key: cord-1045545-rsnn6q78
authors: Bygballe, Lena E.
title: Toward a conceptualization of supplier-switching processes in business relationships
date: 2017-01-31
journal: Journal of Purchasing and Supply Management
DOI: 10.1016/j.pursup.2016.04.007
sha: 5ee6f0a9583884368770842a17b0aef55e2ab391
doc_id: 1045545
cord_uid: rsnn6q78

Abstract This paper employs a practice perspective to study and conceptualize supplier-switching processes in business relationships. It is based on case study research of a company in the international marine industry, which intended to switch one of its key suppliers for a new one as a result of cost-cutting strategies and dissatisfaction with the old supplier. The case study describes the process, which ended with the old supplier being only partially switched. The findings show how (partial) switching from one supplier to another happened via three key sub-processes and associated practices, which the involved actors in the case drew upon in the switching process: (1) initiation – a process enabled by legitimizing and search practices; (2) substitution – a process enabled by transfer, translation, and transformation practices; and (3a) stabilization – a process enabled by institutionalizing practices; and (3b) restoration – a process whereby the old supplier is retained as a result of new conditions, this retention being enabled by certain repair practices. By identifying the processes and practices that enable switching to happen, the findings offer an initial conceptualization of supplier-switching processes, which comprises an important and heretofore underexplored aspect of supplier switching. The research highlights the importance of recognizing how relationships embedded in interorganizational routines are produced and reproduced in the switching process through the actions and interactions of the actors involved.

The formation and development processes of business relationships have received considerable interest in the supply management and marketing literature over the years (Håkansson, 1982; Dwyer et al., 1987; Ring and Van de Ven, 1994) . More recently, the ending-processes of supplier relationships have also received attention, both in the business-to-business (e.g. Alajoutsijärvi et al., 2000; Halinen and Tähtinen, 2002) and businessto-consumer relationships (e.g. Roos, 1999) . This paper investigates the switching aspect of the ending of business-tobusiness relationships, which refers to "such endings where the supplier (or the customer) is substituted for another alternative. The emphasis of the term is on one actor's decisions and actions related to relationship ending as well as the actor's formation (or strengthening) of another relationship" (Tähtinen and Halinen, 2002, p. 183) .

Although research on relationship termination has increased in recent years (see, for an overview, Pick, 2010) , supplier-switching remains largely unexplored (Wagner and Friedl, 2007) . Previous studies of supplier-switching have focused primarily on the switching decision. Key driving forces for switching suppliers reported by researchers include a desire to reduce purchasing costs (Friedl and Wagner, 2012) , the desire for independence from a particular supplier (Geiger et al., 2012) , and dissatisfaction (Ferguson and Johnston, 2011) . However, the perceived costs of switching and the value of established relationships often preclude a buyer from switching (Geiger et al., 2012) . Switching costs can be attributed to the following factors: information asymmetry (Wagner and Friedl, 2007) ; partner-specific investments and asset specificity (Heide and Weiss, 1995; Wathne et al., 2001) ; previous adaptations and the creation of actor bonds, activity links, and resource ties (Håkansson and Snehota, 1995) ; and the embedded nature of relationships and existence of personal relationships (Granovetter, 1985) .

While the literature has offered important insights into the determining factors for supplier-switching decisions, critics have claimed that these studies primarily present snapshots of the decision rather than capturing the dynamics of supplier-switching (Pfeiffer, 2010) . As Tähtinen and Havila (2013) observed, studies of termination tend to focus on the decision to end the relationship and not on the choice of initiating a new relationship to replace the first; it leaves the initiation of the new relationship "in its shadow" (p.12). Therefore, we lack knowledge of important aspects of the supplier-switching process. This paper addresses such gaps by providing insight into the processes of supplier-switching. It aims to answer the following research question: How do supplierswitching processes occur, and what actions and interactions do actors actually perform in this process?

To answer this question, the paper draws empirically on longitudinal case study research of a company and the involved actors' efforts to switch one of the company's key suppliers for another. It illustrates both the ending of the old relationship and the initiation of the new relationship. Theoretically, the research is built on a practice perspective (e.g. Schatzki et al., 2001) , meaning a focus on what the involved actors actually do when they engage in the switching process, as well as the practices upon which they draw. One central assumption within the practice perspective is that social life is an ongoing production that emerges through people's recurrent actions and interactions (Feldman and Orlikowski, 2011) . Rather than focusing merely on the process itself, this perspective argues for delving into the everyday activities of people involved, and how these influence organizational processes and outcomes (Johnson et al., 2003) . Another central assumption is that these everyday actions and the social order (structures, institutions, routines, etc.) in which they take place are mutually constitutive, meaning that social regularities are always 'in the making' (Feldman and Orlikowski, 2011) . Therefore, focusing on practices means focusing on both structures and agency, not either or.

Applying a practice perspective in this research on supplierswitching processes entails paying attention to what the actors involved actually do in the switching process, and how their everyday activities and the practices they draw upon influence the process and the resulting outcomes. As such, the research contributes to opening the black box of supplier-switching processes, and in doing so, extends existing research on supplier-switching, which has predominantly focused on identifying the variables that affect the supplier-switching decision and the different costs preventing the switch from taking place. The practice perspective also supports a deeper understanding of how supplier relationships are enacted by the people involved. Switching of supplier relationships involves a (re) enactment of structural conditions and day-to-day activities, such as the various interorganizational routines that make up relationships. In line with Whittington (2003) observation of organizational studies in general, studying what people really do in these relationships and how they engage in these routines, as well as how they go about the switching process, is an important step toward creating practical wisdom of supplier-switching processes.

The remainder of the paper is structured as follows. Section 2 reviews the literature on supplier-switching, and presents a theoretical basis for understanding switching processes based on a practice perspective. Section 3 describes the research design and Section 4 presents and analyzes findings from the case study. Section 5 discusses the findings in relation to the theoretical basis, and combines the insights into a conceptual framework for understanding supplier-switching processes in business relationships. The Section 6 includes conclusions, key managerial implications of the study, and also presents study limitations and directions for future research.

Business relationships develop as the partners engage in interaction processes in which they establish contact patterns and implement adaptations in the elements and processes of exchange (Håkansson, 1982) . These adaptations produce something unique by interlocking activities and resources of the two involved partners, and this uniqueness cannot be produced by either of the two actors alone, nor easily duplicated (Håkansson and Snehota, 1995) . However, relationships are not static, but change over time. These developmental processes have been captured in several frameworks. For example, Dwyer et al. (1987) described the relationship development process through a linear model involving five distinct stages (awareness, exploration, expansion, commitment, and dissolution). Ring and Van de Ven (1994) , on the other hand, conceptualized this as a cyclical process, suggesting that difficulties in a relationship might just as well lead to reconciliation and new development as to dissolution (Vanpoucke et al., 2014) .

While the overall developmental processes have been conceptualized through different multi-stage and lifecycle models, the dissolution stage and termination of relationships have also been the subject of various models, even though empirical research in this field remains sparse (Pick, 2010) . In a conceptual paper, Halinen and identified the different stages in the business relationship-ending process: an assessment stage, in which the parties start to evaluate the relationship; a decisionmaking stage, during which the parties execute either a voice strategyto restore the relationshipor an exit strategy, to terminate it; and a dyadic communication stage, in which the parties use different strategies to communicate the potential ending. After this stage, if the parties have elected not to restore the relationship, it moves into a disengagement stage, during which business exchange starts to decline, and a network communication stage, at which the decision is communicated to other network actors. Finally, the ending reaches the aftermath stage, in which the parties jointly develop a narrative of the relationship to explain the dissolution within their respective companies and their networks. While this model explains how relationships end, it does not explain the process by which the ending of one relationship takes place at the same time as it is replaced with a new one; in other words, the switching process. Most of the research on dissolution of business relationships still focuses on the ending processes of the old relationship, and largely ignores the establishment of the new relationship (Tähtinen and Havila, 2013) . However, switching refers not only to the conclusion of an existing relationship, but also to the establishment of a new one.

