration of the Cluster Lifecycle can be seen in Figure 1 where the position in the lifecycle is dependent on both size and heterogeneity of the cluster, as well as its maturity.

Size is directly linked to the amount of companies in and related to the cluster. Furthermore, heterogeneity relates to the knowledge, activities and competencies that are present in the cluster. It relates to whether companies have very different core businesses and knowledge or if they are more similar. On the horizontal axis, maturity indicates the age of the cluster as well as its level of development. Ideally, a cluster would have several adaptions and renewals during its lifetime. In the event of being in the declining phase, clusters can attempt transformations.

By doing so the growth curve is reignited and the cluster reinvents itself in a way and stays relevant for the industry over a longer time. In conclusion, by increasing its heterogeneity, a cluster can create the ideal situation where it oscillates between left and right sides of the model preventing the transition to the decline phase. (Menzel & Fornahl, 2009)

Figure 1 The Cluster Lifecycle

Retrieved from Menzel & Fornahl, 2009, p. 218

(14)

2.4. Creating an Analytical Framework

The constructed analytical framework incorporates cluster characteristics and stylizes them into the Cluster Canvas. It is important to note that the proposed canvas is heavily influenced by the Business Model Canvas (BMC) designed by Osterwalder et al. (2010). In summary, the BMC s purpose is to visually simplify the business model concept to influence discussions while describing nine components of the business model concept (Osterwalder et al., 2010, p.

15). The discussion is meant to improve the mutual understanding of the business model and to provide the discussants with the tools to analyze the viability of the business. The authors describe the potential for various benefits which arise from using the BMC such as creating a shared understanding within a company to increase aligned strategic direction (Osterwalder et al., 2010, p.155). This is made possible by the canvas because of its visual storytelling capabilities in going through each component of the canvas and articulating them and their internal relationships as a story (Osterwalder et al., 2010, p. 159).

However, the BMC is not void of critique. Ching & Fauvel's (2013) summary of predominant critique relates to the lack of competitive analysis as none of the BMC s components relate to competition. Furthermore, lack of goal setting as well as numerous suggestions on replacing some building blocks of the business model for others more relevant are some of the descriptions of flaws. Some of the suggestions for section replacements in the canvas relate to the oversimplification having only nine components to the canvas. Maurya (2012) has a clear example of this where he describes the need for taking uncertainty and risk into account when discussing the viability of businesses. Therefore, Maurya replaced some of the components to focus on the entrepreneur and called the new adaptation of the BMC, the Lean Canvas. These adaptations of the BMC to harness the utility of the systematic approach of the canvas as a tool influenced the decision to create the Cluster Canvas.

The intended method of filling out the BMC is to print out the canvas on a large paper and to have participants of the discussion use post-it notes to populate it (Osterwalder et al., 2010, p. 42). However, the analytical framework created in this thesis does not require this method because of the Cluster Can as focus on systematic analysis, in contrast to business model generation. For the same reason, the Cluster Canvas does not require a one-to-one mapping between key services and diversity. Again, contrasting the BMC s one-to-one relationship between the value proposition and market segment components (Osterwalder et al., 2010, p. 22).

The purpose of the Cluster Canvas, shown in Figure 2 The Cluster Canvas, is systematic analysis by visually distinguishing different components of a cluster s characteristics to simplify the cluster dynamic while including the context of the central actor.

B doing so the frame ork s intended outcome is to increase the perception of the o erall harmony between the different components of a cluster s paradigm.

The characteristics shown in the canvas have been specifically selected from the literature to reflect ten important components to systematically analyze a cluster. The different characteristics are extracted from literature and their intended representations in the cluster canvas described.

(15)

Figure 2 The Cluster Canvas

The cluster s borders are defined b using Enright s (2003) geographic scope in combination with the geographic reach of the cluster. The reach of the cluster is therefore defined by the different suppliers and consumers as well as all the actors in the cluster s network. This contrasts the locale of the cluster itself and puts it into the context of being a node in a global network (Martin & Mayer, 2008, p. 273). Additionally, the borders could be influenced by the central actor s network building, e.g. by building relationships with linked clusters in foreign economic areas.

