Twists and turns of triadic business relationship recovery

Industrial Marketing Management 38 (2009) 618–632

Contents lists available at ScienceDirect

Industrial Marketing Management

Twists and turns of triadic business relationship recovery Anniina Salo ⁎, Jaana Tähtinen, Pauliina Ulkuniemi Department of Marketing, University of Oulu, P.O. Box 4600 FIN-90014, Finland

a r t i c l e

i n f o

Article history: Received 20 November 2008 Received in revised form 6 May 2009 Accepted 7 May 2009 Available online 25 July 2009

a b s t r a c t This study focuses on triadic business relationship recovery processes through a single case study. We address the question of what kind of process takes place when a business relationship on the verge of ending is recovered and what roles a third actor can play in the process. As a result, we model a process through which a triadic business relationship is recovered and attraction, trust, and commitment are restored. © 2009 Elsevier Inc. All rights reserved.

Keywords: Recovery process Process model Abductive approach Triadic relationship Coalitions

1. Introduction Relationships have been studied from several viewpoints for a long period of time. The development of relationships (e.g. Ford, Håkansson & Johanson, 1986), relationship management (e.g. Andersson & Narus, 2004), problem solving and crisis situations in business relationships (e.g. Jehn, 1997; Lam & Chin, 2005), and dissolution of relationships (e.g. Alajoutsijärvi, Möller, & Tähtinen, 2000) have all been topics of interest. The question of why companies should try to save their business relationships from ending has also been an area of study. It has been acknowledged that sometimes companies cannot end even a problematic relationship because of high exit barriers (Tähtinen & Vaaland, 2005). Even though variation of content, quality and intensity of cooperation in business relationships is normal (Ploetner & Ehret, 2006), a relationship may face a situation where its continuation is in danger. Conducting business in a problematic relationship is neither comfortable nor profitable and it is likely to increase both parties' willingness to end the relationship. If this happens, investments made towards building the relationship are lost. Another inconvenient consequence is that the company has to start searching for an alternative business partner. There is also a risk that problems can arise in the new relationship. A vicious cycle of badly managed relationships causes a variety of problems, while continuous searching for alternative business partners is likely to become a burden that overshadows a company's ability to successfully conduct its daily business activities. Moreover, the other actors in the network are likely to be exposed to ⁎ Corresponding author. E-mail addresses: anniina.salo@oulu.fi (A. Salo), jaana.tahtinen@oulu.fi (J. Tähtinen), pauliina.ulkuniemi@oulu.fi (P. Ulkuniemi). 0019-8501/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.indmarman.2009.05.009

their own share of inconvenience due to problems in the relationship (see Alajoutsijärvi et al., 2000). Hence, when a business relationship is struggling with problems, but ending the relationship is either impossible or would cause severe damage, there is only one reasonable solution: attempting to recover the relationship. Therefore, companies have to learn how to recover problematic relationships. Despite of being an issue of managerial importance, relationship recovery is still an under-researched topic and it represents a considerable gap in the literature on problematic business relationships. Thus, research on recovering relationships on the verge of dissolution is needed. Currently, though, only a few attempts (e.g. Paparoidamis, Chumpitaz & Tähtinen, 2007; Tähtinen, Paparoidamis & Chumpitaz, 2007) have been made to uncover the phenomenon, although Tähtinen and Vaaland (2005) argue that a strategy of recovering a relationship has to be different from a maintenance strategy. Moreover, while extensive research on service recovery exists, it is of limited value in the context of business relationships. Business relationships are dynamic entities formed by at least two active parties. The actors are not only responsible for the failures (usually more than one) but also perform the recovery efforts, as well as, the evaluation of the efforts. Service recovery research, instead, focuses on a single actor and a single incident, i.e. the service provider's response (see e.g. Grönroos, 1988) to a service failure (see e.g. Mattila, 2004) and to consumer's subsequent evaluation of the recovery (see e.g. Tax, Brown & Chandrashekaran, 1998). Investigations of multiple service failures (Maxham & Netemeyer, 2002) and customers' participation in co-created service recovery (Dong, Evans & Zou, 2008) are recent but rare additions to service recovery research. Moreover, the few existing studies on business relationship recovery (i.e. Paparoidamis et al., 2007; Tähtinen et al., 2007) focus

A. Salo et al. / Industrial Marketing Management 38 (2009) 618–632

on dyadic relationships and leave the influences that third actors have on the process unstudied. As all dyadic relationships are embedded in a network of actors (Anderson, Håkansson & Johanson, 1994) their influence also exists during the troubled phases of the relationship and deserves to be studied. Moreover, taking a triadic view opens a network view to the phenomenon in two aspects. Firstly, as the smallest possible network, a triadic view allows us to investigate different net flows and effects (see e.g. Smith & Laage-Hellman, 1992, 40). As Caplow (1968, 10) puts it: “the elucidation of social processes within a triad is sufficiently challenging in itself.” Secondly, a triadic approach offers a way to study coalitions and their formation, areas in which investigations in a business context have yet to be conducted. As the sociologist Simmel (1950, 135) shows, a third actor can affect the other two actors, either positively (keeping the triad together) or negatively (disturbing the relationships). There is, thus, a tendency for a triad to segregate into a dyad and an isolated actor (Caplow, 1968, 3; Simmel, 1950, 135; Thibaut & Kelley, 1968, 193). In other words, the triadic relationship has a tendency to dissolve. The aim of this study is to create a model for a triadic relationship recovery process and detail the actions which may save the relationship. The main question that this study answers is what kind of a process is triadic relationship recovery? The research question is further divided into three sub-questions: i) What are the restoring actions through which a relationship can be recovered?, ii) How does the atmosphere change during the relationship recovery process?, and iii) What kind of recovering roles can a third party have? The phenomenon is studied by using an abductive approach (see Peirce, 1998, 226–241) applying an explorative single case study. The aim of the research design is to generate a theory by emphasizing the empirical data. The empirical part of the study is a case of a triadic relationship of supplier–customer–end customer that drifted into problems which almost ended the relationship. Most of the data was collected by retrospective interviews with all actors in the triad. Thus, the abductive approach was applied by first exploring the empirical data and thereafter using theoretical knowledge to better understand the data. Following the way the study was conducted (first empirical data, then theoretical domain), the following sections first present the methodology applied in the study. Next, we shortly describe a narrative of the case of a recovery process of a triadic relationship. In line with our abductive research process, the empirical data and the existing knowledge on relationship recovery are closely entwined in the paper. During the analysis, we build a model of a triadic relationship recovery process, one sub-process at a time. Theoretical and managerial discussions with avenues for future research conclude the study. 2. Research design When starting this research project in 2006, no studies focusing on the recovery process of business relationships were available. Therefore, the study began with a pre-understanding that a dyad on the verge of ending can be restored (Halinen & Tähtinen, 2002; Tähtinen 2002), and that recovery efforts differ from relationship maintenance (Tähtinen & Vaaland, 2005). Tähtinen (2002) proposes that open communication, trust and commitment increase the success of restoring actions. Thus, using the meager research available as a starting point, the researchers became interested in the case and started to collect empirical data. Moreover, the empirical phenomenon, the recovery process, was considered as relatively easy to define and, thus, a manageable phenomenon to study empirically without a strong theoretical framework. In practice, this meant outlining the process and the course of events from the state of a problematic triadic relationship to the state of a recovered one. Another reason for using a research approach in which the empirical knowledge is taken as a starting point was the fact

619

that we were able to identify the particular empirical case. A problematic business relationship had been a common topic in the local newspaper for quite some time and we were also aware of it through personal information channels. Thus, the existence and recovery of the triadic relationship was public knowledge and captured our attention. We characterize our research approach as an abductive one. The preliminary understanding of the research phenomenon, which abductive research requires (see Kovács & Spens, 2005), was acquired through the above-mentioned research on business relationship dissolution and, also, by forming a pre-understanding of the empirical case. We had already followed the events of the case through local newspaper reports, but had neither a priori hypothesis nor any advanced theorizing of the research topic, which is often the case in abductive research (see Levin-Rozalis, 2004). A case study strategy was chosen, since it is an appropriate tool in the early phases of theory development (Eisenhardt, 1989). Moreover, Yin (2003, 39–40, 42) suggests that a single case study is a justified research design if the case is of a revelatory nature and allows the uncovering of a previously inaccessible phenomenon. This is the case in this study, since relationship recovery remained a relatively unknown phenomenon and no process descriptions of triadic relationship recoveries were available. In addition, as Kovács and Spens (2005) report, abduction enables not only the exploration of new phenomenon but also the tying of new results to different contexts. In this research, we started from the empirical data and analyzed it in order to describe the triadic relationship's recovery process. In other words, narrative data and empirical findings are emphasized when building a model which describes the recovery process. A single case study requires careful investigation of a potential case in order to minimize the chances of misinterpretations and to maximize access to allow the collection of case study evidence (Yin 2003, 42). To minimize these risks, multiple sources of data were used, which allowed both source and data triangulation. In addition, it enabled the building of a broader picture of the course of events. The majority of empirical data was gathered retrospectively from eight informants through seven narrative interviews (see Table 1). The interviews were carried out with the service supplier, the customer and the end users. The interviews were audiotaped, verbatim transcribed, and analyzed with the help of nodes arising from the data. In addition to the interviews, newspaper articles addressing the events of the case were used. One of the researchers had intensively followed the case through the local media, especially the events that brought the relationship to the verge of ending. This enabled the researchers to identify the particular relationship for the purposes of this study, while also providing good background knowledge of the case. The newspaper articles were stored and

Table 1 Sources of data.