In recent years, supplier-switching in business relationships has received increasing attention in different fields: these include marketing (Heide and Weiss, 1995; Wathne et al., 2001; Ferguson and Johnston, 2011; Geiger et al., 2012) ; purchasing and supply management (Demski et al., 1987; Gadde and Mattson, 1987) ; and operations research (Wagner and Friedl, 2007; Friedl and Wagner, 2012) . Most of this literature has focused on the supplier-switching decision, and whether a buying firm should switch an existing supplier for a new one or instead choose to develop further the relationship with the old supplier (for example, Friedl and Wagner, 2012) . It is acknowledged that buyers are reluctant to switch suppliers, and that firms tend to maintain previous practices and decisions. Factors that have been found to influence the switching decision include a salesperson's customer orientation (Jones et al., 2003) , relationship equity (Low and Johnston, 2006) , customer dissatisfaction (Ferguson and Johnston, 2011) , and perceived relationship value (Geiger et al., 2012) . Another factor is information asymmetry (Wagner and Friedl, 2007) , which relates to an assumption of bounded rationality and the notion that even when actors attempt to behave rationally, they will encounter cognitive limitations. A lack of information about a potential new entrant's hidden characteristics (Wagner and Friedl, 2007) and uncertainty about the cost structures of both the incumbent supplier and the alternative supplier are likely to induce the buyer to retain an existing supplier (Friedl and Wagner, 2012) .

According to the literature, switching costswhich typically refer to a customer's perceived costs in switching suppliers (Campbell, 1985) are a key factor influencing the switching decision. An analysis based on transaction-cost economics indicates that a buyer would tend to stick with an existing supplier if the perceived transaction costs associated with the switch exceed the perceived benefits of finding a new supplier (Heide and Weiss, 1995) . Switching costs can be ascribed to prior commitment and partner-and transaction-specific investments (ibid.) such as physical assets, organizational procedures, employee training (Wathne et al., 2001) , technologies (Low and Johnston, 2006) , and the embedded nature of relationships (Granovetter, 1985) . Therefore, switching costs consist of both a hard-asset dimension, comprising investments in product modifications and dedicated plant facilities, and a soft-asset dimension, which relates to the quantity and quality of the working relationships between individuals from both parties and the communications system underpinning the interaction (Nielson, 1996) . According to Benito et al. (1999) , some switching costs are easily measurable, while others are less quantifiable, including outlays and expenses and the potential loss of more qualitative revenue and costs. Nielson (1996) found that investments in hard assets affect switching costs to a greater degree than soft investments do. He asserted that identifying the hard assets and assigning them an economic value is relatively straightforward compared to assessing soft assets, which are less tangible and harder to evaluate (ibid.).

Asset specificity results from partner-specific investments, and reflects the degree to which an asset can be employed across relationships without losing its value (Williamson, 1985) . Benito et al. (1999, p. 222) noted that "in order to take full advantage of an ongoing relationship, companies gradually adapt their organizational resources and routines to the specific needs of the relationship, which in turn represents a barrier to exit." Business relationships develop through exchange episodes that allow parties in a relationship to increase their mutual knowledge and experience (Ford et al., 2003) . In the exploration and development stages, in order to get to know each other, the parties engage in a detailed learning process to develop social, technical, and economic bonds (Håkansson, 1982) through coordination, adaptation, and integration (Håkansson and Snehota, 1995) . The resulting structure of connections reflects the parties' experiences with problem-solving and experimentation. Over time, routines are established that enable the interaction process and stimulate the relationship to mature (Håkansson, 1982) . These are valuable investments that make it worthwhile to continue a relationship, and avoid changes and switches (Gadde and Mattson, 1987) . Therefore, from the perspective of a customer, a high-involvement supplier relationship is "both the impulse to development and the cage that imprisons it" (Gadde and Snehota, 2000, p. 316) .

The literature has paid considerable attention to the motives of a buying firm in terminating a relationship, to the variables that discourage firms from termination, and to an array of different management strategies to handle relationship termination (Pick, 2010) . We have also seen that various stage models exist for describing the ending processes in business relationships (e.g. Halinen and Tähtinen, 2002) . However, far less attention has been given to the dynamics of switching (Pfeiffer, 2010) and the supplier-switching process itself (Tähtinen and Havila, 2013) . According to Pick (2010, p. 101) , little is known about who is influencing the process of termination and who is responsible for making the final decisionand subsequently, what kinds of interactions and network effects are in play in the process of reactivating former relationships. Furthermore, as Pick (2010) noticed, few previous studies have investigated the win-back process initiated by the supplier or the self-return process initiated by the customer.

This paper aims to fill the gap identified in previous literature, and to open the black box of supplier-switching processes. In line with the wider practice turn in contemporary social theory and management research (Whittington, 2006) , this paper draws on insight from the practice perspective. A practice lens allows the authors to delve deep inside what goes on in an organization and examine the micro-level managerial activities and roles in producing processes and outcomes, such as supplier-switching. At the same time, however, a practice perspective acknowledges the connection between these actions and the structural conditions and larger social orders surrounding them (Whittington, 2006) . The following section elaborates on the implications of a practice perspective on the study of supplier-switching processes.

A practice perspective on supplier-switching processes is likely to reveal many intriguing aspects. For example, from this perspective a supplier relationship is seen as an active accomplishment, whereby its structural features are the result of the actors' recurrent actions and enduring interactions (Feldman and Orlikowski, 2011) . It acknowledges a link between the structure of a relationship and its enacted dimensions (Jarzabkowski et al., 2012) . As is the case with most organizational life (Nelson and Winter, 1982) , key structural features of business relationships, such as connections between activities and resources, are embedded in inter-firm routines (Håkansson and Snehota, 1995) . Routines have long been seen as fundamental, since they enable organizations both to accomplish work and to economize on scarce resources (March and Simon, 1958) . A routine is defined as "a repetitive, recognizable pattern of interdependent actions, involving multiple actors" (Feldman and Pentland, 2003, p. 96) . Traditionally, changes in routines were attributed to exogenous drivers, and considered to be "mutations in the genes" of the organization (Nelson and Winter, 1982) . Another possible cause of changes is dissatisfaction, which leads to modifications in behavior in order to achieve a desired aspiration level (Levitt and March, 1988) . However, routines have typically been seen as stable and as a basis for inaction (Levinthal and March, 1993) .

More recent contributions have recognized routines as a source of both stability and variety, and have found that they change endogenously as people engage in their performance (Feldman, 2000) . Using a practice lens, these contributions have viewed routines as effortful and emergent accomplishments, which means that they can be reproduced in ways that are similar to existing routines or in ways that call for a more radical alteration of previously used routines (Feldman and Orlikowski, 2011 ). An important aspect of organizational routines is that they entail both an ostensive and a performative dimension (Feldman and Pentland, 2003) . The ostensive dimension refers to the abstract, generalized idea of the routinethe routine in principlewhich is easily transferred. The performative dimension refers to the actions of specific individuals that embody the routinethe routine in practiceand this, on the other hand, cannot be easily transferred. Taken together, the performative and ostensive dimensions allow for an understanding of routines as generative systems, which are created through a mutually constitutive and recursive relationship between the two dimensions (Feldman and Orlikowski, 2011) .

Conceptualizing routines as comprising both an ostensive and a performative dimension provides interesting insights into our understanding of what happens in the supplier-switching process, where similar routines are likely to be reproduced in the process of substitution. However, conceding that the performative aspect of routines is embedded in the practices of individuals, only ostensive aspects are likely to be reproduced, whereas new elements and the performative aspects will also need to be produced for the switch to occur.

General management literature has long recognized the problems of transferring knowledge across different boundaries, particularly focusing on the "stickiness" of knowledge and how it is embedded in the context in which it was first created (Szulanski, 1996) . This reflects the notion of asset specificity discussed earlier. For example, Dyer and Hatch (2006) showed how the embedded knowledge in Toyota's supplier networkreflected in a variety of routines and resource combinationscreates relation-specific capabilities that are difficult to transfer and very challenging for competitors to copy.