Clustered firms can be heterogeneous in terms of age, employee size, economic weight and maturity (Van Klink & De Langen, 2001). This is described by using the term diversity in the canvas. Its purpose is to map out the different types of firms in the cluster and to categorize them by their differences. These characteristics are also a part of the canvas to distinguish the different types of companies inside the cluster to aid with contextual information, such as the relations between certain characteristics and type of organization.

Breadth relates to the horizontally related industries which share commonalities such as technologies and end-users (Enright, 2003, p. 102). By defining the industries which make

(16)

up the cluster, the central actor can observe the boundaries for which firms belong in the cluster.

Furthermore, observing the breadth of industries could influence discussion on the opportunity for the cluster to seek participation of other industries.

On the other hand, depth describes the cluster s different vertical industries and their share of the value chain entities (Enright, 2003, p. 102). However, the activity base describes the steps of the value added chain activities present in the clustered firms (Enright, 2003, p.102).

These activities describe the level of value created within the cluster itself and can give insight into both where added value is draining or even give opportunities for innovative value creation.

Furthermore, missing activities can give the cluster the opportunity to determine which activities should be added to the cluster itself, by participation of previously external firms or an entrepreneurial activity to fill in the void. Therefore, depth in the canvas relates to the value chain and combines these two of Enright s characteristics. B isuali ing the cluster s alue chain, it is possible to give insight into how the cluster wants to proceed in terms of strategic direction. This could mean trying to fill a void in a supply chain by recruiting relevant companies to join the cluster.

Inter-firm relations describe the act of cooperation by the development of coordinated strategies between firms in the cluster (Van Klink & De Langen, 2001, p. 451). These relations describe the degree of cooperative capacity. Interactions are a vital part of the cluster dynamic.

Therefore, the characteristic is immensely important to the central actor and the quality of relationships should be monitored closely throughout the lifecycle.

Co-operative competition between firms in a cluster relates to how clustered competitors cooperate in the cooperative domain (Van Klink & De Langen, 2001, p. 451).

Therefore, the characteristic maps out both the existence and degree of competition in the cluster as well as relationships between the competitors. Competition is a vital part to increase productivity growth and cluster success (Porter, 1998b, The Adam Smith Address). Therefore, the central actor has a crucial responsibility to monitor the competitors cooperative domain.

Technological activities describe the clustered firms technological aspects in terms of using, adapting or developing new technology (Enright, 2003, p. 102). The different firms involvement in technological evolution can be shared among the different organizations. For example, one developer of new technology inside or outside the cluster can influence another firm to adapt to that same technology. How these firms operate in terms of technological development and usage could influence e.g. cooperation and even buyer-supplier interaction within the cluster.

Key services represent the services which the central actor offers to different entities inside and outside the cluster. Services offered to clustered firms can be matched to fit the different types of firms. For example, some services can be fitted to the needs of small to medium sized firms and others to larger organizations. The key services can also provide value propositions to clustered firms as part of the perceived benefits of actively participating in the cluster. In this paper, key services are proposed to be a characteristic which is important for the central actor and its role in the cluster.

To illustrate how the central actor operates as a business, the analytical framework includes financial aspects, namely the cost structure and revenue model of the organization.

These characteristics are extracted from the BMC (Osterwalder et al., 2010) to the Cluster Canvas. The cost structure entails costs surrounding the key activities and operation which the cluster has which relate to the central actor. These costs include the costs for the resources and

(17)

activities which will bear the largest toll on the organization. Just as in the BMC, the cost structure is compared to the revenue streams, i.e. how the company generates revenues from different sources. These can come in many forms, such as revenues from its services and funding. The cost structure and revenue streams are included in the canvas to help with visualizing how the central actor operates as a business.

