12

Click here to load reader

Capturing processes in longitudinal multiple case studies

Embed Size (px)

Citation preview

Page 1: Capturing processes in longitudinal multiple case studies

Industrial Marketing Management 41 (2012) 235–246

Contents lists available at SciVerse ScienceDirect

Industrial Marketing Management

Capturing processes in longitudinal multiple case studies

Lise Aaboen ⁎,1, Anna Dubois, Frida LindDepartment of Technology Management and Economics, Chalmers University of Technology, SE-41296 Göteborg, Sweden

⁎ Corresponding author.E-mail addresses: [email protected] (L. Aaboe

(A. Dubois), [email protected] (F. Lind).1 Tel.: +46 31 772 10 00.

0019-8501/$ – see front matter © 2012 Elsevier Inc. Alldoi:10.1016/j.indmarman.2012.01.009

a b s t r a c t

a r t i c l e i n f o

Article history:Received 6 November 2010Received in revised form 8 December 2011Accepted 14 December 2011Available online 28 January 2012

Keywords:Multiple case studyLongitudinalProcessesTimeNetworks

This paper discusses and suggests a methodological approach to capture processes in multiple case studies.While single case studies may address processes in an inductive or abductive manner, multiple case studiesentail a stronger need for analytical frameworks and methodological tools developed ex ante in order tomaintain focus on the same phenomena across cases and over time. This is, however, difficult to pursue instudies that focus on processes. The aim of the paper is to suggest an approach to longitudinal multiplecase studies. We use an example of an on-going multiple case study aiming at capturing the processes ofstrategizing in terms of how the networking of start-up companies interplays with their ideas of future net-work positions. The paper concludes by suggesting the use of a combination of narratives and network draw-ings. Network drawings can be instrumental in capturing the past, the present and the future at differentpoints in time for the individual cases. Based on these descriptions, change patterns and how these evolvecan be analyzed and compared across cases and over time.

© 2012 Elsevier Inc. All rights reserved.

1. Introduction

In this paperwe suggest and discuss amethodological approach thatcan be applied to investigations of patterns in inter-organizational pro-cesses. We address the combined challenges involved in studying mul-tiple cases and processes that unfold over long time periods. Thetheoretical framework chosen to address the empirical phenomenonis rooted in the industrial network approach (see e.g. Håkansson,Ford, Gadde, Snehota, & Waluszewski, 2009); it thus relies on thebasic assumption of companies as interdependent and interactingwith other companies. This choice has implications both for the discus-sion of the methodological approach per se and for the multiple casestudy that we use as an example.

Halinen and Törnroos (2005) distinguish four major challenges ofcase research for research on industrial networks: the problem of net-work boundaries, the problem of complexity, the problem of time,and the problem of case comparison. It can be argued that everycase study approach must offer ‘solutions’ to these problems. First,the problem of network boundaries exists because network settingsextend without limits through connected relationships, whichmakes any network boundary arbitrary. Second, the problem of com-plexity is, according to Halinen and Törnroos (2005), attributable tostructure and embeddedness and to describing a network with allits actors and all the characteristics of the links between them.Third, the problem of time is always involved in industrial network

n), [email protected]

rights reserved.

research since the networks are subject to constant change. To pro-vide valid descriptions of network processes, the concept of timethus has to be incorporated into the research. Fourth, the problemof case comparisons arises because each network case is unique as aresult of context specificity and historical background, and thereforeit is difficult (or even impossible) to compare with other cases.

The theoretical grounding of the industrial network approach hasmainly been developed on the basis of a large number of in-depth sin-gle case studies permitting thorough description and analysis of net-work processes and structures (see e.g. Baraldi, 2003; Gressetvold,2004; Håkansson & Waluszewski, 2002; Holmen, 2001; Lind, 2006;Lundgren, 1995;Wedin, 2001). Quantitative studies have been scarce,although the whole research tradition took off from a study of 900cases of business relationships (Håkansson, 1982). Some multiplecase studies embedded in single case studies (see e.g. Hjelmgren,2005; Hulthén, 2002) have been carried out, but these typically donot rely very much on case comparisons, instead emphasizing the va-riety of (sub)cases in a way that relates to ‘casing’ them as singlecases, i.e. instances of a particular phenomenon.

One question then is whether it would be possible and useful to de-velop methods that allow case comparisons while still maintaining thebasic theoretical assumptions of industrial networks as consisting ofinteracting and interdependent firms. Another related question iswhat we would have to ‘sacrifice’ in order to make such case compari-sons possible.

Any study can be described in three dimensions: the empiricalphenomenon, the theoretical framework and the method applied.The choices made in each of these dimensions are always interrelated(see e.g. Dubois & Gibbert, 2010). In this paper we use an ongoingmultiple case study as an example and suggest a methodological

Page 2: Capturing processes in longitudinal multiple case studies

236 L. Aaboen et al. / Industrial Marketing Management 41 (2012) 235–246

approach that would enable a longitudinal study of the cases. Thestudy focuses on a particular empirical phenomenon (the strategizingprocesses of start-up companies) and applies a particular theoreticalframework. Discussing the methodological approach without consid-eration of the empirical phenomenon and the theoretical frameworkwould be problematic, because of the ways in which these dimen-sions are interrelated. Still, we aim at using this particular setting toillustrate a methodological approach that might be used for other lon-gitudinal multiple case studies of processes taking place in open sys-tems such as industrial networks.

The aim of the paper is to propose an approach to longitudinalmultiple-case studies. The suggested approach relies on a combinationof narratives and network drawings in the analysis of individual casesand on comparisons of evolving change patterns across cases and overtime. At the end of the paper we will return to how the suggested ap-proach relates to the fourmajor challenges of case research for researchon industrial networks pointed out by Halinen and Törnroos (2005).

The paper is structured in the following manner. First, we elaborateon three key notions in a theoretical review: multiple case studies, con-ceptualizations of time andprocesses, andnetwork pictures andnetworkdrawings. Second, the example of a multiple case study is presented anddiscussed. Third, we discuss some findings from the study and how oursuggested methodological approach can capture processes in longitudi-nal multiple case studies. Finally, we re-address four major challengesof case research and some of the problems associatedwith our suggestedapproach.

2. Theoretical reviews

Three theoretical notions are vital to the methodological approachthat we suggest. First, multiple case studies are addressed with regardto how these differ from single case designs that have been recognizedfor their suitability in studies of processes and change. Second, concep-tions of time and processes are vital in all longitudinal approaches. Here,we put special emphasis on how to capture perspectives on the past, thepresent and the future. Third, we coin the concept of network drawings,that is essential in our approach, and discuss how these differ from no-tions on network pictures.

2.1. Multiple case studies

Case studies, single and multiple, are generally described as usefulfor theory development. Furthermore, they are described as a researchstrategy which focuses on understanding the dynamics present withinsingle settings (Yin, 2003; Eisenhardt, 1989). According to Eisenhardt(1989, p. 546): “… attempts to reconcile evidence across cases, typesof data, and different investigations, and between cases and literatureincrease the likelihood of creating reframing into a new theoretical vi-sion.” In a rejoinder to Eisenhardt, Dyer and Wilkins (1991) argue thatbetter stories, and not better constructs, are needed to generate theory,and that to achieve this the rich context of each case needs to be includ-ed. They argue that by contrasting observations frommultiple cases theattention is focused on general constructs, and not on the context of theconstructs and the role these constructs play in a particular setting.Ragin (2000) suggests ‘small-N’ research as a compromise in whichmultiple comparative case studies are carried out with a focus onexplaining causal paths that produce particular outcomes for eachcase. Hence, maintaining a focus on in-case analysis in the study of sev-eral casesmay have its virtues in particular circumstances. For instance,it could facilitate identifying and analyzing particular patterns in certainprocesses, and explaining differences across cases by contrasting themto each other.