However, for a company to prosper and innovate, it must utilize previous learning investments such as routines. Carlile (2004) found that there are three basic approaches to managing knowledge across boundaries: (1) an information-processing approach, where knowledge is seen as an object that can be stored and retrieved, and where the creation of a common lexicon is seen as easing transfer processes; (2) an interpretive approach, which emphasizes the importance of creating shared meanings, and requires that knowledge is translated via cross-functional teams, colocation, and the use of particular individuals as translators; and (3) a political approach, which focuses on different interests, and requires negotiations, trade-offs, and the transformation of knowledge. This kind of knowledge transfer is aided by a pragmatic capacity to represent and model consequences of different actions, and these different approaches are useful under different circumstances. While reusing knowledge by transferring it across boundaries is helpful when the context is stable, novel situations require translation and transformation (ibid.).

The theoretical insights discussed above provide us with a useful lens with which to address the research question posed in the introduction: How do supplier-switching processes occur, and what actions and interactions do actors actually perform in this process? It offers an analytical framework to study these processes, which focuses on their internal life and the practices by which they are actually performed (Whittington, 2003) . The practice lens has two key implications for the study and the conceptualization of the supplier-switching process: first, it implies that it will be useful to examine how the relationships involved in the switchrelationship-specific knowledge, connections, and inter-firm routinesare produced and reproduced through the practices of actors; and second, it implies that these practices should be examined in detail, appreciating the connection between the involved actors' day-to-day activities and the structures in which they are embedded. The concept of routines is central in this respect. In sum, investigating the supplier-switching process from a practice perspective is likely to offer a valuable contribution to the understanding of relationship termination, which remains not very well understood (Tähtinen and Havila, 2013) .

This paper presents a case study of a company that has specialized in the distribution of products and services to the international marine industry, and its efforts to replace one of its most important suppliersin terms of relationship length, volumes, and purchasing costswith a new supplier. The case was considered suitable as it reflected the phenomena under investigation (Eisenhardt and Graebner, 2007). The case represents a typical event (Yin, 2009) in which a northern European company in a mature industry (marine) decides to pursue low-cost purchasing effects by replacing a long-term relationship with a high-cost European supplier with a significantly less expensive producer in Asia. Therefore, even if the study of this particular case is not able to make general claims, the findings might be seen as transferrable to similar contexts. Transferability is a key aspect of trustworthiness, which in turn is a key quality criterion in qualitative research (Halldórsson and Aastrup, 2003) .

A single case study design was chosen because it allows for detailed investigations and rich descriptions of relational phenomena and their dynamics and complexity (Dubois and Araujo, 2004) in their own contexts (Yin, 2009) . Given that switching processes are complex relational phenomena, rich descriptions were seen as important to provide in-depth understandings. This choice is in line with Johnson et al., (2003) , who argued that indepth studies are a necessary feature of micro-level explanations of organizations. These authors cited the significant contributions of small sample in-depth studies, particularly the seminal works by Henry Mintzberg (1977) and Andrew Pettigrew (1985) , in providing rich and enduring insight. Another example of single case studies are the studies by Jeffrey Dyer and colleagues of Toyota (e.g. Dyer and Nobeoka, 2000; Dyer and Hatch, 2006) , which have offered important insight into the strategies and inner life of the car manufacturer and how it deals with its supplier network. These types of studies have enabled the development of the holistic and contextual understanding that is essential to grasp the complexity of strategic change and stability (Johnson et al., 2003) . Thus, a longitudinal study of the particular case at hand was seen appropriate to provide new insight about the supplierswitching process. Based on the theoretical discussion earlier, such processes are poorly understood phenomena within the field of purchasing and supply management, and they are therefore suitable for case study inquiries, including single case studies (Dubois and Araujo, 2007) .

The case study was longitudinal and process-oriented, meaning that it was conducted in real-time, and was theory-driven and contextual (Pettigrew, 1997) . Qualitative data were collected between 2003 and 2006, providing rich information about the existing relationship and the switching process from its inception until the time when the new relationship was considered by the buying company as fully functional. Data were based on formal interviews, several hours of informal conversations during visits to the company's headquarters and international distribution center (IDC), observations in six supplier meetings, and the analysis of documents, including meeting memos, strategy documents, standard operating procedures, specifications, and e-mail correspondence. Several data sources were used to check the consistency of the data, and to help ensure the quality of the study (Eisenhardt, 1989) . In total, 28 semi-structured interviews were conducted (23 with representatives from the buying company, and five with people from the original supplier company), each lasting from 90 to 120 min. Table 1 presents an overview of the formal interviews and interviewees.

In total, 23 interviews were conducted with 10 different people from the buying company; three interviews with the chief procurement manager of the company, 10 with the procurement manager responsible for gas welding products (and thus responsible for the switch), two interviews with the product-technical manager responsible for gas welding products, one with the product engineer of these products, one interview with the international distribution manager of the IDC, one with the network supply coordinator, and five interviews with four operational staff members responsible for ordering, expediting, and stock planning (two interviews with the same planner). Serial interviews were conducted with a few key actors over time; for example with the procurement manager responsible for the switch.

The limitations of personal accounts reported in interviews as a means for understanding practice should be acknowledged (Alvesson, 2003) . However, the serial interviews with key actors allowed for tracking emerging perceptions of the switching process over time, which is important in processual analyses (Pettigrew, 1997) . Five of the 28 formal interviews were conducted with representatives from the old supplier: three individual interviews with the key account sales manager, one joint interview involving this sales manager, a product manager, and an order-handling staff member, and one interview with a logistics manager.

Even if supplier-switching represents an aspect of relationship termination seen from the customer's point of view (Tähtinen and Havila, 2013) , these interviews were nevertheless important sources of information about the switching process. No formal interviews were conducted with the new Chinese supplier, though, the main reason being practical concerns, since this supplier was located in China. It was also perceived by the contact persons in the buying company that little would come out of such interviews, since the Chinese were not likely to be open about issues concerning the business and the relationship as such. However, e-mail correspondence and meeting memos between the buying company and the new supplier were accessed and reviewed, allowing for interpretation of different events in the supplier-switching process from the new supplier's point of view as well. The interviews followed a "snowball" effect, where interviewees suggested subjects to be contacted for further interviews.

The interviews started off with two general interviews with the chief procurement manager of the buying company concerning the company's supplier relationships in general. The questions centered on the developmental aspects of the relationships in terms of time, involvement, connections, and adaptations in activities and resources. Insights from relational marketing (Dyer and Singh, 1998) and industrial marketing and purchasing, or IMP (Håkansson and Snehota, 1995) perspectives informed these interviews. In the interviews, the manager was asked to provide examples of long-term relationships, and the relationship with the old supplier in the case was mentioned. Contact was then established with the procurement manager responsible for this particular supplier relationship and respective product range. After an initial interview with this manager produced his recommendations for interview subjects, further interviews were conducted with product-technical managers responsible for the specific products and members of the operations staff.

This procurement manager also established contact between the authors and the sales manager of the old supplier. All of the key managers and operations staff involved in the switching process were interviewed. The interviews focused on questions about relationship development and management in the old relationship, including contact patterns, connections between activities and resources, routines, and trust. Then, when needs for cost reductions occurred in the buying company concurrent with a growing dissatisfaction with the old supplier, the interviews soon centered on the search for a new supplier and problem-solving activities with the old supplier. The questions now focused on the actions of those persons involved in the switching in particular, and how they perceived their actions and interactions throughout the process. To ensure the study's reliability, a database was created that included transcriptions of each interview and narratives of the case (Ellram, 1996) . To ensure validity, the interview notes and evolving case description were circulated to representatives of the companies for comments (Yin, 2009) .

In addition to the formal interviews, informal meetings and observations in several supplier meetings with the old supplier were conducted, as well as document analyses allowing for triangulation that also helped ensure the validity of the study (Halldórsson and Aastrup, 2003) . The supplier meetings primarily involved the same people that were interviewed, but sometimes others participated. Observations in these meetings allowed for verification of consistency between what the managers said in interviews and what they actually did in the interactions with the supplier.

The research presented here relies on analytic generalization, rather than statistical generalization. According to Yin (2009) , such generalization means that results are generalized to some broader theory, which in this context is theory about business relationships and switching processes. However, analytical generalization is not automatic, and requires substantial commitment to theory to ensure the explanatory power of the case study (Dubois and Gadde, 2002) . Therefore, the data analysis relied on an abductive research logic, which means that theoretical and empirical insights were intertwined and informed each other (ibid.). The analysis started with the relationship as the unit of analysis, operationalized as connections between actors and their activities and resources embedded in inter-firm routines (Håkansson and Snehota, 1995) . However, it soon became apparent that the success of the supplier-switching process relied on the actors engaging in a number of routines and practices. Therefore, the unit of analysis changed to practices, which are defined as recurrent, materially mediated, and situated social activities (Schatzki et al., 2001) .