2.5. Critique of the Analytical Framework

The Cluster Canvas is not free from imperfections and shares many of the same flaws as the Business Model Canvas. One such critique is the fact that the different components of the canvas are on different levels of abstraction as Kraaijenbrink describes the BMC (Ching &

Fauvel, 2013, p.29). In the cluster canvas, the components which rest on different abstraction levels are the three different sources of components. Enright s (2003) characteristics which are imported into the canvas, as well as diversity, describe the characteristics of the collection of clustered firms. Ho e er, Van Klink s characteristics describe the clustered firms internal relationships. Finally, cost structure and revenue streams which are taken from the BMC as well as key services are on the final abstraction level of internal workings of the central actor in the cluster. Despite this abstraction misalignment we believe that the canvas provides an analytical method of describing a cluster while benefiting from the contextual relations of components at different abstraction levels.

In the creation of the canvas, some characteristics from literature have been intentionally left out for different reasons. Most commonly though, the reason for exclusion was that we believed some of those characteristics were not in line with creating a systematic tool to analyze the cluster dynamic. For example, density and ownership structure were excluded.

Density describes the cluster s market share in the different associated industries (Enright, 2003, p. 102). Anal ing the cluster s density can give insight into where its industrial focus is and an opportunity to assess their strengths and weaknesses when it comes to collaborations in spinoff creation. However, for the sake of simplification in the canvas, this characteristic was left out since this analysis requires a different methodology altogether and should done independently from filling out the canvas.

The ownership structure reflects how the foundation for the clustered companies are built and hat stakeholders there are in the cluster s immediate icinit in regards to equity (Enright, 2003a). O nership structure can also characteri e the central actor s monitoring of primary stakeholders of participating firms. Ownership structure was not deemed canvas material on its own because of its lack of more practical use cases. However, ownership structure is considered in the context of the inter-firm relations component of the canvas.

The canvas is not intended to be a tool which can give an exhaustive description of clusters. It takes its ideology from the BMC to facilitate mutual understanding, encourage discussion and facilitate codification of previously tacit knowledge within clusters.

(18)

3. Methodology

The approach to the process of this research is through a combination of methods. The methods are combined with the aim to result in a fitting answer to the research question, while maximizing the utilization of the available data within the scope of the research. The applied methods and process for data collection are explained in this chapter and aim to give insight into the reliability and validity of the research.

3.1. Research Design

First, the research adopts a single opportunist case study method. A case study is considered to be a suitable strategy for analyzing, understanding and describing a phenomenon in its context and uses a variety of methods to obtain in-depth knowledge (Bonoma et al., 1981). In this thesis, only the case of the IOC is investigated, and no other clusters are included in the scope of the research. Hence, we can speak of a single case study. More specifically, this research adopts an opportunist case study, i.e. the opportunity arises to examine a phenomenon because there is access to a particular business, namely the IOC (Otley & Berry, 1994). Furthermore, the empirical data is analyzed using the analytical framework, i.e. the Cluster Canvas, that was derived from theory at the end of the literature review. In an attempt to analyze and explain which lessons can be learned from this specific cluster, the strategy of Building Explanations is used. Yin (1994) suggest this analysis consists of three parts:

An accurate rendition of the facts of the case

Some consideration of alternative explanations of these facts

A conclusion based on the single explanation that appears most congruent with the facts

Single-subject designed research, such as the single case study, have certain limitations which will be elaborated on in section 3.3 Reflection on Methodology. Despite the limitations and taking the nature and depth of a Master s thesis into consideration, the results can still be extremely stimulating and original (Collis & Hussey, 2014).