The advantage of single case designs in relation to multiple cases isthat single case studies permit inductive or abductive approaches inwhich the theoretical framework can be adapted and developed in in-teraction with the empirical case (Dubois & Gadde, 2002). The

disadvantage, on the other hand, may be that the single case becomestoo context-specific, which limits the possibilities to generalize toother settings. That is, the boundary between the phenomenon andthe context may become obscured. In contrast, multiple case studiesmay, based on the variety between the phenomenon and contexts, con-tribute both to a better understanding of the interfaces between phe-nomenon and contexts and also to identification of different patternsin the interplay between them. However, multiple case designs requirepre-structured frameworks to enable some form of case comparison.Whilemultiple case designs can be approached in differentways, it can-not be subject to the same process of direction and redirection as singlecase designs (Dubois & Gadde, 2002).

2.2. Conceptions of time and processes

Time is central to any study of processes and can be captured inmany ways. Clock time or absolute time is the interpretation of timethat has evolved through history to become referred to as the objec-tive view of time (Bluedorn & Denhardt, 1988). The objective viewof time is the basis for research that regards time as a resource, suchas studies of efficiency (Das, 1991). Time is then viewed as somethingthat can be spent, saved, wasted and invested (Lee & Liebenau, 1999).Within the objective view of time, events can be organized as takingplace one at a time or as doing several things at a time (Bluedorn &Denhardt, 1988).

Fried and Slowik (2004) found that an increasing number ofscholars are arguing that clock time should be complemented witha more relativistic view of time that is often referred to as subjectivetime. This is socially constructed with a view of time as the core ofculture as well as social and personal life (Lee & Liebenau, 1999). Insubjective time the past can be recollected, the present perceivedand the future anticipated (Shipp, Edwards, & Lambert, 2009). Intheir review of research about time, Bluedorn and Denhardt (1988)also found related divisions between objective and subjective time,such as even time versus event time. Even time is what is referredto above as objective time, while event time is framed by meaningfulevents related to seasonal variations.

Butler (1995) discusses time frames as affecting decision-makingand learning processes. Different time frames can be used at differentgeographical locations. However, there can also be different timeframes in use at different levels in an organization. For instance, thetime spans associated with an assignment may be shorter at thelower levels of the organization while they are longer at the morestrategic levels. Furthermore, individuals may differ in their orienta-tion towards past, present, future and length of time horizon(Bluedorn & Denhardt, 1988). In order to plan for the future, individ-uals extrapolate the past and the present. The time span whose dyna-misms and evolution the individual can foresee will then differ.Consequently, the planning horizon will differ (Das, 1991). Accordingto Shipp et al. (2009) the differences in temporal focus, in otherwords how much attention is devoted to perceptions of past, presentand future, will affect attitudes and behaviors that in turn will be vis-ible in performance, sense-making and strategic choice. Lowe andHwang (2010) add some aspects when discussing Mead's (1934)conception of time and temporality. Experiences and actions are in-terwoven and involved with identity construction processes at indi-vidual and organizational levels, and past and present are bothreflected in the actions of the present.

In line with this, Araujo and Harrison (2002) point out structure andprocesses as developing over time in relation both to history and to anuncertain future: “A simultaneous attention to action and structurehighlights the importance of understanding path dependence not sim-ply from a retrospective perspective but also from an actor-centeredstandpoint, focusing on how actors react to an uncertain future in realtime” (Araujo and Harrison, 2002, p. 5). Orlikowski and Yates (2002)similarly suggest that time is experienced through a process of

Page 3: Capturing processes in longitudinal multiple case studies

Fig. 1. Main components of the theoretical framework.Modified from Wedin 2001, p. 168.

237L. Aaboen et al. / Industrial Marketing Management 41 (2012) 235–246

temporal structuring, meaning that the temporal conditions areconstructed and reconstructed by the actors involved.

According to Medlin (2004, p. 186) any consideration of time aspast, present and future immediately raises problems: “How long isthe present, or what are the boundaries of the present? When doesit begin and when does it end?” Concerning interaction and time,Medlin suggests that “…the present can be defined as the period be-tween two events, one past and the other anticipated, that changesthe ‘state’ of reality for at least one actor, and this period consists ofmore than two moments. This definition provides for a present thatis relative to the nature of the problem under study, with the presentmoving onwhen a significant event marks a change of future possibil-ities.” Furthermore, Medlin (2004) argues that a time perspective ofinteractions indicates the importance of events that change the futurepotential for managing relationships, and that such events can alsochange one's perspective on the past, present and future.

2.3. Network pictures and network drawings

In recent years there has been a great deal of interest among scholarsin the IMP tradition in exploring the uses and importance of network pic-tures (see e.g. Ford & Redwood, 2005; Henneberg, Mouzas, & Naudé,2006; Leek & Mason, 2009). Ford, Håkansson, Gadde, and Snehota(2003, p.176) present network pictures as “the views of the networkheld by participants in that network”. Geiger and Finch (2010) suggestthat the drawing of network picturesmay simplify complex business set-tings as well as shedding light on temporal and spatial dimensions ascompared with linear measures. In some studies, managers have beenasked to draw network pictures as a tool for visualizing the companyand where it fits into the network (cf. Leek & Mason, 2009).

According toHåkansson et al. (2009, p. 194): “Eachmanager has theirown subjective interpretation of theworld around them and of the inter-actions taking placewithin it, whether or not they are involved in them”.Network pictures are thus concerned with subjective images of theworld surrounding a company in a network wherein every actor hasher own pictures: “Such pictures reveal the companies’ or the indivi-dual's perception of what is happening around them, and guide in asses-sing usefulness of various actions and reactions they may undertake”(Holmen, Aune, & Pedersen, 2008, p. 2). However, it can be argued thatwhen managers are asked to give their subjective views of ‘how theirnetworks look’ they need some idea of what ‘components’ such a pictureshould contain, such as counterparts, relationships, etc. According toÖberg, Henneberg, and Mouzas (2007), network pictures have mainlybeen used to describe actors, activities, resources, network boundaries,network power, and network center/periphery.

In contrast to the notion of network pictures as they have recentlybeen defined, industrial network researchers have since the early daysof the research tradition been using drawings of networks to analyzeand illustrate particular network aspects of their empirical cases. Oneexample of such use is Lundgren (1995) who, based on his empiricalcase material, developed drawings of emerging networks to identifypatterns in this process. That is, unlike network pictures that capturethe subjective perceptions of the actors that are subject to study, net-work drawings made by the researcher have other purposes. In thispaper we will further the idea of network drawings as theoreticallybasedmethodological tools to capture patterns in processes that unfoldover time. Bynetwork drawingswe refer to illustrations of situations in-volving theoretically defined components put together by the research-er as a means to compile and analyze specific data.

3. An example of a multiple case study in progress

In this section, an ongoing study is used as an example of longitu-dinal multiple case studies. Both the study and the firms included arein an early stage of development. The study concerns strategizing ofnew technology-based firms (NTBFs), also referred to as “start-ups”.

The section starts with an introductory note on the framework usedin the study, followed by description of the data collection, three par-ticular cases, and some preliminary analysis of the results so far in thisongoing study.

3.1. A brief note on the theoretical framework used in the study

In a network perspective, strategic action concerns the efforts of afirm to influence its position in the network of which it is a part(Gadde, Huemer, & Håkansson, 2003). A network position, in turn, isdefined by the relationships a firm is involved in at a certain point intime (Johanson &Mattson, 1992). Harrison and Prenkert (2009) identi-fy adaptation within relationships as one type of network strategizing.The relationships of firmsmay be direct, i.e. with customers, or indirect,e.g. with a customer's other suppliers or a customer's customer. Thesubject(s) of business exchange, i.e. the product or offering, is a vitalpart of the relationships. These offerings may to differing extents be de-veloped in interaction with the customers and/or other parties.