Interviews

Service provider

Business customer

Key account manager 1, 1 h Key account manager 2, 1 h

CEO 1, 1 h

End users

A heavy user, 1 h 10 min CEO 2, 1 h Normal end users 2 Computer mechanic, and 3, 1 h 20 min 50 min Newspaper articles Article 1: “Problems with university students' internet concerning all parties connections” Article 2: “University students considering legal actions against the housing company because of poor quality of service” Article 3: “Housing company looking for new broadband connection provider” Other Report in regional TV news

620

A. Salo et al. / Industrial Marketing Management 38 (2009) 618–632

analyzed by the same procedure as the interview data. As a result, a multifaceted understanding and thorough analysis of the phenomenon became possible. The analysis of the empirical data was started by first writing down the chronological description of the case. This enabled us to identify the important events in the process. Furthermore, we were able to map how the important events and actions in the recovery process were connected to each other. After piecing together what had happened in the case, we went back to the existing studies that mentioned the possibility of recovery. For the reasons already mentioned, we found the research stream on service recovery to be of limited value. Instead, the stages of business relationship ending (Halinen & Tähtinen, 2002; Tähtinen 2002) and an analysis of the elements that attenuate relationship ending (Tähtinen & Vaaland, 2005) were used as the main guidelines in the preliminary analysis of critical events and actions. This was followed by re-examining the empirical data and re-reading and re-evaluating the existing and emerging literature. As time had passed since the beginning of the study, two contributions in relationship recovery were at this point available, namely Paparoidamis et al. (2007) and Tähtinen et al. (2007). Moreover, we became aware that the role of the end users and the media needed to be understood. This prompted us to use the theory on triads (e.g. Simmel, 1950). Thus, the main analysis was conducted with the help of the emerging theory on dyadic business relationship recovery and the theory on triads. In practice, the coding of events and analysis was conducted manually through underlining and making notes on the printed interview transcripts and newspaper articles. Next, the case of a triadic business relationship recovery is described as a narrative, starting with the three main actors. 3. Description of the case 3.1. Actors The case under study is a triadic relationship that encountered problems which almost destroyed it. The business-to-business relationship takes place between a service provider and its business customer. The triadic relationship also includes the end users, who are customers of the business customer, as Fig. 1 shows. The service provider (SP) is a telecommunications company offering broadband and data transfer services. The business customer (BC) is a foundation offering housing services for students. The SP supplies broadband connections for the BC's apartments. The end users consist of over 3000 university students who live in the BC's apartments and for whom a properly working internet connection is highly important. 3.2. The story of the problematic relationship The SP was chosen to replace the BC's existing broadband service supplier through competitive bidding. The negotiations, network

Fig. 1. The triadic relationship.

building, and connection testing all proceeded smoothly. The problems started when the internet connections were transferred into the newly built SP's network. There were breaks, the service was slower than expected, and it collapsed easily. Thus, the end users started complaining heavily to the housing foundation (i.e. BC), since the internet connection was part of the rent they paid. The BC informed the SP of the huge amount of complaints, but despite the SP's efforts to solve them, the problems continued, and so did the students' complaints. Especially the heavy users (i.e. engineering students), who were knowledgeable about data transferring and broadband technology, demanded high-quality service. Thus, the BC confronted the SP and demanded that they start providing the service promised in the contract. As the situation continued, some of the BC's employees advised the end users to complain directly to the SP, although this was against the contract. However, the SP did not wish to have anything to do with the end users, because for them, the BC was the customer, not the end users. The angry end users contacted the media, which meant that another third party came along, influencing the business relationship (see Fig. 2). This was extremely inconvenient for both the SP and the BC and both were willing to end the relationship, even though the contract period was not even close to termination. Dissatisfaction with the business relationship was high in the whole triad. However, being in the public eye, the companies felt that a failure of the relationship would severely harm their reputation and business. Communication between the SP and the BC was very difficult, but when the contact persons (i.e. KAM and CEO) on both sides changed, the communication problems eased. The BC's new CEO made the service quality of the broadband connection their main concern and started searching for an alternative supplier. The SP was not happy, but did not really have a say in the matter. As a result of the competitive bid, the BC tested an alternative supplier's broadband service. However, the end users rated it worse than the current service and, therefore, it was not an alternative. Hence, the SP started to trust the BC's willingness to continue the relationship. The BC organized the first end user satisfaction survey, which helped both parties to understand the end users' needs better. As a result of intensified discussions, both companies understood that investments in new technological solutions would be needed if the service quality were to be increased. These investments considerably helped to improve the service quality, resulting in more satisfied end users, which further eased the tension in the business relationship. There were still some practical problems, but since the communication procedures had been rebuilt and both companies understood each other better, the problems were solved. Finally, the triadic relationship reached a stage where all parties perceived it to be working well. 4. Theoretical analysis: building a model of a triadic relationship recovery process In this chapter, we build a model of the actions, atmosphere, and actors of the triadic relationship recovery process, starting from the

Fig. 2. Third parties influencing the business relationship.

A. Salo et al. / Industrial Marketing Management 38 (2009) 618–632

definitions. Since studies on relationship recovery are rare, the terminology is only emerging. In this research, a triadic business relationship that is heading towards dissolution is referred to as a problematic relationship, and the process that aims at saving it is called a recovery process. For the purpose of this study, we modify the definition of Tähtinen, et al. (2007) and define a recovery process as follows: all the actions the actors undertake to change a problematic relationship into a beneficial one. Moreover, we apply the concept of sub-process, and define it as connected actions. Thus, the recovery process consists of sub-processes which consist of connected actions (see Holmlund, 2004). In a well-functioning business relationship, there is mutual dependency, trust and commitment, which are the cornerstones of a partnership (Harris, O'Malley & Patterson, 2003). Moreover, they are also the key elements of the business relationship's recovery process and must be restored in order to normalize and recover a problematic relationship. We separate the actions to recover a problematic business relationship from the elements of the relationship's atmosphere (attraction, trust, and commitment). The reason for the two-level analysis is the two-way relationship between the recovery actions and the atmosphere. An action can be perceived by the counterpart(s) as a sign of, for example, a level of trust. Likewise, any change in the atmosphere influences the future behavior of the parties in the relationship. Thus, the action and atmosphere levels are closely entwined with each other. In addition, we also use a triadic view on the relationship recovery process. The triadic analysis reveals how the actors' positions and roles changed during the recovery process. In the following discussion, the visual presentation of the model is done step by step, and the end result comprises all the sub-processes and their various connections. The discussion follows the abductive manner in which the study actually proceeded. This means that we will present the existing theoretical knowledge that was applied during the analysis concurrently with the results of the empirical analysis. Thus, the analysis of a triadic business relationship recovery process merges the empirical data of this study with its theoretical sources: firstly, relationship ending studies already discussed, secondly, a dyadic relationship recovery model (Tähtinen et al., 2007), and thirdly, research on coalition formation in triads. However, since our process model is based on the notion that a process is non-linear, the analysis does not form a narrative that the reader could follow from the first events to the last. In other words, the recovery process consists of sub-processes that can take place partly simultaneously and the process may proceed to a sub-process more than once. 4.1. Voicing Voicing may play an important role in the maintenance of business-to-business relationships (Ping, 1997). Companies have the possibility to use voicing when experiencing problems in a relationship, even though they do not always use this potential (Ping, 1997; see also Ping, 1993; Hirschman, 1970; Rusbult, Zembrodt & Gunn, 1982). According to Hirschman (1970, 33), voicing is the only option in cases where satisfaction in the relationship is low, but the customer is not really able to exit the relationship. When the number of alternative suppliers is low, companies have higher exit barriers (Tähtinen & Vaaland, 2005). Even in such situations, however, the active options of voicing and/or exit and the passive option of neglect are still available. We have taken voicing as a sub-process in our relationship recovery process model, based on the only process model of a dyadic relationship recovery that we found and applied during the data analysis, namely Tähtinen et al. (2007). The model divides the recovery process into three sub-processes, voicing/blackmail, attenu-