Second, the research entails a qualitative study, where empirical data is collected that is not expressed or analyzed in numbers. Three levels and two methods are distinguished in collection of the empirical data. On the highest level, data is collected through semi-structured interviews with the management of the central actor. A semi-structured approach is used to be able to customize the interview to the interviewee, leaving out irrelevant questions and adding questions to specific topics when deemed necessary. The aim during the interviews is always to create a reflection of the most complete situation. Parallel to the interviews, surveys are sent out to several tenant firms connected to the cluster, these firms were specifically selected based on diversity. The survey is used as a tool to efficiently collect primary data from a sample of the tenant population. The population in this case are all the tenant firms which have resided in the OC House. In the past three years around 40 start-ups have been created in the OC House, today there are 25 start-ups present. The sample on this level of the cluster consists of 5 small companies which are active in different fields in the blue economy. To contrast the findings of tenant firms, a sample is taken of the large companies which have played a major role in the Icelandic fishing industry for decades. These companies are also referred to as big players in the industry and belong to the peripheral firms, i.e. firms in the cluster residing outside of the OC House. Examples of these are large fisheries and fish processing companies. The questions asked to these companies are similar to those asked to the start-ups to be able to analyze and

(19)

compare their perspectives. However, these large corporations require a slightly different and, in a way, a more delicate approach. Therefore, the questionnaire is converted into a semi- structured interview format and will be conducted via video conferencing applications in either Icelandic or English, depending on the individual inter ie ee s preference. To summarize, although the questions are similar for all companies in the cluster, their respective data collection methods are distinguished into tenant firms and peripheral firms.

After data collection, all the information is analyzed qualitatively. A survey can be analyzed quantitively. However, the survey was designed to include multiple open-ended questions in combination with Likert scale questions regarding perceptions on various characteristics. This makes it possible to compare the outcomes of the survey to the outcomes of the interviews efficiently (Collis & Hussey, 2014, p.215). The research question focuses on finding unique characteristics inside the cluster. Therefore, the research does not aim to generalize the findings to a higher level, but rather identify and analyze unique examples of these characteristics in the cluster. Taking the analytical level of the cluster and comparing it with examples from the individual firm level. Consequently, the research takes a qualitative approach to gather data from subjects by putting insights into context and enables them to elaborate further. Quantitative methods cannot answer the research question in the same in- depth way and were therefore omitted.

In order to analyze the data using a non-quantifying method, it is necessary to reduce the data. This research uses continuous data reduction as method for distilling the findings with a focus on the analytical framework which was created in the literature review. Although this is a systematic method for analysis, we must be attentive to stay non-biased and let go of our own frame of reference when assessing the data. (Collis & Hussey, 2014)

3.2. Research Classification

The logic of this research moves from general to specific, following a deductive approach.

This approach allows to first analyze the existing body of knowledge. After a broad introduction into the thesis topic, the literature review gives insight into the current body of knowledge and where there is a potential gap in the literature. The literature review starts general and funnels down to specific theories which can be directly implemented in the analysis of this thesis, for example cluster characteristics and the Cluster Lifecycle theory. Next, the methodology chapter builds an argument for validity and reliability of the research by elaborating on the methods used. After methodology, the thesis dives further into the research topic in section 4. Findings, where the outcomes of the analysis are discussed. Finally, the findings are condensed into conclusions and an outlook to future research which is the most condensed and topic specific part of the thesis.

3.3. Reflection on Methodology

Several aspects of the research have a positive effect on reliability and validity of the outcomes.

However, it is also important to be reflective of certain choices and recognize limitations of the research design. Hence, both positive and negative aspects of the research are reflected on in order to acknowledge its limitations and strengths.

On the one hand, triangulation of data is present throughout the research, strengthening the reliability and validity of the research. Data is collected from different sources in the study of the phenomenon, augmenting the reliability of the research. Triangulation is apparent when

(20)

information is retrieved on different levels within the cluster, more specifically on the levels of management, tenant firms and peripheral firms. Gathering different perspectives on the research problem and cross analyzing the findings make the conclusions of the research more accurate (Easterby-Smith et al., 1991). Additionally, triangulation is present in the way data is collected. Primary data for the research is collected using multiple methods, namely by conducting in-depth interviews in a structured and semi-structured format in combination with a survey. Secondary data for the research is collected from the book, The New Fish Wave (Sigfusson, 2020), b the IOC s founder and used to enhance perceptions of certain general information. This approach to collecting data is referred to as methodological triangulation.

Finally, after collection, the outcomes are cross analyzed in a qualitative manner.