Typically, start-ups begin their ‘life’ with limited resources andfew contacts and business relationships. Their ideas of a future net-work position are influenced by interaction within the relationshipsthe company develops (Aaboen, Dubois, & Lind, 2011). The interac-tion with the first counterparts of a start-up company is thus assumedto have a salient impact on how they form their ambitions in terms ofa future network position, which also relates to how they try to devel-op their resources.

To capture the strategizing processes of start-ups, the main com-ponents or building blocks of our framework build on the four-resource-entities model (Baraldi & Waluszewski, 2005; Håkansson &Waluszewski, 2002; Jahre, Gadde, Håkansson, Harrison, & Persson,2006; Wedin, 2001). The four resource-entities developed throughinteraction are products, production facilities, business units andbusiness relationships. This model enables a focus on the adaptationswithin relationships and how the business units and their relation-ships are developed through interaction with other resources.Hence, the focus on strategizing as a process is directed at analysisof how the business units, products, production facilities and relation-ships develop through interaction (see Fig. 1). And since business re-lationships are keys to strategizing in networks, the formation anddevelopment of relationships and the connections between these rela-tionships become of focal concern in the analysis.

In a longitudinal study of the NTBFs we want to be able to capturethe strategic actions taken (in interaction) in order to achieve thefirms' envisioned future position, the actions taken in the relationshipsthat did not become considered ‘strategic’ until later, and the interplaybetween the different actions as well as the development of the viewsof future positions. As Sørensen, Mattsson, and Sundbo (2010) arguein relation to innovation, development processes tend to be more com-plex, disorderly and interactive than is visible in presentations of con-secutive stages. In order to capture the strategizing patterns of NTBFswe argue that descriptions of, and reflections on, the past, present andfuture can be captured using network drawings. These network

Page 4: Capturing processes in longitudinal multiple case studies

238 L. Aaboen et al. / Industrial Marketing Management 41 (2012) 235–246

drawings are instrumental not onlywhen collecting data, but alsowhenanalyzing strategizing patterns of resource interaction, since the net-workdrawings contribute to capturing the development in some select-ed key dimensions over time. By studyingmultiple cases, we are able tocapture variety and similarity among strategizing patterns as interac-tion among resource entities.

3.2. Data collection and data analysis

The study began from an interest in understanding how start-upsdevelop their first customer relationships. While start-up companieshave received attention in terms of relationships with venture capi-talists, universities, incubators, etc., very little attention has beenpaid to their first customer relationships. Since getting paying cus-tomers is a key to the survival and future growth of any firm, we setout to make open-ended interviews with ten start-up companies, fo-cusing on the development of their initial relationships.

We identified the start-up companies by using snowball sampling.Previously we had been in contact with a few firms at early stages intheir development, and we were now able to re-connect with them.We also interacted with a member of the innovation system that pro-vides services such as coaching on the business plan and mediation ofcontacts to start-upfirms.We asked himaboutfirms that he had recentlyencountered which were in the early phase of forming customer rela-tionship and which had at least one paying customer, and we alsoasked him for other persons who might know such firms. The meetingresulted in contacts to half of the firms and a second contact personwho was part of a network of start-ups with a particular interest in sus-tainability. Through this person we received a couple of additional con-tacts. The firms were approached and ten start-up firms agreed toparticipate in the study. In the mail we sent to the start-up firms, weasked them to confirm that they were start-up firms with at least onepaying customer and also briefly mentioned the forming, managementand development of early customer relationships as the focus of the in-terview. The start-up firm then decided on who was the most suitableperson in that firm to answer the questions. In most of the start-upfirms this was the manager, as this person had been in the firm sincethe beginning and often had ‘been’ the firm more or less alone duringthe first years. Hence, in this first round of interviews one person percompany had an overview of the developments so far andwas thereforethe only one interviewed.

A semi-structured interview guide was designed with the intentionto get a full understanding of how relationships had been initiated anddeveloped, as well as how the relationships were connected with eachother. Our probing was thus focused on understanding what tookplace within the relationships during the different phases of develop-ment and why. During the interviews we also asked about the futureplans of the firms. The ten interviews were recorded on mini-disc andtranscribed in full by one of the authors. Due to the rich narrativeseach transcript became around 20 pages long. The 200 pages of tran-scripts, notes from the interviews and additional material such as bro-chures, material from web sites, industry press and newspaper articleswere compiled for the analysis. When analyzing the material wemade in-case analyses where we first wrote descriptions of the individ-ual start-up firms based on all material collected in relation to them.

In order to structure our data further we started to work on draw-ings to capture the changes in terms of the key components in ourframework. These drawings included both the events that the inter-viewees had emphasized and drawings of the networks they had striv-en to create, or become part of, at that time. Since we did the drawingsourselves, we had the possibility of using the same principles for allcases. This made it easier to compare patterns of development betweencases. If we had instead asked the interviewees to make the drawings(or network pictures) there is a possibility that wewould have receivedricher drawings including aspects of the development that we had notthought to ask about. Moreover, the interviewees may find it easier to

relate to pictures that they have drawn themselves at a later stage. Ifin-case analysis had been our focus, this could have been a better proce-dure, assuming that we would have been able to interpret the picturescorrectly.We noticed that the drawings of the actual networks at differ-ent stages in time tended to become rather complex after only a fewstages of development, while the networks that the firms aspired to be-comeparts of tended to be organized in a rather orderly fashion, and the‘aspired-to’ networks changed when the given NTBF realized that thestrategy it had used so far did not hold for diverse reasons.

Both variety and similarity could be observed among the cases in re-lation to the process of developing their first customer relationships. Inparallel, and in interaction with our impressions from the interviews,we developed the framework based on industrial network notions ofbusiness relationships. This eventually led to framing the processeswe had started to study in terms of ‘strategizing’. However, we alsogot the idea that this study could be developed into a longitudinal ap-proach, i.e. by capturing the processes during a longer time period. Bycollecting data regarding the past, present and future at several pointsin time,wewill be able to revealmore of the “dead-ends, chance events,and controversies” that Hoholm and Araujo (2011, p. 935) argue to beimportant to uncover in studies of development processes. The patternsthat we could identify in our drawings indicated some possible generalpatterns in the development of the companies' networking or strategiz-ing, but we also realized that we were only scratching the surface ofmore interesting long-term processes of strategizing. Since this is nottrivial from a methodological point of view, we decided that in orderto continue the study we first needed to develop the method for it. Inthe present paper we therefore focus on amethod thatmay enable cap-turing of processes in this and other studies of the dynamics in industri-al networks. We will use our study, and three of our cases, as anexample. The three included cases of strategizing processes in start-upfirms were chosen since they represent three different patterns.Hence, the other cases displayed similarities with these three cases inthe development patterns observed. By discussing the suggestedmeth-od based on these three cases, we illustrate that themethodological ap-proach can be useful to capture patterns in development processes.

3.3. The three cases of strategizing

In our examplewehave collected data at one point in time so far.Wehave interviewed one person in each start-up firmat this data collectionpoint since these interviewees have been able to maintain a completeoverview of the activities involved in the interactionwith the customersso far. At later data points there may be other persons at the firms whowill be equally important to interview and then we will do so.

In the paper three network drawings, of the past, present and futurerespectively, are presented in relation to each case. Each set of threenetwork drawings represents the structuring of data related to thekey constructs captured in one interview. This structuring and drawinghave taken place over a period of time where the authors have dis-cussed interpretations of the transcripts and how the network drawingsshould balance to be comprehensive enough without becoming toocomplex. When network drawings from additional data points havebeen collected, the sequences of network drawings can be combinedand the patterns reconsidered. The network drawings presented inthe following case descriptions can be viewed as tentative patternsbased on the first data point.