The case study is a typical example of what Dubois and Gadde (2002) referred to as an evolving case. Qualitative techniques were used to analyze the data (Glaser and Strauss, 1967) , with a focus on relationships, routines, and practices. A chronological history of the case was created, highlighting dominant themes and events expressed by the interviewees, in combination with notes from the supplier meetings and archival material (Miles and Huberman, 1994) . The stories were communicated not only to the procurement manager in charge of the relationship, but also to research colleagues; among others, one professor who had worked with the buying company for several years on purchasing issues. Utilizing the NVivo qualitative software program, coding centered on recurring themes and the building of categories. This analytical process helped identify three key processes and a number of practices employed by the actors that enabled the switching process in the case. These three key processes were the following:

(1) initiation, a process whereby the actors engaged in legitimizing and search practices; (2) substitution, a process whereby the actors engaged in transfer, translation, and transformation practices; (3a) stabilization, a process whereby the actors engaged in institutionalizing practices; and (3b) restoration, a process whereby the old supplier was retained because of new requirements, and (3) individual interviews with one sales manager, and one (1) joint interview with this sales manager, a product manager, and an order-handling staff member A product manager participating in the joint interview referred to in the previous column

An order-handling staff member participating in the joint interview referred to in the previous columns where the actors engaged in repair practices. Table 2 describes the coding and how the categories were identified in the analysis.

The case company has a long tradition of serving the global marine industry via its worldwide distribution network of hubs and an international distribution center (IDC) located in a central European port. It has the reputation for high quality and reliable products, which over the years have been appreciated by the ship owners who represent its conservative customer base. Until early 2000, the supplier base was primarily located in Europe near the IDC, but at that time the company encountered a downturn in its markets, and needed to save money urgently. A new customer type had also arisen, involving companies that were most interested in the lowest possible prices. Top management urged the corporate purchasing department to find ways of reducing purchasing costs. They implemented a new sourcing strategy that focused on buying from low-cost countries in Eastern Europe and Asia. However, the product-technical and sales managers were reluctant to adopt this new strategy, arguing that it would harm the company's reputation as a high-end provider. Furthermore, the logistics were centered on the IDC, and buying from Asia was likely to increase the logistical costs and eventually force a change in the existing distribution system that had been in place for years. Nevertheless, Asia represented an attractive supplier market, and a significant portion of the sales were made to customers based in ports and shipyards in the same area. Therefore, the new strategy emphasized not only low costs, but also local sourcing.

This findings reveal what happened when the company decided to switch one of its long-term suppliers of gas welding equipment to an alternative supplier from China. The process was administered by the corporate procurement department, but to ensure that the switch ran as smoothly as possible, other departmentssuch as sales, the product-technical department, and logistics at the IDCwere also involved. At the outset, they expected the switch to be a simple task, although they anticipated issues related to cultural differences and geographical distance. Following a visit to the Chinese producer, the procurement manager responsible for the switch reported, "The quality shouldn't be a problem. Walking through their production facilities was just like walking through the facility of [the old supplier]." Referring to the termination of the existing relationship, the manager continued, "The dissolution process will come down to the remaining stock, and we will have hardly any contacts afterwards." Table 2 Analytical coding.

Drivers and activities Empirical extracts

Initiation of the switch Driven by cost-cutting strategies and dissatisfaction with existing supplier.

The buying company had a need for cost-cutting, and all units were asked to look toward low-cost countries for sourcing. Gas welding equipment was identified as one type of product suitable for switching, since it is relatively simple, and the old supplier was not performing well. Substitution of the relationships Driven by a trial-and-error process in which knowledge, connections, and routines are produced and reproduced.

An alternative source was chosen, and a process of establishing the new relationship took place simultaneously with planning the gradual demise of the existing relationship. The buyer tried to reuse knowledge, but needed to produce and reproduce new relationshipspecific knowledge.

Driven by a need to create stability and to enable returns on investments.

After a year of hard work, the new relationship was considered as "up and running," even if monitoring and adaptations were still needed. Restoration of the existing relationship Driven by exogenous changes; e.g., new safety requirements. New regulations concerning safety on some of the products required their adaptation. In the judgment of the buyer, only the old supplier was capable of doing this.

Legitimizing and search Providing information and visualizing (legitimizing), pointing toward and identifying alternatives in the supplier market (search).

The responsible procurement manager gathered statistics and price information for legitimizing the switch, with regard to both the internal organization and the old supplier. Procurement and product managers searched for an alternative source in the worldwide supplier market by using internet resources and by in-person visits. Transfer, translate, and transform

Using standards and written material to specify requirements (transfer), establishing cross-functional teams and using a "broker" (translate), interacting to develop new knowledge required (transform).

A standard operating procedure was used in the switching process, and substituted with knowledge codified in product specifications and other written procedures. To create a common understanding of needs, a team of managers from all involved functional areas worked together, and an exclusive sourcing unit was established with a key contact to mediate between the buying company and the new supplier in China. The team had to visit the supplier several times, since written information did not suffice. The buyer discovered that the existing knowledge could be reused only to a certain extent, and new relationship-specific knowledge had to be created, even if it was considered to be a standard supplier relationship.

Embedding new knowledge into routines and ICT systems. The visits enabled the settlement of agreements about delivery and contact patterns, and operational routines were finally in place. Knowledge concerning the relationship, such as product information and scheduling, was included in the parties' ICT systems.

Re-establishing connections through continued buying-delivery and product development discussions.

The relationship with the old supplier never actually dissolved.

Because of delivery problems with the new supplier and new safety requirements, the relationship continued for certain products.

The company had made several switches of European suppliers a couple of years earlier in a supplier development project called SWOP (Supplier Worldwide Opportunity Project). The old supplier of gas welding equipment survived this exercisebasically because it agreed to lower its prices, but also due to the perceived switching costs. The fact that some suppliers were actually switched came as a surprise to the supplier base. At the time, the corporate procurement manager said, "Earlier, everyone knew that if you became a supplier to us, you were in for life." The project was considered quite successful, but had invoked internal resistance, primarily due to different objectives between the sales and product-technical managers on the one side, and the procurement managers on the other.

A few years previously, the company had also attempted to set up a sourcing unit in China, which had not been successful because of, among other reasons, the outbreak of Severe Acute Respiratory Syndrome (SARS). However, the timing was now considered right, and the involved actors were determined to learn from their previous experiences during SWOP.

Switching the old supplier to the new Chinese supplier took approximately two years, from the time first contacts were made to the time the new relationship was considered "up and running." During this process, routines that connected the company with the old supplier were dissolved, at least partly, and new, similar ones were established in the new supplier relationship. Therefore, as we shall see in the case description, this turned out to be only a partial switch (Wagner and Friedl, 2007) , since the company eventually decided to keep the old supplier for some of the product range. The following sections presents the case.

The relationship between the buying company and the old supplier dated back to the time the two companies were born from one parent company in the early 1900s. The relationship was often referred to as a "family" relationship, at least in the old days. The parties had been doing business for over 20 years, and were well-connected through personal relationships and a number of routines that enabled the day-to-day operations. The supplier's sales manager noted the following:

"Many of the same people have been around for years. Both me and [the product manager at buying company] have been here since the early 1980s. When I was away for a couple of years, he actually demanded that I come back! " Such types of personal relationships are common in business relationships (Dwyer et al. 1987 ). On the one hand, they are positive, since fewer resources are needed for monitoring and control, but on the other hand, they may prevent thorough evaluation of the relationship and lead to lock-in situations (Gadde and Snehota, 2000) . The formal meetings usually took place as part of annual negotiations. By the end of the 1990s, the relationship had changed into a more standard and transaction-oriented one. Explanations for this development were the establishment of a corporate purchasing function by the buyer, personnel turnover, and a joint perception of the relationship as being stable and unproblematic.