621

ating analyzing, and normalizing. In the first sub-process, voicing refers to the requests to change something in the relationship, whereas blackmail refers to threatening the other party to perform the desired changes. In the triadic relationship analyzed in this study, the data did not reveal any actions that would refer to blackmail, and, therefore, the sub-process is labeled here as voicing. In the analyzed triadic relationship, the actors first performing voicing were, for the most part, third parties. The end users were the ones that first complained to the BC which was, from their point of view, supplying the housing service including a broadband connection. The BC informed the SP of poor service, but the SP didn't manage to take actions that would have resulted in improved service quality. The number of complaints was enormous and the BC had to work extremely hard in order to respond to all the complaints. This was a critical period, since, for some of the students, a fast and trustworthy broadband connection was the main reason they lived in the BC's apartments. Since the broadband connection was included in the rent, some students threatened to pay only part of the rent due to the nonfunctioning internet connection. On top of everything else, students started to question whether they should move out somewhere where they would have working network connections. Another thing causing friction between the SP and the business customer was that some of the business customer's employees advised the end users to complain directly to the SP. Frustrated with end users complaints, the BC hoped that if the employees at the SP were to receive feedback directly from the users, they would better understand the problems of poor service quality. However, this procedure was against the agreement, since it had been agreed that, in case of service failures, the end users contact the BC that will handle its own half of the network before contacting the SP. This accelerated the decline of atmosphere and increased tension in the whole triad. Voicing has a role in alerting a firm or organization of its failings, but this alerting role is only achieved properly if the management has enough time to respond to the voicing (Hirschman, 1970, 33). In addition, this case shows that the management also has to have a will and means to respond to the voicing. At this point, the end users and the BC used voicing as an alarm of bad service quality, but it did not result in any improvements. Next, let us open up the triadic view on the voicing sub-process. It is distinctive to the dynamics of a triad that a third actor may play different roles within the triad. Simmel (1950, 145–169) distinguishes three functions of the third party: mediator, tertius gaudens, and oppressor. At different times, any of the triad's members may fulfill one of these functions and be a third actor. In other words, we viewed the triadic relationship as consisting of three actors instead of a priori defining it as a dyad (in other words a coalition) and a third actor. Moreover, this view also enabled us to analyze all the network actors as third actors, as shown in the following. According to Simmel (1950), a mediator can keep the triad net together in spite of events precipitating triad ending that may occur in the triad or network context. Moreover, the mediator is involved with both “sides” and may defend the unity of a triad against the particular interests of any single actor (Caplow, 1968). The positive effect of the mediator may, therefore, engage the actors in a restoration process. The negative effects of a third actor can segregate the triad (to a coalition and a single actor) if the actor is more interested in satisfying its own interests than keeping the triad together (Caplow, 1968; Simmel, 1950). Tertius gaudens utilizes situations where the other two actors hold each other in check. It fulfills its own interests by supporting one of the other actors and forming a coalition with it. Thus, the tertius gaudens does not actively seek a coalition, whereas an oppressor intentionally produces friction between the other two actors in order to gain a dominant position. The incentive to act as a tertius gaudens or an oppressor depends on the connections of the three relationships, i.e. whether they are positively or negatively connected. In triads where all the relationships

622

A. Salo et al. / Industrial Marketing Management 38 (2009) 618–632

are positively connected, as they were in this case, there is no need to play the role of tertius gaudens or oppressor unless a more attractive third party is available. An oppressor may deliberately want to switch the third member in a triad if the new triad would produce better results. However, success in the short run may not equal success in the long run. For an actor choosing to act as an oppressor, it is important not to create a strongly negative image in the eyes of other network actors. The victim of an oppressor is bound to suffer and may communicate negative aftermath stories to the network actors. As nets and networks are constantly changing, a company never knows whether it will need to cooperate directly or indirectly with the current adversary company sometime in the future. Therefore, the acts of an oppressor may produce successful long-term results only in exceptional situations. We found in our triadic analysis that coalitions can be internal or external. Internal coalitions take place within the triad, and external coalitions when an actor joins forces with a fourth actor, outside the triad, in order to influence the other two actors within the triad. In our case, both internal and external coalitions took place, internal being the first, as the continuing case description shows. At this point, the BC had a much better understanding of the problems in service quality than the SP. Being overloaded with complaints, the BC had the same goal as the end users, to get through to the SP and demand improvements. All this resulted in the end users and the BC forming an internal coalition as illustrated in Fig. 3. We label the coalition as internal, since it involved two of the original members in the triadic relationship. In spite of the complaints, the SP still refused to acknowledge their relationship with the end users. Thus, the internal voicing coalition was unsuccessful. “We tried to arrange a meeting with the service provider, but they said that the housing company is the customer, not us. We felt that they didn't understand the situation at all.” End User

Thereafter, the end customers decided to seek leverage in order to get their message through. They turned to another third party, the media, which voiced their concerns to not only the two companies, but also to the other actors in the network. Newspapers and a local television station publicized the story and announced it in the regional news, causing bad publicity for both the BC and the SP. The voicing backed up by leverage from the media was finally powerful enough to arouse the attention of the SP. The SP wanted to get rid of the bad

Fig. 3. Voicing by internal coalition.

publicity and started to pay attention to the situation. Still, the atmosphere within the relationship was too inflamed for any real recovery efforts to take place. ”After a while and some newspaper articles, we finally managed to organize a meeting. The BC helped us and lured the SP to participate. And we invited the representative of the old broadband service provider, too. It ended pretty badly and we [the end users] and the BC both accused the SP of bad service quality.” End User Thus, after the failure of the internal coalition (the BC and end users) the end users used leverage voicing, contacted a network actor, and formed an external coalition as illustrated in Fig. 4. Both the internal and external voicing coalitions created pressure that finally forced the SP to reconsider the situation. In addition, because of the publicity, ending the relationship would have caused serious damage to the SP's reputation, which made them less willing to end the relationship. However, no recovery actions were performed at this point. Although voicing was initiated by a third actor, it also has to be noted that during the rest of the restoration process, the BC needed to continue voicing towards the SP. This was because the root causes were not found and, therefore, no resolution to the end users complaints was offered. Actually, the root causes for the service failures were hidden in the interface of the SP's and BC's networks, which neither company felt responsible for. In other words, neither the SP nor the BC detected problems in their own network and, consequently, both perceived themselves to be acting according to their contract, while the end users suffered from bad service quality. The BC had no other choice but to continue voicing the SP. “Even though the service provider had technical expertise, it didn't have the skills to look at the situation from the end users' perspective. For us this was the most important thing and we noticed that we have to emphasize the importance of the end users' satisfaction to the service provider.” Business Customer In Fig. 5, we present the first part of our model of the recovery process of a triadic business relationship, namely the sub-process of voicing.

Fig. 4. Voicing by internal and external coalitions.

A. Salo et al. / Industrial Marketing Management 38 (2009) 618–632

623

“We thought that if we could solve these problems with the SP, a lot of harm would be avoided. It would be easier to continue the relationship than switch suppliers, because with another operator it might take another two years to make things work. Even though the experience had left a bad taste in our mouths we thought that if we could manage to solve these problems, we would be better off.” Business Customer

Fig. 5. A sub-process of voicing within recovery process.

4.2. Analyzing the triadic relationship and its role in rebuilding attraction In a troubled relationship, attraction, trust and commitment are low. Therefore, the affected parties must either face the prospect of ending their relationship or initiate recovery processes aimed at rebuilding all three relationships. In the case under study, high exit barriers (Ping, 1993, 1994, 1999) acted as factors attenuating relationship ending (see Tähtinen & Vaaland, 2005) that decreased the parties' willingness to end their relationship and, therefore, increased their willingness to restore it. Factors attenuating relationship ending are categorized into relational investments, dissolution process costs, possible sanctions for future business, network limitations, and set-up costs (Tähtinen & Vaaland, 2005). Tähtinen et al. (2007) suggest that in a sub-process called “attenuating analysis”, the actors evaluate the factors and events that attenuate relationship ending and decide how to continue, whether they will try to recover or not. We chose to label this sub-process “Analyzing the triadic relationship” to highlight the three actor relationship. As a result of the external and internal coalitions' voicing, the SP felt that their reputation as a broadband operator was on the line (i.e. sanctions for future business) and that their capabilities in a wider perspective were being questioned, as the following quotation reveals. “We were seriously considering ending the relationship, but we are a local operator and the BC is also local, it is against our values to abandon a local customer, or give it to a competitor. Another reason was that we simply didn't want to give up, and we just decided to grit our teeth and not back down.” Service Provider The BC, on the other hand, started to perceive high exit barriers. Changing the supplier would have required changes both in the existing hardware and software, not to mention guiding and informing both their own employees and the end users. All 3000 end users would have had to change their computer settings, which would have required setting up a help desk. Personnel and other resources would have been needed and the whole process would have been very time consuming. Even though the situation in the relationship was intolerable from the BC's viewpoint, high exit barriers reduced the company's willingness to end the relationship.