Since the presentations of our study are based on one data point sofar, there is a chance that the patterns are based on the stories thatthe firms prefer to remember rather than on what actually happened(Hoholm & Araujo, 2011). It is not the network drawings in themselvesthat are the main aim. Instead, drawings and narratives together illus-trate how business units, relationships, products and production facili-ties evolve in interaction. The set of network drawings will beinstrumental at the next data collection point in order to be reflected

Page 5: Capturing processes in longitudinal multiple case studies

239L. Aaboen et al. / Industrial Marketing Management 41 (2012) 235–246

upon as an illustration of how the interviewees viewed past, present,and future at the previous point in time.

Below, we present three of the cases of start-up companies andtheir development in relation to the framework in terms of past, pre-sent and future positions at one point in time. Finally, we discussideas about how to continue the study and thus how to design a lon-gitudinal multiple case approach in an extended effort to capture theevolution of these companies' strategizing processes.

3.3.1. EpsilonThe product of Epsilon is amethod of packaging food that is ready to

eat after heating in a microwave oven, including a particular patentedvalve that is part of the packaging system. The only part that is devel-oped and produced by Epsilon is the valve, but the customers need toinvest in an entire packaging line including equipment for filling theplastic troughswith food, adding plastic covering films and valves, cool-ing the food, etc. In order for the product to be developed, a facility atEpsilon that would produce the valves needed to be established. Fur-thermore, the facilities of the suppliers needed to be developed inorder to be able to produce components that were adapted to the spe-cific system of which the valve of Epsilon is a part. There are various de-grees of supplier collaboration to adapt the parts of the system to thevalve as well as to the particular customers' needs.

The initial focuswas tomake the valvefit togetherwith complemen-tary products offered by other suppliers, to ensure integration betweenthe different parts of the system. Initially, Epsilon developed relation-ships with some suppliers and venture capitalists, but did not have afunctioning product. Receiving venture capital resulted in collaborationwith one of two potential customers, and during the interaction withthe potential customers (both of them large food companies) the prod-uct (valve) was fully developed. The idea was that the customers in-volved in the development process would also buy the finishedproduct. However, that did not happen. Instead Epsilon was able to ini-tiate relationships with one of its first (paying) customers, C1, based onconnections of one of the potential customers. A key person, a chef, hadbeen engaged by one of the potential customers to develop suitable rec-ipes for the new technology. As a chef he could see the potential in thepackaging solution and in 2005 he started a new company, C1, to pro-duce and sell ready-made high-quality meals in a new chain of foodshops. The relationship with C2 was also established in 2005 and initi-ated as a result of a connection to one of the board members of Epsilon.C2 has relationships with national retailers in Sweden and with potatosuppliers in the Netherlands.

At the interview in May 2009, Epsilon was at the end of an eventfulperiod of its development. With two customers and a developed prod-uct, Epsilon maintained the relationship with the suppliers and theyapproached additional potential customers together. Epsilon also real-ized that the recipes, and thus the chefs, are keys to success with thecustomers. Therefore the idea of what potential customers to approachhad changed from ready-made food producers in general to high-quality ready-made food producers. The third customer, C3, is aNorwegian company. C3 is the first firm that Epsilon approached thathad actually become a customer. Bringing packaged food produced byC1 to the first meeting with C3 was important in terms of getting thediscussions going. The fourth customer, C4, is a large Finnish food firmthat contacted Epsilon. One of Finland's most famous chefs was in-volved in this business. The fifth customer, C5, is a Dutch firm originat-ing from a potato supplier, with which C2 has a relationship. Epsilonrealized that its customers were likely to know more than it did aboutthe product in large production settings and especially themanufactur-ing process. Following this realization, Epsilon mediated contacts be-tween its customers. The customers were interested in exchanginginformation about how to set up the food manufacturing processessince, from the customer perspective, the recipes are the secret partsof their respective ready-made foods. Relationships with two Americanfirms were in the process of being built up, and in this process contacts

were being mediated so that representatives of the American firmscould visit the production sites of the Norwegian and Finnishcustomers.

Epsilon felt that in addition to having a network of customers thatare connected to each other, it would be beneficial in the future ifthere were one key customer in each country. The system sold by Ep-silon develops the production facilities of its customers. Epsilon real-ized that it was approaching a limit in terms of how many customersit could handle, given the current size of its business unit. It thereforeplanned to increase its mediation of contacts between customers. Byletting the customers develop each other's facilities further by solvingproblems with the system and learning how to operate the systembetter, Epsilon was able to add to the customers' facilities without de-veloping its business unit further. Epsilon had also decided to try tostrengthen its relationships with its suppliers, by using the same sup-pliers for all customers. In the future, Epsilon hopes that this will ren-der the suppliers more willing, in turn, to promote Epsilon ininteraction with their customers. Fig. 2 illustrates the developmentof the business network of Epsilon, with important stages of networkdevelopment to the left, and Epsilon's idea of strategic action in theparticular network situation to the right.

3.3.2. DeltaFounded in 2006, Delta bases its work on research in organic catal-

ysis at Stockholm University and the Swedish University ofAgricultural Sciences. This knowledge is used to attach molecules tocellulose materials to create certain features, relating, for example,to water resistance and mechanical properties. The original connec-tion with these universities and the continued relationship withthem is central to Delta, since it gives access to certain equipmentand awareness of ongoing related research projects.

Delta started out thinking in terms of developing a customer port-folio strategy with customers in different parts of the value chain, orstages of refinement, of cellulose products. It was not clear where inthat chain Delta's products could best be used from technical andbusiness points of view. During autumn 2006, the CEO of the compa-ny made a number of cold calls to potentially interesting customers.Three of these yielded agreements about starting pre-studies in thespring 2007. The three are referred to as C1, C2 and C3. Later, a fourthcustomer, C4, also signed a pre-study agreement based on these ini-tial contacts.

In May 2009, at the time of the interview, Delta had seven cus-tomers in different stages of a three-stage business model: First,pre-study/lab tests, where the basic material features are developedin interaction with the customer. Second, pilot tests, in which the fea-tures are further developed in interaction with the customer at a pilotplant. Third, full-scale industrial testing, where the material featuresare developed in the customer's own production plant. In total, tenmaterials were tested. Two did not work out and these projectswere ended. Delta bore the costs of the lab tests, while the customerpaid for the pilot and full-scale tests. The substances needed to per-form the tests are standardized and the quantities were small, andDelta bought them from suppliers without much interaction.

C2 is a Swedish paper manufacturer, mainly interested in usingthe new technology for its water resistance features. At the time ofthe interview, Delta and C2 were jointly working with a pilot studyperformed using a small paper machine, in order to conclude wherein the paper manufacturing process it would be appropriate to addthe fluid for organic catalysis. C2 already worked in a structured man-ner and had, for instance, a pilot set-up that was very useful for run-ning Delta's tests. Delta considered C2 skilled and knowledgeable, andDelta could learn about the process of paper manufacturing from C2and then use this knowledge in relation to other customers, for exam-ple C4. The relationship with C2 was developing to involve more peo-ple from different C2 divisions and also to take the customers intoconsideration.

Page 6: Capturing processes in longitudinal multiple case studies

Fig. 2. Network development and strategizing of Epsilon.

240 L. Aaboen et al. / Industrial Marketing Management 41 (2012) 235–246

C4 is a French company that develops and manufactures plaster-boards. A plasterboard has a paper sheet on each side of the plasterand C4 was interested in developing certain features of the paper.One of the mills of C4 is located in central Sweden, and at this millDelta performed its first full-scale tests, mainly based on what ithad learned in interaction with C2. At C4 the project went directlyfrom the pre-study lab tests to full-scale testing, partly because thiscustomer did not have a pilot machine to use. In relation to a Spanishfabric manufacturer, C5, it was also possible to shorten the pilot testtime based on Delta's understanding of the paper manufacturing pro-cess learned from C2 and C4.