The relationship was considered as "business as usual" until 2002, when, encouraged by the SWOP, the buying company established a performance-evaluation routine. This soon revealed that the supplier was not performing as well as once thought. The corporate procurement manager commented, "This illustrates that business as usual is not necessarily good business! " Great efforts were made to improve the delivery performance, but at the same time, the corporate procurement department implemented a new low-cost sourcing strategy, and started to search for alternatives in Asia. The combination of the new sourcing strategy and dissatisfaction with the supplier sparked the switching process. Therefore, the initiation was due both to endogenous reasons such as dissatisfaction, which is a common motive for switching (Ferguson and Johnston, 2011) , and reasons exogenous to the relationship, in that the competitive situation accentuated the need for cost savings. The latter explanation shows the importance of acknowledging the context in understanding these processes (Johnson et al., 2003) .

The procurement manager, as the person responsible for the gas equipment product range and the associated supplier relationships, was a key driver of the process. He was relatively new, and was very eager to challenge the suppliers and to reduce purchasing costs. He had no personal relationships with representatives from the supplier, and an important task for him was to provide the internal organization with arguments for switching the targeted supplier, particularly to managers within the sales and product-technical functions, who were more emotionally involved. This illustrates Ferguson and Johnston's (2011) point that different people in the buying company may act differently in their dissatisfaction responses. Top management issued a clear mandate to save overall company coststhereby ensuring competitiveness by reducing purchasing costs, particularly through low-cost sourcing. Few disputed the legitimacy of the need and the optimal means to the end, but still there was a job to do in terms of creating full commitment to the decision of switching the supplier of gas welding equipment. The procurement manager expended great efforts into collecting and communicating price information from alternative suppliers, thereby showing the reductions that could be achieved. To underscore the fact that the old supplier was underperforming, reports from the performance evaluations of the old supplier and complaints about technical and delivery issues were discussed across the departments.

Similar legitimizing was used toward the supplier. Delivery problems led to more frequent meetings throughout 2003. In these meetings, numbers and figures were presented to the supplier illustrating the buyer's dissatisfaction. At one of the meetings, the procurement manager provided some figures, and said, "We do not want to lose you as a supplier and you do not want to lose us as a customer, but the performance measurement graphics look like a roller coaster! " He was clear on the point that if the supplier did not improve, he would take it personally, and mentioned that they were looking to low-cost countries for alternative supply. The supplier considered this a real threat. Even if its managers expected the buyer's product and sales departments to delay low-cost sourcing because of quality issues, they admitted that it would probably just be a matter of time before the customer found a low-cost alternative that could meet quality standards, at least with the simpler products.

The company used internet search engines to find information about the supplier market, and the procurement and product managers attended exhibitions around the world, where producers of gas welding equipment demonstrated their products. They also contacted a Taiwanese producer that they had discovered during the SWOP, and managers from the buyer went there for a visit. As a result of the conventions and visit, the managers returned home with samples of products produced in Asia, as well as price information. The samples showed that the Asian producers were able to deliver the quality needed.

The procurement manager referred to this initial phase as "image building." As part of legitimizing the switch, the procurement department also made reference to a previously conducted blind test of two products, one produced in Europe and one in China, where people within the company were unaware of which product had been produced in which continent. Important factors for initiating the switch were legitimizing the switch internally and searching for real alternatives. These activities centered on collecting and providing information about price and quality issues, and recognizing the problems with the old supplier.

Representatives from the buying company first made contact with the Chinese supplier at an exhibition in the fall of 2003. It turned out to be a large corporation, supplying a wide range of products to European customers. The buying company had been ready to sign a contract with the Taiwanese supplier, but after lengthy discussions the fact emerged that the price savings would not outweigh the perceived switching costs; in addition, a visit to the production facilities in Taiwan did not provide an overall impression of high quality.

Next, contacts were made with the Chinese producer, and inquiries were made regarding some of the old products. The Chinese company returned with a price that was only 25% of the price the company was paying the old supplier. This encouraged the procurement manager to visit the supplier in early 2004, and he was impressed by the modern production facilities. A key concern in terms of buying from China was the high rate of manual work, compared to the automatized production activities used by Western companies; however, this facility seemed to be able to match the production of European suppliers. The product manager, the logistics manager of the IDC, and quality-assurance staff made several visits. A few problems were anticipated, but the new relationship was expected to be similar to the company's regular relationships with big suppliers. As the product manager explained:

"If there is a relationship with a supplier, this supplier supplies the products, which are then further distributed. Questions and requests about these products will come from customers to our local network, which will then contact me, and I will contact the supplier. It's more a question of day-to-day interaction."

The process of substituting the old supplier for the Chinese supplier soon began, a learning process driven by trial and error. The logic here follows Ferguson and Johnston's (2011) findings that when the type of product falls within the commodity category and there are alternative sources, customers are more likely to actually exit a relationship. Since the new supply chain varied from that of the old supplier, the new relationship required the accumulation of new knowledge, connections, and inter-firm routines. However, managers drew on previous experience to reproduce what the company considered to be standard routines for ordering, delivery, and expediting. According to the logistics manager, changes in the routines were considered primarily as a "question of scheduling and changing numbers in the system." Thus standard routines for ordering and delivery were establishedonly with different scheduling.

The basis for the switch was a standard procedure for switching processes that had been developed during the SWOP. The procurement manager prepared an implementation plan with 56 steps, which was then given to staff in the product-technical department so that they could add their own steps. Based on the experiences from the SWOP, the managers recognized that certain things could likely go wrong, and took precautions against them. To avoid running out of stock required a well-coordinated plan to decide when to stop buying from the old supplier and when to start buying from the new supplier. The procurement manager did not want to tell the old supplier about the switch before he was certain that the Chinese supplier would be able to provide the quantities and quality required. He described the dilemma between informing the company's internal organization and the risk that someone would tell the supplier:

"Some of our employees have been in contact with [the old supplier] for many years. Thus, it is always a risk that someone 'happens' to mention the issue to them. However, everybody understands that this replacement needs to be done, and that otherwise we will lose our business." However, other representatives from both the buying company and the old supplier expressed that they were able to discern what was going on, even if it was not explicitly communicated. In some ways, this strategy resembles what Alajoutsijärvi et al., (2000) termed a "silent" exit strategy. The specifications and drawings of the existing products and the delivery terms were handed over to the Chinese, but this information-sharing proved insufficient. The supplier complained that the products were much more complex than anticipated, and not only kept raising the price, but asked for help.

The need was clear for the development of a shared understanding of the relationship and what it required. One means to achieve this understanding involved the decision to run the substitution phase as a cross-functional project, where people from various functions worked closely together both internally and with the Chinese producer. There were initially some different views among the groups with regard to the products. While procurement considered the products to be relatively "simple," and believed switching would primarily require standardized components to be put together in new branded packaging, producttechnical managers acknowledged that at least some of the products in the range were more complex, and had been developed through various adaptations in the old relationship over the years. This illustrates a sensitivity to what researchers have referred to as stickiness of knowledge (Szulanski, 1996) and asset specificity (Heide and Weiss, 1995) .

The divergent views among the involved actors were acknowledged, and the procurement manager requested that as many employees as possible get involved, to be sure that all aspects were covered. Direct interaction with the new supplier was considered a necessity, as written instructions and agreements proved to be a poor mode of communication. A team consisting of the procurement manager, product-technical managers, and the IDC manager visited China regularly during the spring of 2004 to establish the relationship. This approach was both to ensure a mutual understanding of the agreement and because the team recognized that personal relationships are highly important in the Chinese business context. In addition, a sourcing unit was established in China with a person responsible for daily contact with the supplier. This person was seen as an important broker to mediate information between the buying company and the new supplier. In May, the agreement was finally signed.

Despite efforts to create a mutual understanding with the new supplier, it soon became clear that the parties had quite different views on doing business, and that they had different interests. To solve these issues, the actors engaged in direct interaction and negotiations in order to create the new relationship-specific knowledge necessary for success.

When the supplier was unable to deliver the first shipment, scheduled for mid-October 2004, on time, the company did not inform the buyer until the day before the due date, admitting that the shipment would be delayed by several weeks. This forced the buyer to continue buying from the old supplier; however, the old supplier had learned of the switch in a meeting earlier that fall, and had reduced stocks of those products. Now, when the buyer needed the products again, the supplier fortunately agreed to resume production. In the words of the sales manager, the supplier made this decision "because of our good relationship" and "because we don't want to shut the door." This illustrates that not only is it important to consider the customer's responses to dissatisfaction, but also the behavior of the supplier in trying to win back the customer (Pick, 2010) .