High exit barriers and a fear of sanctions in future business operations, as well as network limitations, acted as attenuating factors. Not only did they encourage all the members of the triad to act to recover the relationship, but also enabled attraction to appear within the relationship. The sub-process of analyzing the relationship within a recovery process, highlighting the role of attraction, is presented in Fig. 6. Trust and commitment are recognized as essential ingredients of a beneficial long-term business relationships (Garbarino & Johnson, 1999) These, along with attraction, are the three elements of a relationship's atmosphere that need to be restored during the relationship recovery process. Attraction influences the development and maintenance of exchange relationships (Halinen, 1997; Harris et al., 2003). Moreover, attraction precedes trust and commitment, and is therefore discussed next, starting from how recovery subprocesses influence the atmosphere. Halinen (1997, 59) defines attraction as “a company's interest in exchange with another, based on the economic and social reward–cost outcomes expected from the relationship over time.” In this case, factors attenuating relationship ending seemed to be the main reason for the increase in attraction. In addition, existing research confirms that attraction has a motivating role in relationship development (Harris et al., 2003). The case shows that attraction has a significant role also in the relationship recovery process, because it restores the interest to remain in the exchange with the partner. However, this interest is based on future value, since the value of the relationship has not met expectations in the past. If the reason for the problematic relationship lies in service failures, as was the case here, the customer begins to question the supplier's capability to provide quality service. This means that unless the customer changes its perception of the supplier, relationship recovery is unlikely to occur (unless the ending of the relationship is not an option). In other words, in order to be willing to recover the relationship, the customer needs to be convinced that the supplier has potential as a supplier to provide future value. The supplier, on the other hand, needs to be convinced that investing in a recovery process is worthwhile and that the buyer is also willing to recover the relationship. In order to increase the trust and commitment required to recover the relationship, it is, therefore, highly important that both parties show signs of attraction towards each other ”Many negotiations were needed before we started to believe that mutual willingness existed.” Service Provider In this case, the recovery process would have been impossible without the knowledge that the BC had gained when it tested an alternative supplier. We consider this as an external coalition between the BC and the Alternative SP that was meant to help the ending of the relationship with the original SP, but, instead, turned out to provide a reason to restore it. The tests with the Alternative SP revealed that their service was worse than the current one. Thus, the BC started to trust the SP's technical skills and, on the other hand, the lack of an alternative increased the BC's willingness to recover the relationship. In addition, the SP became convinced that the BC was ready to commit to a recovery process. This not only increased their mutual attraction but also was the first step on the way of regaining trust and commitment.

624

A. Salo et al. / Industrial Marketing Management 38 (2009) 618–632

Fig. 6. A sub-process of analyzing the relationship within a recovery process.

“We tested another broadband operator's service in one of our apartment buildings and asked the end users living there to give us feedback. Even though the SP's broadband connection was rated bad, the one we tested received even worse feedback from the end users.” Business Customer “It became very clear that even though we had some problems, fundamentally our service is of a good technical quality and we perform well in a competitive situation. The bid proved to us and to the BC what we are made of and what our technical skills are.” Service Provider The competitive bidding convinced the BC, the SP and the end users that their relationship had potential to provide future value and it was worth recovering. Thus, the external coalition of the BC and the Alternative SP positively influenced the recovery process as illustrated in Fig. 7.

is reliable and an expert in its area and, at least, an intention to rely on the partner. Inspired by literature on social psychology, Doney and Cannon (1997) highlight the basic contingency of trust, which is the existence of vulnerability between parties, since without vulnerability, trust would not become operational. In other words, in situations where one party or some parties may exercise power, uncertainty is diminished and, thus, trusting the opposite party is no longer needed. Moorman et al. (1993) also indicate that trust results more from a relationship between partners than from the individual factors (in this case from the managers and end users). All this can also be looked at from the angle of missing trust. Logically, when trust is missing, the actor(s) try to increase their power in order to be able to continue their relationship without being too vulnerable. This is also the case here, when coalitions were used to increase power in a situation of low trust. In this case, trust has an important meaning on at least two levels. Firstly, the BC's trust towards the SP's capabilities was very low.

4.3. Normalizing and its role in rebuilding attraction, trust and commitment Tähtinen et al. (2007) describe a normalizing sub-process as referring to actions to make the requested changes to recover the relationship. In reality, collaboration without a minimum level of trust is impossible, because the risk of potential opportunism is never totally absent (Ploetner & Ehret, 2006). According to Ganesan (1994), trust is a necessary element for parties to have a long-term orientation. Thus, trust building is not only an essential antecedent to the development of partnerships, but rebuilding it plays an important role in the recovery process of business relationships. “The most difficult thing in the beginning was to learn to trust that we were both looking for a win–win situation, instead of just maximizing benefits in the short term.” Service Provider Trust is defined as “a willingness to rely on an exchange partner in whom one has confidence” (Moorman, Zaltman & Deshpandé, 1992). This definition aims to cover both the cognitive and the behavioral aspects of trust. Therefore, as supported by Moorman, Deshpandé and Zaltman (1993), when trust exists, it includes a belief that the partner

Fig. 7. External coalition increasing attraction.

A. Salo et al. / Industrial Marketing Management 38 (2009) 618–632

625

Secondly, both counterparts had no trust in each other's willingness to stay in the relationship. Testing the SP's skills by competitive bidding increased trust between the parties, not only because it proved that the SP is capable of providing the service, but also because it became clear that changing the supplier would have been very laborious for the BC. “We had been arguing about quality for some time, but at this point, we defined clearly what we wanted. It is a little bit funny, but the competitive bid we arranged helped us to define what service quality we really wanted to have in practice. And the supplier did the same. It was tough but we both paid an effort. And it helped a lot in clarifying the situation.” Business Customer According to Morgan and Hunt (1994), trust exists “when one party has confidence in an exchange partner's reliability and integrity.” When the counterparts perceive each other to be capable and willing to stay in the relationship, trust building begins. After that, the discussion to form a concerted goal may follow. This task may be difficult, since the communication links may have suffered from longrunning problems and need to be restored. In order to agree on a concerted goal, companies are forced to improve their dialogue. Being able to listen to each other and describe their business environments to their counterpart helps them in this process. “In the beginning it was difficult to understand all the technical talk. We are professionals in the housing business, but we don't have a lot of expertise in broadband technology. But after a while we learnt, and we began to understand the operator and we were able to participate in the technological conversations.” Business Customer As a result of increased communication and a willingness to understand each other the companies began, little by little, to understand each other's business environments and problems. In this case, the SP had refused to see the end users as customers, but eventually, a willingness to open communications and familiarizing with the BC's business environment opened the SP's eyes, and it also started perceiving the whole triad. This “user-focused” insight was new for the SP, who had only just now begun to understand the situation from the end users' point of view. This allowed the companies to find common interests and, after a while, they were able to agree upon a concerted goal. “The beginning was very difficult and there didn't seem to be a light at the end of the tunnel. Meetings were hard and it felt like we were stuck in a rut. But once we found a common interest the situation changed completely.” Business Customer “We defined a certain service quality level and most, but not all, end users were happy with this. Some [heavy users] would have wanted more but we had to set the service quality level somewhere and think about the majority of the end users.” Service Provider In other words, the companies in the relationship started negotiating about service level and defined it clearly in order to set a concerted goal. From this point on, the SP and the BC were able to form a new internal coalition. The members in the new coalition both explained, especially to the heavy end users, why it was important to make an effort to recover their relationship, rather than end it. Fig. 8 illustrates this internal coalition. Both the SP and the BC had heard the end users' voice before, and they didn't let the end users interfere with setting the service level. On the other hand, at this point, they both had good information of the end users' service requirements and managed to set the service at a

Fig. 8. Formation of internal coalition (SP and BC).