By getting to know the customers' business contexts, Delta nowsees a future potential in growing by developing several productsfor each one of its customers. Therefore, instead of aiming primarilyat developing more customer relationships, Delta stated at the time

of the interview that it intended to grow further within each relation-ship. On the basis of the realization that all products developed arehighly specific to the particular end-product and production facility,Delta stated that it planned to use its new knowledge about productfeatures in a more systematic way across all the customer relation-ships. In short, growing and developing individual customer relation-ships, and utilizing knowledge about product features acrosscustomer relationships without contacts between the customers,will be the focus for future business development. In other words,during the process of developing a product in interaction with its cus-tomers, the facilities of the customer are developed using knowledgefrom other customers' facilities in combination with the knowledge ofDelta's business unit and the customer's business unit.

See Fig. 3 for an illustration of the network development of Deltaand the focus of strategizing in different phases. It should be noted

Page 7: Capturing processes in longitudinal multiple case studies

Fig. 3. Network development and strategizing of Delta.

241L. Aaboen et al. / Industrial Marketing Management 41 (2012) 235–246

that the current product of Delta is undergoing tests, illustrated witha dotted line around a focal triangle, and it is still quite a long wayfrommass production and use of the product, i.e. the organic catalysissubstance. When they start large-scale production, the suppliers willbecome more central than they are today.

3.3.3. BetaBeta is a company that develops software to detect child pornog-

raphy on the Internet. To date, the software has been developedinto four different products. The first product, P1, helps the policeforce investigate crimes by identifying users of child pornographywebsites. The second product, P2, is intended for use by companies,

to reveal employees surfing on child pornography pages. P3 isdesigned for Internet operators that want to block such contents forall its users. P4, the latest addition, is a filter to be used by parentswho want to protect their children when using the Internet.

The initial idea was to develop the four products for four differenttypes of customers. Beta started with the two products, P1 and P2. P1was developed together with the Swedish police force, who became a(non-paying) customer/user (C1). P1 helped the police force investi-gate this type of crimes and thereby improved its own facilities. In ad-dition to using P1, the police also provided information about whattype of content was illegal, and this information was used to developthe basic software. In other words, the police instead of paying

Page 8: Capturing processes in longitudinal multiple case studies

242 L. Aaboen et al. / Industrial Marketing Management 41 (2012) 235–246

improved the facility of Beta. Beta has also drawn the conclusion thatthe Swedish police force (C1) could be a door opener towards policeforces in other countries. Another actor with whom Beta initiated arelationship on a more general level is A1, a charity focusing on chil-dren, and A1 has helped mediate contacts to potential customers.

A1 and the relationship with the police force provide legitimacy forBeta and its products. Beta was, at the time of the interview in March2009, learning that this is important in order to motivate potentialcustomers. The first major customer of P2 is C2. Through contacts onthe board of A1, Beta was able to contact the CEO of C2 directly, afterits initial attempts to approach the firm failed. The CEO wanted tohave P2 installed and made her decision within a week. Previously, P2had only been tested on ten computers at a time but at C2 the numberof computers was in the range of 10,000, which resulted in the occur-rence of a number of unforeseen bugs. However, these problems weresolved through hard work at Beta in collaboration with C2 and the inci-dent actually turned out to improve communication, and to build afoundation for a closer relationship with C2. Furthermore, Beta learnedthat the productworks in large networks and this led to an ‘all problemsare solvable’ attitude. C2 has become a valuable reference customersince it is involved in the development of the security departments atmany companies and is therefore well-known. The manual C2 wroteabout how to handle alerts from the program was edited and couldthen also be offered to Beta's other customers. Since it was the CEOwho decided that the software should be bought by C2, Beta hopedthat all customers for P2 would be accessible through their CEOs. Thisassumption changed, however, after a couple of failed attempts. Thenext customer of P2, C3, was instead approached through the identifica-tion and persuasion of members of both human resources and IT, andthis model for approaching new customers then remained in use.

P3 was developed in collaboration with another firm, a supplier (S)with Beta as its only customer. P3 is sold by distributors (D), and theirpersonal networks and understanding of their markets is of great im-portance. P3 is often sold abroad to large Internet providers or entirecountries. Through feedback from early distributors, Beta has beenable to make necessary adjustments, in addition to translations, oftheir marketing material. New Zealand and Monaco were the first cus-tomers to buy P3. Both New Zealand and Monaco were given favorablecontracts since Beta planned to use them as reference customers. Beta isdependent on distributors recognizing windows of opportunity, e.g.when child pornography is making headlines. Beta had introduced itsfourth product, P4, at the time of the interviews, andwas about to mar-ket this to existing customers as a complement to the products they hadalready purchased.

For future customer interaction, Beta will consider P4 as a comple-ment to the other products since P4 helps the user, rather than servinga preventive or monitoring function like the other products. P4 willtherefore hopefully facilitate the implementation of P2 in other organi-zations. Having initially considered each one of its products separatelyin relation to different potential customers, Beta intends, in the future,to increase its focus on expanding within current customer relation-ships. C2, for example, is a large provider of network access and tele-communication services, and might therefore be interested in being adistributor for P3 and providing P4 to both employees and customers.So far, the product of Beta has not directly improved the facilities ofthe customers (except for the police) but has instead added differenttypes of value that may affect the facilities indirectly. However, oncethe product of Beta is integrated into the product offering by its cus-tomers to their customers, this may be seen as adding to the facilitiesof Beta's customers.

In a long-term perspective, Beta has formulated a vision of reach-ing all computers in the world with its products, and it plans to fulfillthis vision through the customer network of one big actor rather thantrying to approach each of the potential users themselves. However,fulfilling this vision will move Beta towards a more peripheral posi-tion in the network, as its software will become a small part of an

extensive software package. Fig. 4 illustrates the development of thebusiness network of Beta to the left and the strategic changes to theright.

3.3.4. Preliminary analysis of strategizing patterns for Epsilon, Delta andBeta

Althoughwehave, so far, only collected data at one point in time, thecases illustrate how the strategizing focus has changed over time as thestart-up companies' networks have developed. For Epsilon, the focusmoved from developing the product together with two ‘potential’ cus-tomers and suppliers/partners to a focus on what customer relation-ships to develop and how to interact with customers. Havingdeveloped some customer relationships, it next moved to a focus onconnecting the customer relationships bymediating direct contacts be-tween them in order to help them support each other to solve variousproblems, as well as to developing the relationships with the suppliersto make them more active in finding additional customers. In this wayEpsilon handled the limitations of its possibilities to further developits business unit and facilities. Its ideas on how to link relationships toone another (and when not to do so) were taken further to the idea ofdeveloping key customer relationships in different countries.

The strategizing of Delta shows a different pattern. At the time of theinterviews, Delta was performing tests with its customers. While thetests were still rather distant in time from ‘real’ products, they werebeing carried out in close interaction with the customers, with the am-bition of developing new products. Hence the strategizing has empha-sized developing customer relationships. Initially, Delta had quite arandom approach to potential customers, but over time it learned thatthere was a need to get into the production processes of the customersand that this required intense and complex interaction. The scope of therelationships expanded to include connections to other actors such asthe customers' customers, to gain an understanding of the require-ments of the customers. Instead of trying to find more customers,Delta began to put its efforts into including more products (first tests)in its current customer relationships. This reduced the costs of learningabout the customers' facilities. Regarding connections between rela-tionships, Delta saw that what it learned about the product features inone customer relationship was potentially useful in other customer re-lationships. In contrast to Epsilon, Delta kept its customer relationshipsseparate and thus built on indirect connections between customerrelationships.