On the other hand, triangulation poses limitations to the research as well. Since qualitative data are generated with the use of multiple methods it is more difficult to replicate the research and its outcomes. Furthermore, the research follows a single-subject design, this type of research design does not allow for generalizations of the outcomes. Although the findings of the research only apply for the specific case of the IOC, they can still be stimulating and original.

The characteristics in the canvas were selected from the large pool of literature. The criteria for selecting the specific characteristics, which were imported to the canvas, were not empirically tested on applicability for all clusters. They were selected based on applicability for gi ing insight into the cluster s lifecycle and to answer the research question.

The delimitations for our research can be described through the scope of data gathering, time of occurrence and interpretation. As a single case study, the scope for the data gathering is set to the three identified levels within the specific cluster. These three levels were chosen by us to represent the key entities within the cluster which can give insight into the nature of our research question. The scope for our subjects is therefore set to individuals belonging to firms connected to the cluster. The scope of time relates to how examples are interpreted in relation to when they occur. The scope for the questions created for data collection were designed to leave the time of occurrence open to give the subjects space to elaborate. The scope of interpretation is set by analyzing the data through the analytical framework. Each factor found in the research is therefore interpreted through the lens of the characteristics in literature.

3.4. Ethics

When conducting research, it is important to take responsibility for ethical concerns, and prevent moving into unethical practices. Conducting research in collaboration with a company takes an additional perspective on ethics. Generally, researchers are grateful for being able to work with a company and the facilitation of data for the research. Therefore, there is usually a good relationship between company and researcher(s). However, ethics must be added to this equation. For this research specifically, empirical data is collected inside the company itself and within its network. The situation could occur that the research uncovers unethical practices b the compan or companies in the cluster s net ork. During data collection attention has been paid to ethical matters within and outside the company. We were pleased to find no unethical practices by the company or its partners in the network during data collection.

It is important to note that the first connection with the IOC was made through the family relations between the founder of the organization and one of the authors of the thesis, Thorsteinsson. However, the authors feel that this connection has eased communications

(21)

between the organization and the researchers and has not influenced the objectivity of the research.

Furthermore, to ensure that subjects who took part in the research felt they could speak freely and to ensure the highest possible quality of data we anonymized the interviews and questionnaires that were conducted. This felt like the appropriate step towards confidentiality in the research. The names and roles of the subjects are known with the authors of the thesis.

(22)

4. Findings

The empirical findings which are presented in this section are gathered from the primary data which were gathered during the research. This data is qualitatively analyzed and presented in two distinct ways while utilizing the Cluster Canvas as a presentation tool. To illustrate the findings, they are imported into the canvas shown in Figure 3 Populated Cluster Canvas. There, the summary of each factor of each component is represented as a condensed bullet point. The canvas is therefore the presentation tool of the cluster dynamic. Subsequently, the cluster dynamic is explained through what Osterwalder et al. (2010, p. 159) call, the story. This is done with the help of the canvas by going through each bullet point in each component in an algorithmic fashion and describing how they relate to other components.

Figure 3 Populated Cluster Canvas

(23)

4.1. Filling in The Canvas and Telling The Story

The following paragraphs describe the characteristics of the IOC, analyzing the findings from our research. This is done by going through each component of the Cluster Canvas, beginning with its borders.

Throughout the section, references are made to the transcriptions of the interviews and questionnaire, these can be found in the appendix. The letters A-G refer to the interviewees and the numbers refer to the interview questions. The letters U-Z refer to the participants of the questionnaire and the numbers refer to the different questions.

In Iceland s blue econom there resides an industrial cluster, located in the capital region of Reykjavik. Situated there is an economic center of gravity for the Icelandic blue economy, an agglomeration of companies tied to the seafood industry. The epicenter of the cluster is the Ocean Cluster House (OC House), located on the docks of Grandi, Reykjavik.