level that pleased the majority of end users. This was very important and helped to improve the atmosphere within the entire triad. Although finding a concerted goal may be almost impossible, it is, nevertheless, a prerequisite for continuing the recovery process. If companies are unable to find a concerted goal, there is little sense in recovering the relationship. “The end user's service quality was taken as a beacon and it helped us to change our view point from simply arguing and blaming each other, to improving the service quality. This had a great influence on the atmosphere within the relationship as the attention was switched from technical solutions to improving service quality.” Service Provider Finding a concerted goal was crucial in the recovery process, and the means through which the concerted goal is reached could be negotiated. Because problems in the relationship could not be solved overnight, a need remained for clear procedures for acting in difficult situations. “We agreed upon the service quality, how it is monitored and what the sanctions should be if the quality level is not achieved. If problems occur, we'll fix them. As a result the whole bearing of the relationship improved.” Service Provider A commitment toward achieving a mutual goal has a major role in trust building. However, this may take a long time, since this is the stage where solutions to problems are being looked for on a concrete level. It is also possible that the whole organization needs to be involved in the process, at least on some level. Metcalf, Frear, and Krishnan (1992) show that an increase in information and social exchange also increases co-operation and adaptation in a business relationship. Since co-operation and adaptation are crucial in a successful recovery process, both information and social exchange have an important role also in a recovering business relationship. During the recovery process, information exchange needs to be continuous, and it is required at all organizational levels, from senior management to the blue-collar level. Time and effort is needed to build communication routines both within companies and between them. Moreover, communication to other actors in the triad is necessary in order to inform, or even involve, the third actor in the recovery process. Information exchange within a single organization is also important, so that the overall understanding of the recovery

626

A. Salo et al. / Industrial Marketing Management 38 (2009) 618–632

process remains clear, and all employees are able to strive for a common goal. “It was the management's task to inform the whole personnel that the priority at the moment is to solve the problems of the (broadband) service quality.” Business Customer ”Our employees got to know the BC's employees and co-operation strengthened, things started to work better.” Service Provider As a result of improved communicating skills and routines, people from different organizations learn to know each other and communication becomes more efficient. Knowledge of their partner's business environment increases within both organizations and problem solving becomes easier. “We had to discuss things at a grassroots level. I mean, we talked about individual servers trying to figure out why it wasn't working. Why connections in this house were bad and would new cables be required?” Business Customer “We had countless meetings and if we had thought of the situation only from a financial viewpoint, we probably would have given up.” Service Provider The actions that companies perform in order to build trust also play a significant role. The BC invited the SP to visit and test their half of the network, which is unusual in this kind of business and required a lot of trust. “The SP was invited to visit us and we showed them our part of the network. It was then that they realized that everything was not going as planned and that the end users' broadband connections were practically useless. So they realized that things really were in bad shape, even though everything seemed to be fine from their end.” Business Customer A mutual willingness towards conducting business together became higher than before. The SP and BC met frequently and, little by little, trust begun to exist in the relationship. At the same time, the atmosphere of the relationship began to improve. However, there were still several uncertainties that required considerable attention and many problems had to be solved on a grassroots level. Thus, the companies spent a lot of time and effort in communicating with each other, which helped them to acknowledge and understand each other's business environments. Interfirm adaptations, when reciprocal, are demonstrations of commitment and trust (Hallén, Johanson & Seyed-Mohamed, 1991). The first signs of interfirm adaptations could be seen after a while, when the SP allowed the BC to have a look at its broadband traffic. This was a very important step, and made a considerable difference by allowing for a new approach to the situation. “I'd say that it is highly unusual for an operator to allow a customer to view their network. It requires a lot of trust. In the beginning we couldn't even think of anything like this happening. But our trust towards our partner had increased and we thought that this new step could be taken, and to tell the truth, this has been a good solution for both of us.” Service Provider At this point of the recovery process, the first signs of commitment can be seen which, for their part, prodded the companies to continue the recovery process. Commitment is a signal of the desire to develop a stable relationship and the willingness to make sacrifices in the short

term in order to benefit in the long term (Anderson & Weitz, 1992). In addition, both information and social exchange were needed, because two-way communication raises the commitment levels of both parties of the business relationship (Anderson & Weitz, 1992). Even though more concentration was required, finding solutions for the real problems became possible and active information exchange accelerated the recovery process. Understanding the business partner's viewpoint enabled the SP and the BC to plan their activities so that the service recovery became possible. At this point of the process, both had an understanding of each other's business activities and the surrounding networks. These broader insights into the business environment are beneficial, not only during the recovery process, but also in the future of the relationship. Furthermore, when there is commitment in the relationship, the parties are also confident in the stability of the relationship (Anderson & Weitz, 1992). Morgan and Hunt (1994) define relationship commitment as “believing that an ongoing relationship … is so important as to warrant maximum efforts at maintaining it; that is, the committed party believes the relationship is worth working on to ensure that it endures indefinitely.” In short, commitment can be defined as an intention to continue the relationship (see Geyskens, Steenkamp, Scheer & Kumar, 1996). In addition, Garbarino and Johnson (1999) have shown that future intentions strengthen the relationship. If the commitment to the relationship is low, the customer's goal is an overall satisfaction, but if the commitment to the relationship is high, it is the trust, along with commitment, that influences future intentions. The development and improvement of future business opportunities motivates commitment in the relationship. In other words, investments in the relationship increase commitment. On the other hand, as soon as specific investments are translated into value, they are at risk of being lost if the relationship is terminated (Ploetner & Ehret, 2006). This means that the counterparts have to share at least some visions of future co-operation before commitment (in the form of investments) is shown. In this case, both the SP and BC committed in a concerted goal (the service quality of the end users), which improved the vision of the relationship's future and enabled investing in the relationship, creating a positive spin. Finding a concerted goal, to which both counterparts are mutually committed, allows trust to build within a relationship. Trust is also a prerequisite for the emergence of commitment and partnerspecific investments, which for their part further increase trust (Ganesan, 1994). Nevertheless, mutual commitment among exchange partners is a central factor in a successful relationship and produces significant benefits for companies (Geyskens et al., 1996). However, it cannot be achieved without trust and shared views about the future of the relationship. As Håkansson and Snehota (1995, 198) state: “Trust is a necessary condition for commitment and commitment only makes sense if tomorrow matters.” As a result of increased commitment and trust gained by intangible investments (time, negotiations, improvement of communication skills and procedures), the companies acquired enough courage to also make tangible investments that, in this case, improved the service quality. These investments were highly important, since they enabled achieving the mutual goal and also strongly influenced the relationship's functionality. In this case, the end users' satisfaction increased significantly once the investments in the network interface were made. This decreased the number of complaints from the end users, which clearly eased the pressure on the relationship and enabled open discussion. In other words, the companies were able to focus even more on the recovery process and on developing better service for gaining mutual benefits. Empirical findings show that improved service quality also plays a significant role in avoiding future problems. “With the service provider also using the same control system now, we can easily see what is going on in the network and how things are working. It's great; it works so much better for us! This

A. Salo et al. / Industrial Marketing Management 38 (2009) 618–632

is a remarkable change and required a lot of effort from the service provider.” Business Customer In Fig. 9, we present the third sub-process of the recovery process of a triadic business relationship, focusing on the normalizing of the relationship.

4.4. Aftermath and the end result of the process Even though it is hardly desirable that a business relationship should fall into trouble, it may have positive results if the problems are solved and the relationship recovery process succeeds. For example, the parties may realize that their partner is the best available one. On the other hand, they may find out the opposite, but even then the companies may, through attempting to recover their relationship, learn new skills that may be beneficial in other relationships. “We have made a serious effort to get to know each other and to make our co-operation work. We have learnt a lot and I feel that we have also gained important knowledge concerning different ways of co-operation. We have already been able to use these new skills with other customers in other business relationships.” Service Provider In this case, the parties' mutual willingness to amend the relationship and persistent recovery actions eventually resulted in their continued co-operation. Recovering the relationship was both expensive and hard, and required a lot of time, monetary investments, and effort at the relationship management level. Finally, when everything was running smoothly, the BC was very satisfied and felt that they made the right choice by giving the SP time to improve the quality of the service. Therefore, the result of the recovery process resulted in a beneficial and easily maintainable business relationship with satisfied end users. In other words, the relationship recovery was beneficial for the whole triad.

627

“At the moment we have meetings regularly and the atmosphere is very pleasant. We each clearly understand our responsibilities and we don't have to argue over trifles and details.” Business Customer “I remember how we used to think that the quality of the broadband service was horrible and that anything would be better than the one we had. But now I have to admit that we are satisfied users and the broadband connections are trustworthy and very competent.” End User As a result of the recovery process, information and communication channels were working and there was a mutual understanding of responsibilities. Because there was trust in the relationship, even small problems could be openly discussed, preventing them from becoming bigger issues. Fig. 10 presents the fourth sub-process of the recovery process of a triadic business relationship.