The strategizing of Beta was initially concerned with four differentproducts and how to approach potential customers with them. In thisprocess, it learned that it needed to ensure that the technology of theproducts functioned, and also to convince the customers that they hada problem that could be solved. This led to a focus on developing its le-gitimacy by using the relationships with the police force, the charity,and one of its first customers as reference partners. Another insight oc-curred when it was about to launch its fourth product, and this led to anew approach to customers — offering several products to its currentcustomers, similar to the development of Delta's strategizing. Hence,the strategizing of Beta initially had amain focus on the product dimen-sion but developed into a focus on relationships. Furthermore, the prod-uct of Beta is targeted towards the development of business units ratherthan facilities, and potential customers therefore have to be approachedin a different way. Not only pure business relationships, but also otherrelationships and the connections among relationships, have been keyissues in developing new ones.

While the common aspect of the case companies' strategizing is theintention to grow the business units, the three cases display differentpatterns of interaction between the four resource entities that are thekey building blocks of the framework. The importance of creating director indirect connections between relationships can be seen in the inter-action processes in all the three cases. However, the connections con-cern different kinds of parties and have different functions for thestrategizing of the companies.

Page 9: Capturing processes in longitudinal multiple case studies

Fig. 4. Network development and strategizing of Beta.

243L. Aaboen et al. / Industrial Marketing Management 41 (2012) 235–246

4. Discussion

4.1. Findings

In order to observe a process the central focus tends to be on pro-gressions of events, stages and activities that the focal firm undergoesas it develops over time. The events can be viewed as paths that proceedin parallel, or diverging, or converging in a way thatmakes them cumu-lative, conjunctive and iterative. Furthermore, the events may be addi-tions, substitutions or modifications of the previous events (Van deVen, 1992). In the case study, the main components of strategizing ina network (the empirical phenomenon of focal interest) are identifiedin the theoretical framework. Consequently, the events are defined bychanges among the components, i.e. changes in the products, facilities,

relationships, units, the product-relationship link and connections be-tween the relationships.

In the case study we have illustrated how the theoretical frameworkof selected components before the data is collected, as presented in Fig. 1,can be used to capture key elements in the focal process as changes in re-lation to the components in the framework. The key concepts translatedinto components of the framework can be captured in network drawingsin such away that these changes are clearly seenwhen comparing draw-ings fromdifferent points in time. Changes in relation to strategizing thusoccur when the drawings change in one or several ways, e.g. when rela-tionships are added, when connections between relationships change,when products are added to a relationship, etc.

The strategizing process that we study in these early phases offirm development differs on an important point from traditional

Page 10: Capturing processes in longitudinal multiple case studies

244 L. Aaboen et al. / Industrial Marketing Management 41 (2012) 235–246

studies of strategizing. The latter assume that the development is amovement towards a desired end state or goal (Van de Ven, 1992).However, using the identified main components to construct the de-sired end state, we observe that it also undergoes changes, which inturn cause the strategizing to change in a new direction. Thesechanges of goals divided the process into the stages of strategizingthat we observed, e.g. when the firm realizes that the present strategyand goal are not working and decides to work in a different direction.In other words, we suggest that the components and events may bedefined a priori while the stages and goals are realized during thestudy.

4.2. The methodological approach

In the approach we suggest, we include two perspectives of time.The absolute or objective time is complemented with the relativisticor subjective time. The absolute or objective time is related to the actualpassage of time (Shipp et al., 2009). In order to capture the processes in-cluding all events, we proposemaking interviews at certain intervals inorder to gather and analyze narratives in a systematic way over time.Subjective time, in contrast, is related to personality, social culturesand related norms (Fried & Slowik, 2004). For example quick, slowand late mean different things to different people, and while some peo-ple focus on planning for the future, others value traditions and the past.We propose that interviews cover past, present and future but do notspecify the time span, or which of the three parts should be given themost focus, knowing that this will differ between interviews. The narra-tives are then told from the time perspective of the interviewee.

By collecting the narratives based on a subjective time perspectiveat certain intervals in absolute time, the same event will, based on aseries of interviews over time, be narrated as an event in the future,present and past. The different narratives will then aggregate intoan understanding of the events and their role and importance in theprocess. This aggregation is illustrated in Fig. 5 below.

The narrative collected at the first study point is illustrated as Past(1) Present (1) Future (1). At the second study point another narra-tive will be collected. As illustrated in Fig. 5, the events that were per-ceived as the present at the first study point will be in the past at thesecond study point. The series of events that take place around thetime of the second study point will then, after these three studypoints, have been narrated three times but from different perspec-tives — as the future, present and past. However, as explained earlier,the narratives are told from the time perspective of the interviewee,and the equally sized boxes in the illustration are therefore just

Fig. 5. Illustration of a ‘rolling’ present and a ‘rolling’ future.

illustrations. The narratives presented as illustrations in the presentpaper were collected at study point 1 in our study.

The reflections made at the data collection point play an importantpart in the in-case analysis. Hence the network drawing may be usedas a tool to enhance reflections over time in order to strengthen in-case analyses and to reveal longitudinal patterns. The longitudinal pat-terns can then be contrasted between the cases.

The narratives include more than can be captured in the drawings.However, the network drawings give an overview of the events con-sidered essential to development in the past, present and future at thepoint in time of the interview. In the analysis, the drawings can thusbe an embedded part of the narratives. In our approach we suggestthat the drawings from previous study points should be brought tothe interviews. The drawings serve as a tool for remembering the sit-uation at the previous study point, and thereby enable reflectionupon how the situation has developed, as well as reflection on previ-ous situations in light of the present. As illustrated in Table 1, the nar-ratives collected at study point 3 in Fig. 5 above may then actuallyconsist of not only Past(3) Present(3) Future(3), but also reflectionsupon the narratives collected at the previous study points in light ofthe present situation, as well as reflection upon in what ways the pre-sent situation is a function of the previous situations. Furthermore,they enable reflections upon how the current future goals are a func-tion of the future goals of the past and whether the past expectationshave materialized. Hence, the network drawings enable us to relateand contrast the descriptions of past, present and future over time,and thus reinterpretations and reconsiderations of prior situationscan be integrated into the narrative descriptions of the changes stud-ied. In order for the data collected at a series of study points not to be-come infinitely abundant and complex, we suggest that it is not onlythe future and present that are ‘rolling’ as illustrated in Fig. 5, but alsothe focus of attention. At each study point the narratives revolvearound the present study point and the nearest study points beforeand after it. The further away a study point is from the area of focus,the more it is treated as marginal and as a part of the ‘past’ or ‘future’.

In commenting on the industrial networks approach and contempo-rary notions about ‘network pictures’, Lowe and Hwang (2010, p.13)refer to these as using a ‘photographic’ metaphor, since “pictures andtime are relatively still and sequential”. This relates to the emphasison structure versus process and to the notion of path dependence. In re-sponse,we suggest that a series of network drawings and narratives de-scribing their evolution over time could turn the drawings into more ofa ‘movie’ than a set of static structural descriptions.

Over time, cross-case analysis with our suggested method will per-mit identification and analysis of patterns in the companies' develop-ment processes. The firms in our study have at the first data pointdisplayed individual patterns of development based on how the

Table 1The collection of information at study point 3 by using reflection upon the networkdrawings from previous study points as an active part of the interviews.

Narrative collected Reflections captured

The narrative of study point 3. The current view of past, present andfuture.

The narrative of study point 2 and thenarrative of study point 1 using networkdrawings of study points 1 and 2.

Reflection upon the situations in theprevious study point and the studypoint before that in light of the presentsituation.