There are 60 different firms residing in the OC House and 50 more are located in the capital region (A, 2.1). Although there is a clear focal point of companies in the capital region, there are also some others situated in other municipalities (D 1.5; E, 1.5). Besides this concentrated region, the borders of the cluster span internationally. Several firms in the cluster have a global geographic span of sales and this is especially evident with the big fisheries in the cluster. Some of these fisheries own equity in sales offices abroad (F, 2.2; D, 2.2) and even have inventories in foreign countries in order to be closer to potential clients (D, 2.2). One interviewee noted that 99.9% of our production is e ported out of the countr (F, 2.1) when describing the fisher s geographical span of sales. Regarding the central actor s borders, the reach of the cluster s net ork is global as ell. This is illustrated by the relationships with other blue economy clusters worldwide, e.g. the cluster has a strong formal relationship with clusters in the United States one in New Bedford, Massachusetts, which the IOC partly owns and another in New England (A, 2.4). Furthermore, they have relations with clusters in the Nordic countries, such as a seafood innovation cluster in Norway and a transportation and logistics cluster in Denmark (A, 2.4).

The diversity between many of the clustered firms can be redirected back to their location, employee count, maturity, and core activities. These firms have been categorized into two categories. Firstly, tenant firms which are small to medium sized firms with 1-10 employees, some of which are start-ups (C, 4.8). Most of the firms in the OC House are small- to-medium sized firms. These differ in terms of maturity although there is a trend of being relatively young and being founded after the cluster s formation (A, 1.1)(V,X,Y,Z, 2).

Secondly, peripheral firms reside outside the OC House and tend to be larger. Some of these peripheral firms are however very different organizations than the ones residing in the Ocean Cluster House. These are the incumbent peripheral fisheries and fish processors. They are established organizations whose heritage include decades of fishing in the Atlantic Ocean.

Additionally, as incumbents, these organizations are drastically different to the smaller residents of the OC House regarding their number of employees (D, 1.3, F, 1.3). Additionally, the diversity of the clustered firms can also be in terms of industry, vertically or horizontally heterogenous.

Together, the different firms in the cluster span a variety of industries, i.e. breadth, which relate to the blue economy. First, the fisheries are in the seafood industry as their key activities relate to fishing, fish processing and sales (D, 1.2; F, 1.2; E, 1.2). Additionally, the cluster includes consultancies which aid the fisheries in technological endeavors such as the

(24)

creation of new fishing vessels and ocean technology (B, 4.8). However, from that industry there sprout various companies in other related industries, directly and indirectly related to the seafood industry. For example, a collagen manufacturer in the cluster now creates collagen from the fish skin which previously used to be a waste product or only utilized to create fodder.

However, by harvesting the collagen, the company gives other industries the opportunity to utilize previous waste from the seafood industry (G, 3.2). These industries include pharmaceuticals and food industries as the collagen is used to make food glue and food supplements (G, 3.8). Other industries which have presence by the clustered firms include but are not limited to agriculture, transportation and logistics, marketing and distribution, health, beauty and biotechnology. (B, 3.1, C, 3.1)

Of the different depth of industries in the cluster, the steps in the value chain covered by the clustered firms vary between industries. We could not find examples of firms covering the whole value chain, although some firms such as the fisheries sell the fish they caught and processed directly to the retail market leaving only the trivial step of reaching the end-user (D, 1.5, 2.1; F, 3.8; E, 1.2). The depth as seen from the fisheries perspective is demonstrated with the technological developments they buy from other clustered firms (F, 3.9). However, the utilization of the fish is a great example of another depth dependency where the collagen manufacturer depends on the fish skin from the other fisheries to create its product (G, 3.10).

However, of the tenant firms which took the questionnaire, not a single firm answered that other clustered firms were dependent on them in terms of supply chain (V,X,Y,Z, 21). One of the firms cover almost all of the value chain straight by Gro ing and selling fresh asabi (Z, 3), another handles the latest steps in the value chain directly to the consumer market (V, 3).

The other two differentiate themselves by providing a different kind of value to their customers.

Company X is in business development and company Y is in process innovation, extending the shelf life of food (X,Y,3).