To make matters worse, severe communication problems occurred with the new supplier. In a meeting following the delay, the contact person at the sourcing unit started an argument with the president of the Chinese producer, accusing him of not being in control of the situation in front of his whole management team. This accusation made the president very angry because he had "lost his face," and he threatened to cancel the whole agreement. The situation was resolved by the procurement manager, who apologized and agreed to replace the contact person. As he explained afterwards, "I really had no choice… I couldn't afford to jeopardize the whole relationship after all the efforts we'd made." Failure was simply not an option, and this interesting development illustrates the perception of the manager that there could be no turning back once the process had started.

Throughout the process, it became obvious that the new supplier had little knowledge of the technical properties of the products, and would need a lot of extra help from the buyer. The products had been developed long ago in close interaction with the former supplier, which in fact was the more knowledgeable of the two parties. A great deal of time and effort was expended to resolve product-technical issues concerning the right way to connect components and fit them together, and the product managers from the buying company had to go to China several times, since written instructions did not work very well. Furthermore, much of the production process at the Chinese facility relied on manual labor, and no proper quality-control system was in place. Another problem was that the new supplier was reluctant to invest in testing, which was an absolutely necessity for some of the products. The supplier viewed the buyer as a small customer and did not give them a high priority, and the supplier was at first reluctant to invest in expensive quality-assurance systems and test equipment.

Through lengthy negotiations and discussions, agreements and solutions were eventually finalized. The supplier eventually made the necessary investments, considering them as a means to attract other European customers. The buying company acknowledged the need to spend more time on developing the relationship, the interaction processes, and the operational routines, compared with its more standard buying relationships.

After a year of hard work, the managers regarded the switch as a success, and the new relationship as relatively stable. There were still problems, but they were solved on a day-to-day basis. The new relationship was eventually institutionalized (Håkansson, 1982) in terms of establishing operational routines, exercised through alignment of schedules, agreeing on delivery and contact patterns, and use of the ICT system, which helped enhance the stability of the new relationship. The responsible procurement manager recognized that the new supplier relationship needed continuous monitoring and incremental adjustments. Quality remained an issue, and a qualitycontrol routine for shipments from China was incorporated into the IDC. The question now concerned whether the Chinese supplier should start delivering to other hubs in Asia, in addition to Europe. The costs inherent in the existing delivery structure, in which products were shipped from China to Europe and then back to Asian customers, could be offset to some extent by the low prices. However, the competition was increasing month by month, and the entire existence of the IDC was questioned.

The managers admitted that they had underestimated the work involved in establishing the new relationship. It became clear that the standard operating procedures for switching suppliers and establishing new supplier relationships, developed in the SWOP, were not appropriate in this new context, which involved vast geographical distances and cultural differences. This shows that routines are indeed hard-won accomplishments, and that since people are involved, they are always subject to change (Feldman and Orlikowski, 2011) . In the end, the consensus was that there was no way to measure exactly the overall costs of the switching process. Nevertheless, a common perception among the managers was that the switch had been worth the time spent due to the substantial reductions in purchasing costs. In the words of the product manager:

"The key issue is that the products are 'fit for the purpose.' It is of no use that they are far above the standards, because the price will increase accordingly, and our customers will say, 'Why should we pay this extra money for quality that is far above what is necessary? '"

During the switching process, contacts with the former supplier were maintained. One reason for this was that the initial delivery problems with the new supplier prolonged the old relationship. However, the main reason was new safety requirements imposed on some of the products, which required complex product adaptation. The buyer did not consider the new supplier to be knowledgeable enough to achieve such modifications. As a result, new product development discussions took place with the old supplier, headed by the product-technical manager of the buying company and the sales manager of the old supplier, who had known each other for over 30 years and enjoyed a trusting relationship. These new commitments resulted in a restoration of the relationship, at least for the most complex parts of the product range. The actors engaged in direct interaction and meetings to reestablish the relationship and the connections (Håkansson and Snehota, 1995) . Product development discussions took place, and delivery and order routines were reinstated. No one perceived this as a problem since the old routines had always worked quite smoothly on both sides. Furthermore, even if both parties recognized that the relationship had been harmed during the delivery problems and the switch process, it "was just business, and not personal," as stated by the supplier's sales manager. The managers at the buying company admitted that they had also underestimated the value of the old relationship.

The objective of this paper was to examine how supplierswitching processes take place in business relationships and what actors do in this process, and the study aimed to contribute to a conceptualization of this process by combining case study research and insights from a practice perspective. Literature contributions specific to switching have focused on the switching decision (e.g. Wagner and Friedl, 2007) . However, few researchers have investigated the dynamics of switching (Pfeiffer, 2010) , and what happens and what actors actually do in a move from one supplier to another (Tähtinen and Havila, 2013) . The findings presented here provide insight into these processes and form a basis for developing an initial conceptualization of the supplier-switching process and the practices that enable it. This conceptualization includes not only the ending of the existing supplier relationship and the initiation of one with the new supplier, but also the process of substitution itself. Based on the findings from the current study, we may identify three key processes and associated practices that enabled the overall supplier-switching process. Fig. 1 illustrates the initial model of this supplier-switching process and the practices that are likely required for this process to happen. These practices are not merely something the actors do, but represent "embodied arrays of activities organized around a practical understanding or 'way of doing'" (Bjørkeng et al., 2009, p.14) .

The first process, initiation, describes the motivation and efforts that drive the switch, and what actors do in this process. As we learned from the findings section, actors in the case engaged in what we may term legitimizing and search practices. As in many other cases (Friedl and Wagner, 2012) , the switch in this case was motivated mainly by top management's focus on cutting costs, supported by the competitive situation, a new low-cost strategy, and a dedicated and professional procurement department. Dissatisfaction with the existing supplier also influenced the decision. Therefore, reasons both exogenous to the relationship (for example, the competitive situation) and endogenous to it (such as dissatisfaction) motivated the initial desire for switching. Despite a collective agreement to cut costs, legitimizing practices needed to occur in order to prepare for the switch. There was a general internal resistance toward buying from low-cost countries, and the perceived negative consequences for the company's reputation as a high-end supplier of quality products.

Regarding the specific supplier of gas welding equipment, the buyer also acknowledged that there would be some switching costs due to the personal relationships and adjustments in the products and processes that had been made over the years (Dwyer et al., 1987; Gadde and Mattson, 1987) . However, the managers responsible soon succeeded in creating a shared view among the rest of the managers involved that the benefits would outweigh any such costs. The findings agree with much of the existing literature in regard to the switching decision, including results on cost considerations and the importance of previous investments. However, the case illustrates the practices actors employ in the initiation process, such as the importance of legitimizing. The second practice that was particularly important in this process was search practices. Since lack of real alternatives is likely to prevent the decision to end and switch a relationship (Ferguson and Johnston, 2011) , it was critical for the actors, such as the responsible procurement manager, to not only legitimize the switch, but also search for and identify real alternatives. In some respect, the search activities formed an important basis for the legitimizing practices.

The second process, substitution, represents the actual switch from the old supplier to the new one. Here, the findings from the case study reveal that the actors used practices producing and reproducing knowledge, connections, and inter-firm routines. These are referred to as transfer, translate, and transform practices. In the current case, representatives from the buying company considered the involved supplier relationships, products, and routines to be standard ones. However, even if the process involved standardized routines, transfer practicesin terms of handing documentation over to the new supplier and reusing standard procedures and specificationsproved insufficient. The managers had to create a common understanding of differences between the old and the new relationship, and they had to translate knowledge embedded in the old relationship and adapt it to the new relationship. They had to identify exactly which aspects could be reused and which could not, and what needed to be developed in terms of new knowledge. The development of this new relationship-specific knowledge had to take into account the different interests among the parties, and required negotiation and direct interaction.