5. Theoretical discussion In this study, we explored a triadic relationship recovery process through the process model which consists of recovery actions, the elements of the relationship atmosphere, and the role of the actors during the recovery process. Our research strategy was to explore a case of the triadic relationship of a supplier, business customer and end customer, and relate insights from the empirical data to the scarce existing knowledge on relationship recovery and the theory of coalition formation in triads. The process model built in this study conceptualizes the atmosphere of the relationship through the changes in attraction, trust, and commitment. The way the counterparts of the relationship perceive these three elements influences the course of the relationship. A problematic relationship starts to recover only if actions towards recovery are taken. The recovery actions influence the atmosphere of

Fig. 9. A recovery process with normalizing and atmosphere.

628

A. Salo et al. / Industrial Marketing Management 38 (2009) 618–632

Fig. 10. A recovery process with aftermath and the outcome.

the relationship, because they are perceived as signs of the actor's perception of the relationship, for example, whether the actor itself is committed to or trusts the counterpart. In addition, if the actors trust each other, they can invest in the relationship which, in turn, positively influences the atmosphere. Thus, both levels, the atmosphere and the recovery actions, influence each other and are closely intertwined. The model identifies three sub-processes in the recovery process: voicing the concerns, analyzing the relationship and making decisions about the willingness to continue the relationship, and normalizing the relationship. The voicing refers to actions that bring the relationship's problems and the requests of solving them into the awareness of all parties. During the analysis, each party assesses whether the relationship is worth saving or not and determines the possibilities of recovering or ending the relationship. Unless a will to recover the relationship arises at this stage, the recovery process may end here and a dissolution process may begin. The normalizing occurs when the will is proven to the counterparts by normalizing actions, such as, by agreeing on a mutual goal and investing resources into achieving the goal and recovering the relationship. Compared to the model in Tähtinen et al. (2007), in addition to focusing on triadic relationship recovery, this study took a more detailed look at what goes on in the sub-processes; what influences the atmosphere, and who are the actors performing the subprocesses. However, the dyadic recovery model (Tähtinen et al. 2007) supported the analysis, although the empirical data of this study enabled a more detailed and triadic elaboration. Thus, in summary, Fig. 11 presents not only the activities that unfold within the process, but also the dynamics that exists between the process and the structure, i.e. the activities and the elements of the relationship atmosphere: attraction, trust, and commitment. By combining these two views into a single figure, we highlight the inseparability of structure and process, which naturally applies to any phase in a relationship's life, and thus draw on the levels-ofinterdependence model of Thibaut and Kelley (1968) and Kelley (1979). Above all, we draw on the idea of interaction producing concrete (e.g. resource ties) and abstract (e.g. the atmosphere) outcomes which, once formed, influence the interdependence. There-

fore, the influence between the sub-processes and the elements of the atmosphere are illustrated with two-way arrows. Empirical findings show that the need for effective information exchange, between and within the parties, remains high during the whole process, as also does the need for intangible investments, such as spending time on meetings and negotiations. It is obvious that recovering a problematic relationship requires mutual effort and understanding of a partner's business. The less understanding companies have of each other's business and business environment (such as typical characteristics of the industry and the value network of the partner), the more they will need to communicate. In addition, the quality of information needs to be emphasized when the understanding of the partner is low. The quality of communication refers to its richness and constructiveness. Communication has an important role in the process, especially because a lack of it can cause problems. However, these actions are also important at the beginning of the recovery process, since the process will not advance without them. Recovering a problematical business relationship differs from building a new business relationship, even though there are some similarities. In the first situation, the partners have more knowledge of each other, or, at least they think that they are familiar with the partner. The latter kind of thinking weakens the quality of communication, since discussions may lack details that are presumed to be familiar to the partner. Thus, the information may not travel very efficiently, it can be inadequate and, in a way, shallow. There may also be arguments or “bad blood” between people, which is likely to influence the relationship, as well as its recovery process. In addition, the companies have knowledge and hands-on experience of what kind of troubles certain actions may cause. Unless communication is handled with care, there is no possibility of recovering the relationship. Particularly, communication skills and the commitment of boundary spanners are crucial. In order to recover the relationship, recovery actions have to take place and factors hindering the relationship's recovery have to be diminished. In addition, empirical findings show that the causes for the problems need to be detected and solved before the recovery actions can be successful. As a result of a relationship recovery process, the

A. Salo et al. / Industrial Marketing Management 38 (2009) 618–632

629

Fig. 11. Recovery process of a triadic business relationship.

relationship can regain its status as beneficial and can even achieve better functionality than before. As the parties familiarize themselves with each other during the recovery process and build new communication skills, improved efficiency in daily business routines becomes possible. In addition, it is possible that new business skills learnt during the recovery process may be used in improving the companies' other relationships. Tähtinen and Vaaland (2005) point out that factors and events attenuating the reasons to end a relationship play an important role when companies are considering recovering their business relationship. The study at hand shows that the elements of the relationship atmosphere, e.g. attraction between the actors, play a major role in influencing the actions as well as the actors' perceptions of the actions. This influence is emphasized especially in the beginning of the recovery process. According to Tähtinen and Vaaland (2005), the purpose of the restoring actions is to learn to reduce the tension in the relationship. This also applies to the triad relationship recovery process. Therefore, we can claim that the purpose of recovery actions is to improve the atmosphere of the relationship in order to find and solve the difficulties that are causing problems. If this is achieved, new problems can be dealt with before they pose a serious threat to the continuation of the relationship. The role of a third actor in a relationship recovery process was clear and important. Based on the case, any one of the three actors can be the one performing voicing, and, therefore, bringing the problems to the surface and starting the recovery process. Hirschman (1970, 34) claims that the role of voicing is likely to increase as the opportunities of exiting the relationship decline. This case shows that when the

opportunities of recovering the business relationship increase, voicing is no longer needed and beneficial co-operation is allowed to develop. In addition, in all the three sub-processes, all actors were active and performed recovery actions. A major role in the triadic relationship recovery process proved to be coalition formation. Both internal and external coalitions were formed and they influenced the conduct of the recovery process. Internal coalitions refer to situations where two parties within the triadic relationship form a united front in order to influence the remaining actor. The incentive of the coalition formation is the resulting increase in power; if the two actors are more powerful together than the remaining actor, their coalition can influence its actions. External coalitions, on the other hand, are those where a network actor, i.e. an actor outside the triad, is brought in to form a coalition. The purpose is the same, to increase the power of one actor (or two actors) in order to influence either or both remaining actors in the triad. This categorization of internal and external coalitions emerged from the case study and, thus, contributes not only to the theoretical discussion of triadic relationship recovery, but also more generally to the triad theory. Moreover, many of the coalitions taking place in the case were conceptualized as unifying coalitions, as opposed to the negative, dissolving ones identified in the existing triad literature (e.g. Simmel, 1950). When a member of a triad forms an internal or external coalition in order to gain more power for voicing and making changes within the triadic relationship, forming of a coalition aids the recovery process and keeps the triad together. This unifying role was also found in terms of external coalitions, e.g. in the case when the media was brought in. When discussing the roles of the actors, the