The different steps in the value chain have however created the opportunity for clustered companies to interact and grow inter-firm relations. The geographical closeness of firms in the OC House have given them the opportunity to frequently interact with each other and collaborate. This is evident as the central actor s research suggests that 70% of the companies in the OC House have collaborated with one or more companies in the OC House (C, 4.8). An interesting example of this is the collaboration between an established maritime engineering consultancy and a nanotech startup who have been working together to develop electric fishing vessels (A, 4.1). Another factor influencing inter-firm relations inside the OC House is the weekly Friday-Coffee event. Each Friday, one firm is asked to prepare a meetup and invites the other tenants to an introduction of their business as well as to build their network within the cluster (B, 4.6). The tenant firms which completed the questionnaire described the effect of their interactions with each other as relatively fruitful although the quantity of deep relationships was not consistent between them (U,V,X,Y,Z, 10 & 11). The tenant firms show a trend of forming relationships with the other tenant firms because of their willingness to interact with each other. One tenant firm noted how the greatest benefit of being a resident of the OC House was the informal interactions and subsequent collaborations between the tenants.

These informal interactions were the result of the central actor s projects revolved around encouraging collaboration in the cluster, but their perception is that these projects are no longer being executed which may have lowered the overall collaboration within the cluster (Y, 18).

Another firm noted that they wanted to see the emergence of focused meetups, similar to the collaboration projects held by the IOC, but to focus these projects on closely related companies (U, 18). Contrasting the characteristics within the OC House, the peripheral firms in the cluster do not benefit from interactions which stem from the extreme proximity inside the OC House.

(25)

Their interactions appear more explicit and formal when it comes to the interactions with the residents of the OC House (B, 4.6). Interactions between peripheral firms and tenant firms have often been initiated through facilitation by the central actor but also by the peripheral firms themselves by directly contacting other clustered companies (B, 4.6; C, 4.11). The central actor also has an important role of initiating interactions where they see possibilities (B, 4.6).

On the tenant level, the co-operative competition has enabled interactions and collaboration to a certain degree. One firm s e perience as that the cooperation as maintained through implicit methods such as giving advice to competitors for the sake of openness (V, 20). However, the shared venture between the fisheries is perhaps the highlight of the cluster s cooperati e competition. There is a shared understanding between some of the fisheries within the cluster that they are not competing amongst each other, at least not to a high degree (D, 2.1, F, 2.1). The reason for this is that they compete in a vast global market with an abundance of opportunities and therefore the term competition, is not descriptive of their relationship. When asked about competition, one key individual at one of the fisheries noted Iceland is so small on a global scale that there s enough market share for e er one (D, 21). However, these fisheries look at foreign competition through a very different lens. New innovations in ocean technology, developed in Iceland, have given the Icelandic fisheries competitive advantages in the global market. However, because of recent governmental regulation regarding specific fishing taxation, the innovative openness of the fisheries gives foreign competitors the opportunity of copying their competitive advantage and utilizing it more successfully. This has been made possible by the foreign competitors by using the same new technology as the Icelandic counterparts and utilizing it more successfully since they do not adhere to the same regulations. This in turn, may have reduced the fisheries illingness to collaborate openly (D, 1.6). Not everyone in the industry agrees that the fisheries are not competitors, one fishery suggested that the big players have a competitive advantage in pricing which they often make use of when reacting to volatility in the market (E, 2.1). Even so, an example of a company founded within the cluster which combines the involvement of the central actor and the collaboration of the competing fisheries and other ocean related companies is the previously mentioned collagen manufacturer. The company was created through the mutual willingness to collaborate for the sake of innovation (F, 1.6), shared prosperity (D, 2.1), and sustainability (G, 2.3).