If companies are to prosper and innovate, it is essential that they utilize previous learning investments (Carlile, 2004) . However, knowledge is "sticky" (Szulanski, 1996) , and relationshipspecific knowledge can be an important competitive advantage because it is hard to copy (Dyer and Hatch, 2006) . The findings from the study indicate that the managers of the buying company underestimated the complexity and significance of the existing routines, and they treated the products as standardized and easily transferable. In other words, they considered only the ostensive aspect of the routines and neglected their performative aspect (Feldman and Pentland, 2003) . While identifying and transferring the ostensive aspect of a routinethe routine in principlecan be simple, the company must translate the actual performance of the routine into the new situation, and also transform the routines to include new elements (Carlile, 2004) .

The third process, stabilization, consisted of what may be termed as practices for institutionalizing the new situation in which the old relationship was being phased out (or at least the attempt was made) and the new relationship was in place. Routines were established, enabling a return on previous investments (Levinthal and March 1993) . However, as the case study showed, this process was complemented by a restoration process, in which the old supplier relationship continued in parallel with the new one. Previous literature refers to this as "partial-switching" (Wagner and Friedl, 2007) . This restoration process was in turn enabled as the involved actors used repair practices, such as new discussions and re-establishment of interorganizational routines aimed at restoration of the old relationship. As seen in this case, this process took place in parallel with the stabilization of the new relationship, indicating a cyclical process (Ring and Van de Ven, 1994) .

The findings of this case study of the supplier-switching process make two key contributions to the literature: (1) they identify key processes involved in the overall supplier-switching process, which capture the (partial) ending of an old relationship, establishment of a new relationship, and the substitution process occurring between these two relationships; and (2) they demonstrate the usefulness of a practice lens, in allowing us to capture what actors actually do in the switching process, and in allowing us to understand the relationship-specific aspects embedded in interorganizational routines (which, in turn, were found to heavily influence the supplier-switching process). If we consider these contributions in more detail, the findings provide a basis for a conceptualization of the supplier-switching process (Fig. 1) , which focuses not on the development and termination of one relationship per se, but on the process (and the practices enabling them) in which one supplier is substituted for another . If we consider the first contribution, it is important to acknowledge that switching has been on the corporate agenda for decades. However, few studies have delved into the switching process itself. When the process is considered, there is a bias toward looking at the ending of the old relationship, while not much attention is given to the (simultaneous) initiation of the new relationship (Tähtinen and Havila, 2013) . The extant supply management and marketing literature has dealt primarily with the supplier-switching decision and its associated costs, and often from a "one-shot" perspective (Pfeiffer, 2010) . The reasons for switching in our case were both endogenous (Ring and Van de Ven, 1994) , in terms of cost-cutting strategies and increasing dissatisfaction with supplier performance, and exogenous, in terms of a more fiercely fought, cost-focused competitive situation.

The findings are in accordance with Ring and Van de Ven (1994) , who proposed that an imbalance in formal and informal processes is likely to increase the likelihood of dissolution. In this case, the new corporate procurement office and the professional procurement managers were seen as key drivers of the switch. Over the years, with new "business-oriented" people becoming responsible, the personal relationships were weakened. Furthermore, the increased focus on formal measurement and monitoring of supplier performance revealed the performance problems, and use of this evidence in the negotiations further escalated the conflicts in the relationship. Social commitments on an individual level have proved just as important as economic considerations in explaining why companies are likely to preserve a relationship; breaking long-term relationships often causes psychological, emotional, and physical stress (Dwyer et al., 1987) . The case showed similar tendencies.

Even if personal bonds were seen to have been weakened, the fact that the old supplier helped out when problems occurred in the new relationship was explained by the sales manager to be due to "our long-term relationship."

The case turned out to be a case of only partial switching, which is interesting. One reason for restoring parts of the old relationship was because changes in the surrounding network dictated new safety requirements and altered customer needs. According to Pick (2010) , there are few examples and little empirical knowledge about the impact of networks on customer recovery and the decision to restore a former relationship. In the case, we see that the new regulations in some ways compelled the customer firm to enter into new product development talks with the old supplier. In this way, the findings confirm that the relationship development process might be cyclical (Ring and Van de Ven, 1994) ; even if the relationship with the old supplier had gone through a stressful and problematic dissolution phase, the new product development discussion implied that the relationship went through new exploration and commitments. Therefore, contrary to Jap and Anderson (2007) , the findings suggest that the relationship enjoyed a fresh start, and that the "psychological scars" were not influencing the new discussions in a negative manner. The comment by the sales manager indicating that the intended dissolution was just about business, not anything personal, may explain this finding. This finding also expands insight into the trust dimension of relationships, and supports the contention that trust occurs at different levels and is embedded in the personal relationships between managers. The product development discussions took place between the product-technical manager of these products at the buying company and the sales manager, who enjoyed a trusting relationship. Certain business relationships might have different contact points, and while some actors may be dissatisfied and eager to switch, other actors might not share this point of view (Ferguson and Johnston, 2011) . The restoration process might also be explained by what Palmatier et al. (2013) referred to as "commitment velocity"; new commitments were established through the new product development discussions and investments, with the result that the old supplier was only partially replaced in the provision of equipment for the gas welding product range. This illustrates a central argument in the IMP perspective (e.g. Håkansson and Snehota, 1995) , which states that connections in long-term relationships bring opportunities for change and innovation. These established connections in activities and resources might explain why the old relationship in our case was restored; existing knowledge embedded in the products was vital to deal with the new safety requirements. This finding complies with previous literature on switching costs, referring to such connections as idiosyncratic and partner-/transaction-specific investments (Heide and Weiss, 1995; Wathne et al., 2001; Jap and Anderson, 2007) .

If we move further to the second contribution, applying a practice lens allowed the identification of not only the processes enabling the overall supplier-switching process and its stability and change features, but also the actions and interactions through which these outcomes were achieved over time (Feldman and Orlikowski, 2011) . The sub-processes comprising the overall supplier-switching process were enabled by different practices. As Jarzabkowski et al., (2007, p. 9 ) observed, practices might be seen as "resources through which multiple actors are able to interact in order to socially accomplish collective activity." These involve both symbolic and material tools, and when they are utilized in routinized ways that form patterns, they may be studied to understand how organizational activities are constructed (ibid.). In other words, studying practices involved in the supplier-switching process in the case provided us with in-depth knowledge about how the switching activity was constructed, and insight into what brought about the change and the stabilizing dimensions. According to Ferguson and Johnston (2011) , individual members might play different roles in the process, and dissatisfaction responses depend on the influence they possess. In our case, we saw that the responsible procurement manager was the key promoter of the switch, while the product-technical managers were key opponents (Pick, 2010) . To overcome the resistance, the procurement manager used different legitimizing practices, including searching for alternatives and providing arguments for switching to the opposition. An important part of these influencing efforts was to involve the opponents in the switching process, creating buy-in and commitment to the decision.

In some ways, the overall switching process may be seen as a routinized practice. The buying company had previous switching experiences, which were codified into a standard operating procedure for switching suppliers, and which the managers attempted to draw upon in the case. However, as we saw in the case, it is clear that such procedures primarily capture the routine in principle, not its actual performance.

The practice lens also provided insight into how relationships embedded in relationship-specific knowledge, connections, and routines (Håkansson and Snehota, 1995) are enacted by people engaging in them. Business relationships are embedded into different interorganizational routines, including those for ordering, delivery, payments, and meetings. However, we know relatively little about interorganizational routines, as most of the routines literature deals with intra-organizational phenomena (an important exception being Zollo et al., 2002) . Similar to the previous notion about the intra-organizational switching routine, the interorganizational routines in the case were also produced and reproduced through a constitutive relationship between the structural features of the routines and the specific performance of those particular routines (Feldman and Pentland, 2003) . The case showed how the managers of the buying company considered the old relationship and the routines to be relatively standard and therefore easily transferable. However, despite similarities in the ostensive aspects, the switch required translating and transforming routines into the new context (Carlile, 2004) . For example, the delivery routine in the old and the new relationship had certain standard features, but the actual performance of this routine was based on knowledge of and adaptation to the particular relationship and its context. As people perform routines, they are always subject to change (Feldman, 2000) .