630

A. Salo et al. / Industrial Marketing Management 38 (2009) 618–632

initiator of a unifying coalition can be labeled as Unifier, again, opposed to the negative roles of an oppressor and tertius gaudens discussed in the triad theory. The unifier role differs from a mediator: a mediator operates in between the other two actors, whereas a unifier forms a coalition to influence either the single actors (in the case of internal coalition) or both actors in the triad (in the case of external coalition). In addition, the study provides a new concept of leverage voicing by which we refer to the formation of an external coalition that increases the power of voicing within the triad and, therefore, aids in recovering the triadic relationship. 6. Managerial discussion The results of this study provide three types of managerial implications. For firms in business-to-business markets that aim to develop and maintain valuable relationships with their customers, suppliers, and other stakeholders, knowledge about the sub-processes of relationship recovery, the importance of the relationship atmosphere in the process, and all the roles that third parties may play in the process are useful. Next, we will discuss each of these issues in more detail. Firstly, a company can enhance its chances to maintain valuable relationships by providing its partners opportunities for early voicing. Moreover, the company needs to listen to the concerns of its partners and third actors, as well as communicate the concerns within the company. Communication is a tool for ensuring that voicing is heard and responded to. Being able to respond to the concerns early on enables the company to start a recovery process before the relationship is severely damaged. However, a company may not be especially receptive to voicing, but as the study showed, even in this case all hope is not lost. The case revealed that even severe problems such as damaged trust and commitment can be restored, if the actors can establish a strong will to recover the relationship. The willingness can be achieved by analyzing the current state and potential of the relationship/triad and comparing it to the outcome that ending the relationship/triad would have for all parties. The analysis can be done internally in both companies, but the issue can also be discussed jointly (Tähtinen & Vaaland, 2005). The following questions may help the companies to complete the analysis: What will happen to the companies' market shares and incomes? Can ending the relationship harm the reputation of the companies? Are there any alternative partners, and how much time and effort is needed to end this relationship and to establish a new one? If the analysis shows that the particular relationship/triad is worth saving, the actors can make a joint plan on normalizing actions. The plan should include an agreed goal towards which all actors strive, engagement of the right people who can forget the old problems and aim for the goal, as well as ways to find the root causes of the problems, so that they can be attended to. Agreeing on the concerted goal, though, is not enough, but setting smaller and concrete targets gives a task list that helps in evaluating the progress of the normalizing process. Setting up routines and patterns for communication, such as regular meetings, and defining spheres of responsibilities are important parts of the normalizing process. It may be beneficial to change the key persons from all actors, so that they can have a clean start. Finding the root causes for the relationship's problems requires reviewing the relationship's past and asking the right questions, such as, when did the first problems occur, and what enabled the problems to appear? In this case study, for example, one reason for the major problems was that the level of service quality was defined inadequately right from the beginning. Moreover, the final sub-process, the aftermath, also needs attention from managers. By arranging an aftermath session for all the actors, each individual involved can reflect back on what happened and how the relationship was restored, and to plan how all actors can manage the relationship together better in the future. The communication routines

developed during the restoration process may be adjusted to fit the current situation. Moreover, this is a stage where the actors learn from their experiences. The lessons need to be written down and distributed to all actors. This allows for the good practices to be implemented in managing and maintaining other relationships. For example, monitoring the quality of a company's most important relationships can be recommended, not only because it helps to solve problems early on but also because it encourages open communication. In some cases, especially if the problems of the relationship have been in the public eye, the aftermath offers a chance of gaining positive publicity by communicating the good state of the relationship also to the surrounding network. Secondly, managerial implications relate to the ways the restoring actions influence the atmosphere in the relationship. Analyzing the relationship is a step towards increasing the lost attraction. If attraction can be restored, it will be the basis of the will to recover the relationship and, thus, it will help to guide the process forward. Moreover, during the whole recovery process, good communication skills, information exchange, and investments are essential in increasing trust and commitment. Communication may need to take place in all levels within and between the involved organizations. Internal communication is needed in order to clarify the situation to the whole personnel, especially to the boundary spanners. External communication may be needed to analyze the relationship and to respond to the third actors' concerns about the relationship. Thirdly, it is valuable for the manager to recognize the roles that third parties may play in a recovery process. When third parties, such as the customer's customers, are acknowledged as major actors in the relationship, their influence can be taken into account. Being proactive may even save the relationship from entering a troubled stage in the first place, and, at the very least, an understanding of the triad, coalitions, and the role of third parties helps in managing the relationship. Coalitions can be used to unify the triad, in other words, to aid in the relationship recovery process by providing more power. Coalitions can provide the necessary leverage in voicing the concerns in the first place. In addition, the third actors can be useful in defining the root causes of the problems and, therefore, considerably facilitate the recovery. Finally, we wish to stress that even though business relationships are built between companies, they always rely on the engagement of the people acting on behalf of the companies. Once a recovery process is initiated, everyone involved needs to be informed about it and its progress. This is important because the employees are likely to be aware of the problems or, as this case study shows, their work may suffer because of them. It is not enough, however, for only one partner to have people devoted to recovering a relationship/triad. If the counterpart does not see the relationship as worth continuing (i.e. if their reasons for ending the relationship are stronger than the attenuating factors), the relationship is likely to end (see Halinen & Tähtinen, 2002). Thus, managers should try to avoid blaming individual employees for failures in the recovery processes. 7. Evaluation of the study We evaluate the study by using the criteria for trustworthiness developed by Lincoln and Guba (1985). Accordingly, qualitative research can be evaluated through credibility, transferability, dependability, and confirmability of the research. Firstly, we have increased the credibility of this research by data triangulation, in other words, by using multiple sources of data (interviews and newspaper articles as well as other media coverage), and by interviewing representatives of all actors in the triad. In addition, after the interview data was transcribed verbatim, the document was sent to the interviewees to ensure that their narratives matched their memories of how the events took place. Triangulation of data also improves the dependability of the results by allowing congruence to be perceived in the data collected from the different actors of the triadic relationship.

A. Salo et al. / Industrial Marketing Management 38 (2009) 618–632

Secondly, similarly to any single case study, our study describes the phenomenon in its own context. A careful representation of the context improves the transferability of the results. In this study, the context was described in the case presentation so that the characteristics of the context come across and evaluation of the case becomes possible, not only in its own context, but also when relating the results of the research to other contexts. Finally, to improve confirmability, the case analysis is presented with quotations from the interview data. This shows the chain of logic and helps the readers to make their own assessment of the analysis.

631

formation (see e.g. Cook & Emerson, 1978) and it can, thus, be identified as an important future research challenge. Therefore, the way that internal and external coalitions proposed in this study influence recovery processes deserves attention. Furthermore, more research is also needed on the positive role of a third actor as a unifier, another new concept found in this study, related to coalitions. Finally, since a triad is the smallest network (Simmel, 1950), we can even suggest that a triadic view offers a net view on relationship recovery. Thus, although several actors may be involved, their action and their influence on the relationship's atmosphere may be analyzed using the presented triadic model.

8. Suggestions for future research The aim of this study was to model a triadic relationship recovery process and detail the actions to save the relationship. Thus, the model presented here is one step in building a theory on business relationship recovery, while specifically taking into account the third actors involved. Accordingly, the aim was to generalize at the analytical level (see Yin, 2003). Thus, the concepts and their relations should also be helpful in understanding recovery processes in similar contexts. Although the empirical context was professional services, this study contributes to the knowledge on business relationships and nets. However, insights from this model could give service recovery researchers inspiration when studying multiple service recoveries within a relationship and applying the value co-creation approach (see Dong et al., 2008; Maxham & Netemeyer, 2002). The results of this study cannot be generalized to other contexts or populations without further investigations. The theoretical concepts suggested here are all context-dependent and, therefore, they may need to be modified to be useful in, for example, less knowledge and human intensive industrial settings. Therefore, both qualitative studies in different contexts and quantitative studies in professional services would help us to determine the limits of the model. Another limitation that this study entails is its focus on the process, not the antecedents of the process. Thus, more attention could be given on such questions as how do the relationship's characteristics and the type of reasons for the troubles influence the outcome of the recovery process. Moreover, this study does not offer views on the situations when a recovery process should be discontinued. From a financial viewpoint, an interesting question is in which situation should a company abandon attempts to recover the relationship and what are the financial consequences? After all, not every relationship is worth saving. Moreover, an interesting issue is if a relationship or triad can be recovered more than once? The present study does not elaborate on the skills to manage a recovery process in depth, although communications skills are suggested to be essential. More research is needed in order to find out what other capabilities are required and how they can be developed without having to go through actual recovery processes. Thus, one open question relates to how to develop relationship recovery capabilities and how to spread the knowledge in the company and to its focal net. In the present study, the adopted triadic view resulted in a process model that included new concepts describing the roles of the third parties. The particular case could not be understood without the triadic view, since the recovery involved not only the buyer–seller relationship that we are usually interested in, but also the other two relationships, i.e. the ones between the buyer and the end customers and the seller and the end customers, although, the latter is noneconomic. We suggest that multiple relationships may also be involved in other contexts of business relationships, and, therefore, it would be worthwhile to apply a triadic analysis more often. Especially in contexts where the business relationship involves more than two actors (e.g. in international trade, logistics, etc.), the third actor's tendency to form coalitions is an interesting issue. The role of power in a recovery process is clearly evident in coalition

Acknowlegdements The authors wish to thank the anonymous reviewers for their valuable comments. This study is part of ValueNet research project, partly funded by the Academy of Finland, Liike2 research programme.