The narrative of study point 3 usingnetwork drawings of study points 1and 2.

Reflection upon the present situationas a function of the previous and/orseveral of the previous situations.

Comparison of what is presently seen inthe future with previous views of thefuture.

Reflection upon how the goals havechanged.

Comparison of how the same events areperceived when narrated at differentpoints in time.

Reflection upon whether expectationshave materialized.

Page 11: Capturing processes in longitudinal multiple case studies

245L. Aaboen et al. / Industrial Marketing Management 41 (2012) 235–246

product ideas relate to existing structures of business units, business re-lationships, production facilities and other products among their cus-tomers (Aaboen et al., 2011). The firms not only needed to relate tothe resource structures of their customers in different ways, they alsoneeded to learn how to interact with their first customers in order tobe able to ‘fit in’ or ‘add to’ the customers' resource structures(Aaboen et al., 2011). Hence, the kinds of patterns that we search fordo not focus on pre-determined relations between variables. Thereforethe set-up of cross-case analysis has to be adjusted accordingly. In linewith to Eisenhardt (1989) we suggest that similarities and differencesacross cases should be looked for. However, we emphasize that thecomparison among cases need to focus on patterns in the processes oflearning and interaction rather than on relationships between con-structs. The comparisons thatwe aim for relate to series of events ratherthan to snapshots ofmoments in the development. The patterns thatweare looking for in the study of strategising in start-up companies will bebased on how the firms are learning from resource interaction and thuson the changing perceptions of the development of relationships, busi-ness units, products and production facilities and, especially, how theseare connected. At this pointwe can, as a similar pattern across the cases,see a shifting focus from interaction aiming at developing the products,based on ambitions to sell these products to (many) customers, to inter-action aiming at developing the contents (including additional prod-ucts) of the individual customer relationships, and of benefitting fromconnecting customer relationships, directly or indirectly. In later stages,we expect to see other similar (and different) patterns across the cases.

5. Concluding discussion

In this paper we have used an ongoing multiple case study of howstart-up companies strategize in order to elaborate on how to captureprocesses in industrial networks. The suggested approach entails par-ticular solutions to the problems of case research on industrial net-works pointed out by Halinen and Törnroos (2005). First, ‘setting of(network) boundaries’ can be done in many different ways. The ap-proach described here permits continuous study of how networks de-velop, since the network boundary is identified in an inductivemanner. Second, the network complexity is ‘reduced’ by using a the-oretical focus including some specific aspects of network complexity.Using this deductive approach to the complexity dimension, somekey aspects that develop over time, such as those of concern for strat-egizing processes in our example, can be studied systematically.Third, as far as the problem of time is concerned, we have focusedon particular events with regard to perceived changes in the net-works of the companies and of their ideas of future network positions.In the study of strategizing, aiming for continued study of how thestart-up companies and their networks develop, an additional timeboundary concern will be how to ‘condition’ any conclusions aboutthe strategizing in relation to the arbitrary points in time when weupdate and analyze the cases. Fourth, concerning case comparisonsin the suggested continued efforts, the variety among strategizingpatterns will be addressed. A focus on patterns instead of changesamong the components is advocated as a way of balancing the bene-fits from the variety of cases against maintaining an in-case logic.

We therefore argue that the four boundary challenges can beaddressed either by inductive (empirical) inquiry or by relying ontheoretical frameworks that set limits for the boundaries. Decidingwhich boundaries to address, and how, is a key issue for any studyof network phenomena. As a methodological tool, network drawingsenable description and analysis of particular changes over time. In ad-dition, network drawings may support reflection on how, when andwhy particular action has been taken and on change in a longer-term perspective.

There are some other suggestions in the literature of how to com-bine longitudinal studies with multiple case studies. Leonard-Barton(1990) favors a dual method for case studies, with the synergistic

combining of a longitudinal single case study with retrospective(multiple case) studies. Hereby, “the retrospective studies offer theopportunity to identify patterns indicative of dynamic processes andthe longitudinal study provides a close-up view of those patterns asthey evolve over time” (Leonard-Barton, 1990, p. 248). She holdsthat the past and the present can be captured in combination withstudies of one and several cases. We find that aspired future statesare also essential and possible to capture, as these too guide the ac-tions in the present.

In line with Leonard-Barton (1990, p. 260) we agree that “nomethodology is perfect” and follow her example by ending thispaper with a few of the most important problems or possible sacri-fices that we see as associated with the suggested approach.

First, the industrial network approach that is the foundation forthis research emphasizes the unique character of every company.This relates back to the assumption of resource heterogeneity(Penrose, 1959), and particularly the notion that the value of everyresource depends on which other resources it is connected with. Ac-knowledging this notion implies that cross-case comparisons cannotbe made in an atomistic fashion wherein companies, or their re-sources, or the processes they are going through, are compared acrossa range of general features. For the same reason, the basic theoreticalassumptions on interdependence and interaction do not encourage asearch for extrapolating processes into, for example, generalized‘phases’. This naturally sets limits for what kinds of comparisons canbe made across cases.

Second, studies of relationships and networks place an emphasison the dyads and how dyads are connected. Using the individualfirm as the unit of analysis, as in the cases discussed in this paper, istherefore problematic. Moreover, using only one firm's perspectiveon the development of a relationship, and thus also on that relation-ship's connections to other relationships, is also problematic. We donot address the practical problems inherent in the enormity of tryingto capture the perspectives of all counterparts to the firms involvedand over time. This limitation or sacrifice, however, reduces the ex-planatory power in the approach we suggest, since capturing ‘theother side’ of the relationships would be necessary in order to devel-op an understanding ofwhy things happen in the interaction betweenthe focal firm and its counterparts.

Third, one of the questions often asked by single case researchersis: What is this case a case of? Ragin (2000) argues that this questionshould be asked repeatedly during a case study and that this is the es-sence of the ‘casing’ process itself. Whenmakingmultiple case studiesthe idea of what the cases are cases of cannot be adjusted or changedas the study progresses, since ‘casing’ each and every one of the casesin this way could have the result that the cases appear increasinglydistinct and become cases of different phenomena. Hence, the choiceof a set of cases, and what these should be cases of, has to be made inthe beginning of the study and cannot emerge as an outcome of theresearch process as it can for single case designs (Dubois & Araujo,2007). The choice of using multiple cases to capture a variety of pat-terns of particular (ex ante defined) processes thus goes hand in handwith the need to use a more deductive approach, which does not per-mit inclusion of all interesting complexities that emerge in the pro-cess of the study. Instead, the study must remain with the initiallydeveloped framework.

References

Aaboen, L., Dubois, A., & Lind, F. (2011). Start-ups starting up— Firms looking for a net-work. IMP Journal, 5(1), 42–58.

Araujo, L., & Harrison, D. (2002). Path dependence, agency and technological evolution.Technology Analysis & Strategic Management, 14(1), 5–19.

Baraldi, E. (2003). When Information Technology Faces Resource Interaction. Doctoraldissertation, No. 105, Uppsala University, Department of Business Studies.

Baraldi, E., & Waluszewski, A. (2005). Information technology at IKEA: An “open sesa-me” solution or just another type of facility? Journal of Business Research, 58(9),1251–1260.

Page 12: Capturing processes in longitudinal multiple case studies

246 L. Aaboen et al. / Industrial Marketing Management 41 (2012) 235–246

Bluedorn, A. C., & Denhardt, R. B. (1988). Time and organizations. Journal of Manage-ment, 14(2), 299–319.

Butler, R. (1995). Time in organizations: Its experience, explanations and effects. Orga-nization Studies, 16(6), 925–950.

Das, T. K. (1991). Time: The hidden dimension in strategic planning. Long Range Plan-ning, 24(3), 49–57.