This paper falls within the stream of research that deals with relationship dynamics in general (e.g. Håkansson, 1982; Dwyer et al., 1987; Ring and Van de Ven, 1994; Håkansson and Snehota, 1995) , and the processes of relationship ending, including the supplier-switching process, in particular (e.g. Tähtinen and Halinen, 2002; Halinen and Tähtinen, 2002; Pick, 2010; Pfeiffer, 2010; Ferguson and Johnston, 2011; Tähtinen and Havila, 2013) . Furthermore, it is in line with contemporary social science research that uses a practice perspective to study organizational phenomena (Schatzki et al., 2001; Whittington, 2003; 2006; Jarzabkowski et al., 2007) . Today's practice turn in the social sciences is largely motivated by a perceived need to re-introduce the role of human agency in research on organizational work (Jarzabkowski et al., 2007) . This paper adds to the purchasing and supply chain management (SCM) literatureas well as to business marketingby showing how the practice lens may reveal important aspects of the switching of suppliers, how the switching process occurs, and what are the specific actions an interactions of actors when they substitute one supplier relationship for another. As mentioned above, this process is still not very fully understood (Tähtinen and Havila, 2013) .

Previous literature has emphasized that understanding the ultimate stage in relationship lifecycles is paramount for managers in order to develop suitable relationship strategies (e.g. Pick, 2010) . Relationships are valuable assets for firms (Håkansson and Snehota, 1995) , but come with both benefits and costs (Gadde and Snehota, 2000) . Therefore, understanding the complexities of supplierswitching processes is important both from a customer and supplier perspective. A key managerial implication of the case study research presented in this paper is that managers should not underestimate the embedded nature of relationships, even in situations where the relationship is considered to be standard and "business-as-usual." The study emphasizes the importance of looking beyond the routines in principle and in the abstract (the so-called ostensive aspects), and scrutinize the actual performance of these routines (the performative aspects). This is an important insight for any buying firm seeking to replace one supplier with another.

Long-term business relationships are likely to be characterized by adaptations in activities and resources over years of interaction. These connections are not necessarily easy to identify; some people might have moved on, and the relationship might have been taken for granted. In addition, connections transcend the dyadic relationship to include third parties, such as other suppliers, sub-suppliers, and customers of both parties (Håkansson and Snehota, 1995) . For managers, this means that existing relationships must not be regarded too lightly in estimating the true costs of switching. Such connections represent significant switching costs that must be taken into account when considering the tradeoffs between the benefits and costs of switching. Another important implication is that these connections are not easily transferable across relationships. In the case, the managers initially considered the routines to be relatively similar. However, it is important for managers to recognize that such routines are enacted by people in their own specific contexts. Therefore, as the case demonstrates, managers need to engage in translation and transformation practices (Carlile, 2004) , including creating a shared understanding among the people involved in the switch and developing new knowledge of the new context, which the new relationship represents.

Even if switching is seen as a decision considered primarily from the point of view of buying firms, the findings also have implications for supplier companies. Suppliers need to understand the motives of their customers to switch, as well as the actions and interactions of the people involved in the switching process. In this particular case, the old supplier was only partially switched. This can be attributed to the product-specific knowledge that the supplier possessed and also to the existing personal relationships between managers in the two companies, as well as the supplier's behavior in helping out the customer when deliveries went awry with the new supplier. The scope of this study did not include considering "win-back" strategies on the supplier's behalf, but the research nevertheless suggests that suppliers should consider investing in technology and knowledge, particularly when facing competition from low-cost countries. Furthermore, as previous literature has emphasized, personal relationships are important (e.g. Nielson, 1996) . Using these relationships, and influencing managers of the customer company who are likely to be opponents of the switch, is one way in which a supplier might take action; another is by the response behavior in the process, as representatives of the buying firm voice their dissatisfaction (e.g. Ferguson and Johnston, 2011) . Pick (2010) noticed that, often, losing a customer is interpreted as a result of external forces and completely out of the control of the supplier. The findings from the research presented in this paper indicate that the supplier is able to influence the overall switching process.

The study offers several avenues for future research. It is likely that the practice turn that is currently observed in contemporary social theory and general management and strategy research (Whittington, 2006) will be increasingly adopted in research within the traditions of business marketing, supply chain management, and purchasing and supply management. Looking at the micro-practices of the people involved, the dynamics of business relationshipsincluding switching processes and the decision of a buying company to substitute one supplier for an alternativeis likely to provide a fruitful avenue for future research. More specifically, future research could continue to test and refine the proposed process model of supplier-switching in business relationships across different organizations and business contexts. There are still many gaps to be filled in research on the ending of relationships (Pick, 2010) , and the switching process itself remains in the shadows (Tähtinen and Havila, 2013) . The research presented in this paper is just a first attempt to delve into the micropractices of supplier-switching processes, bringing this important aspect of business relationships out into the light. However, there is indeed a need for further research.

The study has several limitations. As a qualitative single-case study, this paper does not provide any statistical generalizations and caveats apply with regard to the findings, even though various quality-enhancing efforts, which were discussed in detail in section three of this paper, have been undertaken to increase their trustworthiness and validity. The case-study approach provides a basis for analytical generalizability (Yin, 2009) , and it is compared and related to theoretical insight. Therefore, even if not generalizable, the findings from the case study in combination with the theoretical insights are likely to be relevant in settings similar to the case context where companies consider to replace one supplier with another.

Beautiful exit: How to leave your business partner

Beyond neopositivists, romantics, and localists: a reflexive approach to interviews in organizational research

Foreign operation methods and switching costs: conceptual issues and possible effects

Becoming a practice

An interaction approach to organizational buying behaviour

Transferring, translating, and transforming: an integrative framework for managing knowledge across boundaries

Managing supplier switching

Systematic combining: An abductive approach to case research

Rethinking Marketing: Developing a New Understanding of Markets

Case research in purchasing and supply management: opportunities and challenges

Developing buyer-seller relationships

The relational view: cooperative strategy and sources of interorganizational competitive advantage

Creating and managing a high-performance knowledge-sharing network

Relation-specific capabilities and barriers to knowledge transfers: Creating advantage through network relationships

Building theories from case study research

Theory building from cases: opportunities and challenges

The use of case study method in logistics research

Organizational routines as a source of continuous change

Reconceptualizing organizational routines as a source of flexibility and change

Theorizing practice and practicing theory

Customer response to dissatisfaction: a synthesis of literature and conceptual framework

Managing Business Relationships

Supplier development or supplier switching?

Stability and change in network relationships

Making the most of supplier relationships

The bonding effects of relationship value and switching costs in industrial buyer-seller relationships: an investigation into role differences

The Discovery of Grounded Theory: Strategies for Qualitative Research

Economic action and social structure: the problem of embeddedness

International Marketing and Purchasing of Industrial Goods: An Interaction Approach

Developing Relationships in Business Networks

Process theory of relationship ending

Quality criteria for qualitative inquiries in logistics

Vendor consideration and switching behaviour for buyers in high-technology markets

Testing a life-cycle theory of cooperative interorganizational relationships: movement across stages and performance. Manag

Strategizing: the challenges of a practice perspective

Toward a theory of coordinating: creating coordination mechanisms in practice

Micro-strategy and strategizing: toward an activity-based view

Firm market orientation and salesperson customer orientation: interpersonal and intrapersonal influences on customer service and retention in business-to-business buyer-seller relationships

The myopia of learning

Organizational learning

Relationship equity and switching behaviour in the adoption of new telecommunication services

Qualitative Data Analysis

The Nature of Managerial Work

An Evolutionary Theory of Economic Change

An empirical examination of switching cost investments in business-to-business marketing relationships

Relationship velocity: toward a theory of relationship dynamics

The Awakening Giant. Basil Blackwell

A dynamic model of supplier switching

Never say never -Status quo and research agenda for relationship termination in B2B markets

Developmental processes of interorganizational cooperative relationships

Switching processes in customer relationships

The Practice Turn in Contemporary Theory

Exploring internal stickiness: impediments to the transfer of best practice within the firm

Research on ending exchange relationships: a categorization

Definitions of relationship ending: A literature review

Triggers and patterns of integration initiatives in successful buyer-supplier relationships

Supplier switching decisions

Choice of supplier in embedded markets: relationship and marketing program effects

The work of strategizing and organizing: for the practice perspective

Completing the practice turn in strategy research

The Economic Institutions of Capitalism

Case Study Research: Design and Methods

Interorganizational routines and performance in strategic alliances