References Alajoutsijärvi, K., Möller, K., & Tähtinen, J. (2000). Beautiful exit: How to leave your business partner. European Journal of Marketing, 34(11/12), 1270−1290. Anderson, C. J., Håkansson, H., & Johanson, J. (1994). Dyadic business relationships within a network context. Journal of Marketing, 58(4), 1−15. Andersson, C. J., & Narus, J. A. (2004). Business market management: Understanding, creating and delivering value, 2nd edition New Jersey: Prentice Hall. Anderson, E., & Weitz, B. (1992, February). The use of pledges to build and sustain commitment in distribution channels. Journal of Marketing Research, XXIX, 18−34. Caplow, T. (1968). Two against one. Coalitions in triads New Jersey: Prentice-Hall. Cook, K. S., & Emerson, R. M. (1978). Power, equity and commitment in exchange networks. American Sociological Review, 43, 721−739. Doney, P. M., & Cannon, J. P. (1997). An examination of the nature of trust in buyer–seller relationships. Journal of Marketing, 61(2), 35−51. Dong, B., Evans, K. R., & Zou, S. (2008). The effects of customer participation in cocreated service recovery. Journal of the Academy of Marketing Science, 36, 123−137. Eisenhardt, K. M. (1989). Building theories from case study research. Academy of Management Review, 14(4), 532−550. Ford, D., Håkansson, H., & Johanson, J. (1986). How do companies interact? Industrial Marketing and Purchasing, 1(1), 26−41. Ganesan, S. (1994). Determinants of long-term orientation in buyer–seller relationships. Journal of Marketing, 58(2), 1−19. Garbarino, E., & Johnson, M. S. (1999). The different roles of satisfaction, trust, and commitment in customer relationships. Journal of Marketing, 63(2), 70−87. Geyskens, I., Steenkamp, J. -B. E. M., Scheer, L. K., & Kumar, N. (1996). The effects of trust and interdependence on relationship commitment: A trans-Atlantic study. International Journal of Research in Marketing, 13, 303−317. Grönroos, C. (1988). The six criteria of good perceived service. Review of Business, 9(3), 10−13. Halinen, A. (1997). Relationship marketing in professional services: A study of agency– client dynamics in the advertising sector. London: Routledge. Halinen, A., & Tähtinen, J. (2002). A process theory of relationship ending. International Journal of Service Industry Management, 13(2), 163−180. Hallén, L., Johanson, J., & Seyed-Mohamed, N. (1991). Interfirm adaptations in business relationships. Journal of Marketing, 55(2), 29−37. Harris, L. C., O'Malley, L., & Patterson, M. (2003). Professional interaction: Exploring the concept of attraction. Marketing Theory, 3(1), 9−63. Hirschman, A. O. (1970). Exit, voice, and loyalty: Responses to decline in firms, organizations, and states. Cambridge, MA: Harvard University Press. Holmlund, M. (2004). Analyzing business relationships and distinguishing different interaction levels. Industrial Marketing Management, 33(4), 279−287. Håkansson, H., & Snehota, I. (1995). Developing relationships in business networks. London: Routledge. Jehn, K. A. (1997). A qualitative analysis of conflict types and dimensions in organizational groups. Administrative Science Quarterly, 42, 530−557. Kelley, H. H. (1979). Personal relationships: Their structures and processes. Hillsdale: Lawrence Erlbaum Associates. Kovács, G., & Spens, K. A. (2005). Abductive reasoning in logistics research. International Journal of Physical Distribution & Logistics Management, 35(2), 132−144. Lam, P. -K., & Chin, K. -S. (2005). Identifying and prioritizing critical success factors for conflict management in collaborative new product development. Industrial Marketing Management, 34(8), 761−772. Levin-Rozalis, M. (2004). Searching for the unknowable: A process of detection– abductive research generated by projective techniques. International Journal of Qualitative Methods, 3(2), 1−36. Lincoln, Y., & Guba, E. (1985). Naturalistic inquiry. Thousand Oaks, CA: Sage Publications. Mattila, A. S. (2004). The impact of service failures on customer loyalty. The moderating role of affective commitment. International Journal of Service Industry Management, 15(2), 134−149.

632

A. Salo et al. / Industrial Marketing Management 38 (2009) 618–632

Maxham, J. G., III, & Netemeyer, R. G. (2002, Oct). A longitudinal study of complaining customers' evaluations of multiple service failures and recovery efforts. Journal of Marketing, 66, 57−71. Metcalf, L. E., Frear, C. R., & Krishnan, R. (1992). Buyer–seller relationships: An application of the IMP interaction model. European Journal of Marketing, 26(2), 27−46. Moorman, C., Zaltman, G., & Deshpandé, R. (1992). Factors affecting trust in market research relationships. Journal of Marketing, 57(1), 81−101. Moorman, C., Deshpandé, R., & Zaltman, G. (1993, August). Relationships between providers and users of market research: The dynamics of trust within and between organizations. Journal of Marketing Research, 29, 314−329. Morgan, R. M., & Hunt, S. D. (1994). The commitment–trust theory of relationship marketing. Journal of Marketing, 58(3), 20−38. Paparoidamis, N. G., Chumpitaz, R., & Tähtinen, J. (2007). Relationship recovery and business loyalty: Conceptual model and propositions. Electronic proceedings of relationship marketing summit 2007, 13–15 December, Buenos Aires, Argentina. Peirce, C. S. (1998). The essential Peirce: Selected philosophical writings. Edited by the Peirce Edition Project Indiana: Indiana University Press. Ping, R. A., Jr. (1993). The effects of satisfaction and structural constraints on retailer exiting, voice, loyalty, opportunism and neglect. Journal of Retailing, 69(3), 320−352. Ping, R. A., Jr. (1994). Does satisfaction moderate the association between alternative attractiveness and exit intention in a marketing channel? Academy of Marketing Science, 22(4), 364−372. Ping, R. A., Jr. (1997). Voice in business-to-business relationships: Cost-of-exit and demographic antecedents. Journal of Retailing, 73(2), 261−281. Ping, R. A., Jr. (1999). Unexplored antecedents of exiting in a marketing channel. Journal of Retailing, 75(2), 218−241. Ploetner, O., & Ehret, M. (2006). From relationships to partnerships — New forms of cooperation between buyer and seller. Industrial Marketing Management, 35(1), 4−9. Rusbult, C. E., Zembrodt, I. M., & Gunn, L. K. (1982). Exit, voice, loyalty, and neglect: Responses to dissatisfaction in romantic involvements. Journal of Personality and Social Psychology, 43, 1230−1242. Simmel, G. (1950). Quantitative aspects of the group. In K. H. Wolf (Ed.), A translation of Chapter II of Soziologie, in The Sociology of George Simmel (pp. 87−177). Glengoe: Free Press. Smith, P. C., & Laage-Hellman, J. (1992). Small group analysis in industrial networks. In B. Axelsson & G. Easton (Eds.), Industrial networks — A new view of reality (pp. 37−61). London: Routledge.

Tax, S. S., Brown, S. W., & Chandrashekaran, M. (1998). Customer evaluations of service complaint experiences: Implications for relationship marketing. Journal of Marketing, 62(2), 60−76. Thibaut, J. W. & Kelley, H. H. (1968). The social psychology of groups. Reprint, Transaction Books, New Brunswich. Tähtinen, J. (2002). The process of business relationship ending — Its stages and actors. Journal of Market-Focused Management, 5(4), 331−353. Tähtinen, J., & Vaaland, T. I. (2005). Business relationships facing the end: Why restore them? Journal of Business & Industrial Marketing, 21(1), 14−23. Tähtinen, J., Paparoidamis, N. G., & Chumpitaz, R. (2007). Business relationship recovery — A process model. Proceedings of the 23rd annual IMP conference, Manchester, UK. Yin, R. K. (2003). Case study research: Design and methods, (3rd ed). Thousand Oaks: Sage Publications. Anniina Salo is a part-time doctoral student at the University of Oulu, Finland. She has a Master of Science degree (Econ.&Bus.Adm.) and she has graduated in 2007. Businessto-business relationships and dynamics of business nets are her fields of interest in research. She is currently holding a managerial position in a private health care company in Finland. Jaana Tähtinen is Professor of Marketing at the University of Oulu, Finland, where she earned her doctorate in 2001. Her main research interests relate to management and value creation in business nets and dynamics of business relationships. She is one of the founders of biannual Nordic Workshop on Relationship Dynamics (NoRD) and a founding member of Business Relationship Dynamics Group (BuRD). She has published e.g. in the European Journal of Marketing, Industrial Marketing Management, International Journal of Service Industry Management, Journal of Business & Industrial Marketing, and Marketing Theory. Pauliina Ulkuniemi is Assistant Professor of Marketing at the University of Oulu, Finland, where she also received her PhD in 2003. Her publications and research interests lie in the value creation in business relationships in different industry contexts, especially in business services and in project business. She has published e.g. in the Journal of Purchasing and Supply Management, International Journal of Logistics Management, International Journal of Technology Marketing and International Journal of Service Industry Management.