These new ocean technologies are one of the many examples of technological activities performed by the companies within the cluster. Several of the residents of the OC House are developers of technology (B, 4.1). In contrast, the fisheries and peripheral firms are focused on their core business and source their research and development to other companies (D, 3.9; F, 3.9). However, the fisheries are adapters of new technology from companies in the cluster when opportunities and needs arise. Furthermore, these investments tend to be process innovations which increase the efficiency of their same core activities (F, 2.4). Through the proximity of firms in the region, knowledge diffuses to other companies within the cluster (C, Extra) and even crosses oceans to friends in foreign countries (G, 3.9). Some of the innovations in the cluster have been applied from other industries. An example of this is the technology used to extract gelatin from pigs and applying it to the extraction of collagen from fish skin (G, 1.4, 2.2 3.9).

Since forming the cluster, the IOC has focused on connecting the clustered firms as the central actor in the network. At the OC House, the IOC provides a portfolio of services to their tenant firms. Those firms receive one of the IOC s key services, incubation (U,V,X,Y,Z, 18 &

36). The incubation includes shared infrastructure which includes shared office space, meeting

(26)

rooms and a shared kitchen. This infrastructure enables the tenant firms to operate their business in proximity with each other, within the OC House. As previously mentioned, this pro imit is here man of the tenant firms interactions within the OC House as they pass each other on the hallways and in the kitchen on a daily basis (B, 4.6).

Some of the tenant start-ups are also part of the central actor s acceleration services (C, 1.1). Of those, the IOC provides seed capital, office space and business support to those operations (G, 1.2, 1.5). Furthermore, they have started a new project surrounding acceleration of the development of ideas. The IOC plans to recruit teams of students over the summer to develop several blue economy business ideas and provide the teams with office space in the OC House (A, 4.4). a key individual within the cluster describes that one of the problems which the organization has been having is the link between idea development and execution. There is no shortage of innovative business ideas and projects floating around in the cluster. However, the capacity for execution and finding the right people to lead them is more important.

Therefore, in this specific case finding the right people for the execution of ideas is of more importance than looking for people with ideas (A, 4.4). The collagen manufacturer is a joint venture between the different ocean related companies in the cluster and the IOC. The idea development process was a collective effort by the different cluster members and the IOC (G, 1.2). Subsequently, the project was kick-started by recruiting key personnel to execute the project in a separate collectively owned spinoff with equity shared evenly between the ocean related companies as well as being partly owned by the central actor. Additionally, the Spanish gelatin plant also received shares in the company for their help in the execution of the idea in terms of knowledge regarding e.g. technological activities (G, 1.4).

The IOC has also led consultancy projects for institutions and ocean related companies (D, 3.11). One of those projects was to find a suitable location for an import-export harbor on the coast of Iceland. The project was done in association with the fisheries, large freight organizations as well as the harbor administrations in the capital region (F, 2.7). These projects have also consisted of doing particular research in the industry and answering questions for the fisheries in the cluster (D, 3.11). Furthermore, some specific projects have been for Icelandic governmental agencies (A, 4.5). In addition to those, the central actor has also been worked with other clusters abroad such as the New Bedford Ocean Cluster (A, 2.4).

The cluster s net ork management is interesting to look at because of net orking s prevalence in the cluster literature. The central actor manages the cluster s net ork in an implicit manner. Connections between the IOC s management and other companies are tacit and personal (C, 4.9). The connections with other clusters worldwide have been made through the relationship between the IOC s management and key individuals in other clusters. This has meant that in the event of staff turnover wherein the former key individuals are no longer part of the foreign cluster, relationships with the clusters also fade away and connections need to be remade (A, 2.4). Another interesting factor to look at is the fisheries collecti el o ned company and its relationship with individuals in the cluster. Key individuals in the company have had people reaching out to them with ideas of cooperation with the fisheries. Subsequently the company has in some cases connected these individuals with e.g. CEOs of the fisheries, making them a different kind of liaison to the incumbents than the central actor (G, 3.7).

However, there is another service which the IOC provides which somewhat structures the cluster s network. This is the newsletter which is sent out to all clustered companies on a regular basis (A, 4.9; B, 4.9). On a critical note, two of the tenant firms suggested that the cluster was lacking an open database consisting of the clustered firms and their core activities,