Dubois, A., & Araujo, L. (2007). Case research in purchasing and supply management:Opportunities and challenges. Journal of Purchasing & Supply Management, 13(3),170–181.

Dubois, A., & Gadde, L. -E. (2002). Systematic combining: An abductive approach tocase research. Journal of Business Research, 55(7), 553–560.

Dubois, A., & Gibbert, M. (2010). From complexity to transparency: Managing the in-terplay between theory, method and empirical phenomena in IMM case studies.Industrial Marketing Management, 39(1), 129–136.

Dyer, W. G., & Wilkins, A. L. (1991). Better stories, not better constructs, to generatebetter theory: A rejoinder to Eisenhardt. The Academy of Management Review,16(3), 613–619.

Eisenhardt, K. M. (1989). Building theories from case study research. The Academy ofManagement Review, 14(4), 532–550.

Ford, D., Håkansson, H., Gadde, L. -E., & Snehota, I. (2003). Managing business relation-ships. Chichester: John Wiley & Sons Ltd..

Ford, D., & Redwood, M. (2005). Making sense of network dynamics through networkpictures: A longitudinal case study. Industrial Marketing Management, 34(7),648–657.

Fried, Y., & Slowik, L. H. (2004). Enriching goal-setting theory with time: An integratedapproach. The Academy of Management Review, 29(3), 404–422.

Gadde, L. -E., Huemer, L., & Håkansson, H. (2003). Strategizing in industrial networks.Industrial Marketing Management, 32(5), 357–364.

Geiger, S., & Finch, J. (2010). Networks of mind and networks of organizations: Themap metaphor in business network research. Industrial Marketing Management,39(3), 381–389.

Gressetvold, E. (2004). Product Development –Effects on a Company's Network ofRelationships. Doctoral dissertation, Norwegian University of Science and Technology,Department of Industrial Economics and Technology Management.

Håkansson, H. (1982). International marketing and purchasing of industrial goods. Chiches-ter: John Wiley & Sons Ltd..

Håkansson, H., Ford, D., Gadde, L. -E., Snehota, I., & Waluszewski, A. (2009). Business innetworks. Chichester: John Wiley & Sons Ltd..

Håkansson, H., & Waluszewski, A. (2002). Managing technological development. Lon-don: Routledge.

Halinen, A., & Törnroos, J.-Å. (2005). Using case methods in the study of contemporarybusiness networks. Journal of Business Research, 58(9), 1285–1297.

Harrison, D., & Prenkert, F. (2009). Network strategizing trajectories within a plannedstrategizing process. Industrial Marketing Management, 38(6), 662–670.

Henneberg, S. C., Mouzas, S., & Naudé, P. (2006). Network pictures: Concepts and rep-resentations. European Journal of Marketing, 40(3/4), 408–429.

Hjelmgren, D. (2005). Exploring the Interplay between Standard Products and CustomerSpecific Solutions. Doctoral dissertation, Chalmers University of Technology, De-partment of Technology Management and Economics.

Hoholm, T., & Araujo, L. (2011). Studying innovation processes in real-time: The prom-ises and challenges of ethnography. Industrial Marketing Management, 40(6),933–939.

Holmen, E. (2001). Notes on a Conceptualisation of Resource-related Embeddedness ofInterorganisational Product Development. Doctoral dissertation, Southern DenmarkBusiness School.

Holmen, H., Aune, T., & Pedersen, A. -C. (2008). Network pictures network potentials insupply management. The annual 24th IMP Conference, Uppsala, Sweden.

Hulthén, K. (2002). Variety in Distribution Networks: A Transvection Analysis. Doctoraldissertation, Chalmers University of Technology, Department of Technology Man-agement and Economics.

Jahre,M., Gadde, L. -E., Håkansson, H., Harrison, D., & Persson, G. (2006). Resourcing in busi-ness logistics: The art of systematic combining. Malmö: Liber & Copenhagen BusinessSchool Press.

Johanson, J., & Mattson, L. -G. (1992). Network positions and strategic action — Ananalytical framework. In B. Axelsson, & G. Easton (Eds.), Industrial networks: Anew view of reality (pp. 205–217). London: Routledge.

Lee, H., & Liebenau, J. (1999). Time in organizational studies: Towards a new researchdirection. Organization Studies, 20(6), 1035–1058.

Leek, S., & Mason, K. (2009). Network pictures: Building an holistic representation of adyadic business-to-business relationship. Industrial Marketing Management, 38(6),599–607.

Leonard-Barton, D. (1990). A dual methodology for case studies: Synergistic use of alongitudinal single site with replicated multiple sites. Organization Science, 1(3),248–266.

Lind, F. (2006). Resource Combining across Inter-organisational Project Boundaries. Doc-toral dissertation, Chalmers University of Technology, Department of TechnologyManagement and Economics.

Lowe, S., & Hwang, K. S. (2010). Time, process and discourse in business network re-search. Paper presented at the 26th Annual IMP Conference, Budapest, Hungary.

Lundgren, A. (1995). Technological innovation and network evolution. London:Routledge.

Mead, G. H. (1934). Mind, self and society. Chicago: University of Chicago Press.Medlin, C. J. (2004). Interaction in business relationships: A time perspective. Industrial

Marketing Management, 33(3), 185–193.Öberg, C., Henneberg, S., & Mouzas, S. (2007). Changing network pictures: Evidence

from mergers and acquisitions. Industrial Marketing Management, 36(7), 926–940.Orlikowski, W. J., & Yates, J. (2002). It's about time: Temporal structuring in organiza-

tions. Organization Science, 13(6), 684–700.Penrose, E. T. (1959). The theory of the growth of the firm. Oxford: Basil Blackwell.Ragin, C. C. (2000). Comparative methods. In W. Outhwaite, & S. P. Turner (Eds.), The

SAGE handbook of social science methodology (pp. 67–81). London: SAGEPublications.

Shipp, A. J., Edwards, J. R., & Lambert, L. S. (2009). Conceptualization and measurementof temporal focus: The subjective experience of the past, present, and future. Orga-nizational Behavior and Human Decision Processes, 110(1), 1–22.

Sørensen, F., Mattsson, J., & Sundbo, J. (2010). Experimental methods in innovation re-search. Research Policy, 39(3), 313–322.

Van de Ven, A. H. (1992). Suggestions for studying strategy process: A research note.Strategic Management Journal, 13, 169–188 Special Issue.

Wedin, T. (2001). Networks and demand: The use of electricity in an industrial process.Doctoral dissertation, No. 83, Uppsala University, Department of Business Studies.

Yin, R. K. (2003). Case study research: Designs and methods (3rd ed.). Thousand Oaks:Sage Publications.

LiEcN

se Aaboen is a PhD and researcher at the Department of Technology Management andonomics, Chalmers University of Technology. Her research interests include incubators,TBFs, early customer relationships, strategy and commercialization of technologybased

ideas. She has published in Technovation, International Journal ofManagement and Enter-prise Development, IMP Journal, International Journal of Technology Transfer and Com-mercialisation and International Journal of Business Innovation and Research.

Anna Dubois is a professor in the Department of Technology Management and Eco-nomics at Chalmers University of Technology. Her main research interests are industri-al marketing and purchasing, supply network structures and dynamics, organizingacross boundaries and case methods. She has published in Industrial Marketing Man-agement, Journal of Purchasing and Supply Management, Journal of Business Research,Supply Chain Management: An International Journal and Journal of ManagementStudies.

Frida Lind is an Assistant Professor in the Department of Technology Management andEconomics, at Chalmers University of Technology, Gothenburg, Sweden. Her main re-search interests are in the areas of industrial networks, resource development, researchcenters and start-up companies and their initial customers. She has published papers inJournal of Business Research, Scandinavian Journal ofManagement, IMP Journal and Tech-nology Analysis & Strategic Management.