Antecedents of CP

11 The antecedents of CP consist of two categories: customer and firm. Five customer antecedents seem key to understanding why customers participate or not: role size, role awareness, role clarity, customer ability, and customer willingness. In addition, firms can trigger or hinder participative behavior by organizationally socializing customers. Studies that consider the nature of these antecedents focus on the manner in which they can increase the levels of CP, but they do not take into account the potential impact of these antecedents on the service process itself or on the coordination that supports the process In addition, as Wu (2011) notes, most of this research is theoretical (Bowen, 1986; Etgar, 2008; Kelley, Donnelly, & Skinner, 1990; Rodie & Kleine, 2000); empirical studies are scarce.

12 Among the latter, a majority focuses on customers’ ability, willingness (i.e., customer antecedents), and on the customers’ organizational socialization (i.e., firm antecedent), as it shows that customers’ actual and perceived ability to participate (Auh et al., 2007; McKee, Simmers, & Licata, 2006) and willingness (Meuter, Bitner, Ostrom, & Brown, 2005) favor participative behaviors. Furthermore, customer organizational socialization significantly and positively influences the level of CP in two ways: by increasing customers’ understanding of the firm’s norms, values, and expectations, and by providing customers with the knowledge and skills they need to participate (Claycomb, Lengnick Hall, & Inks, 2001; Wu, 2011). Such objectives can be achieved through organizational communication literature (e.g., brochures, websites, Frequently Asked Questions) or by emphasizing the advantages of CP through discounts, faster delivery, etc. (Bateson, 1985). Employees can also encourage customers to participate, or help them understand the content of the participation and how to participate (Bove, Pervan, Beatty, & Shiu, 2009). Eventually, firms can also implement customers’ online communities that provide information on how to participate , as this increases the level and efficiency of customers’ participative behaviors (Algesheimer, Borle, Dholakia, & Singh, 2010).

Inputs CP

13 Because CP antecedents trigger customers’ participative behaviors in service processes, they influence both the nature and quantity of inputs that customers bring to these processes. Researchers identify seven kinds of inputs: informational [1] , emotional, physical, financial, temporal, behavioral, and relational. These inputs are nonexclusive, and not all of them are necessarily mobilized by a single participating customer. They have been identified in prior literature, yet research on these inputs suffers from three gaps: (1) it is mainly conceptual, (2) it essentially focuses on CP at large, regardless of the nature of the inputs, and (3) it often puts the emphasis on some specific inputs and gives little attention to others.

14 First, similar to a majority of research on CP (Bendapudi & Leone, 2003), studies on CP inputs are primarily conceptual. Inputs have been theoretically defined and discussed (e.g. Bitner et al., 1997; Grönroos, 2001; Kelley et al., 1990; Song & Adams, 1993) but they have rarely been studied empirically. Second gap in research listed above, conceptual and empirical studies are largely restricted in that they examine the extent of customer participation, rather than the inputs of CP themselves. Research shows that the intensity of CP can influence the overall performance of service processes, whether through service delivery methods (Bolton & Saxena-Iyer, 2009; Shahin & Nikneshan, 2008), new product development (Füller & Matzler, 2007), value creation processes (Chan et al., 2010), or service employees’ performance (Chan et al., 2010; Yi, Nataraajan, & Gong, 2011). In other words, these studies examine the influence of the level of CP rather than the influence of the nature of CP, and they aggregate the inputs to measure their combined influence, without distinguishing different inputs. Moreover, some inputs have been better investigated than others. In particular, the influence of the information that customers provide to firms (i.e., informational inputs) has been studied extensively, empirically and theoretically, both in business-to-business and business-to-consumer contexts (Fang, 2008; Fang et al., 2008; Siehl, Bowen, & Pearson, 1992).

15 Thus, the antecedents and inputs of CP alike relate to what customers do in service processes. For that reason, it is suggested here that antecedents and inputs likely play important roles for understanding how customers might influence relational coordination between service employees involved in this process. The next section presents the conceptual framework I propose to link both literature streams and thereby suggest the manner in which this influence may occur.

Table 2

Antecedents and inputs of customer participation

ANTECEDENTS (C = customer; F = firm)

INPUTS

Role size (C) The proportion of the service produced and delivered by the customer (Mills, Chase, & Margulies, 1983; Mills & Turk, 1986; Rodie & Kleine, 2000). Role awareness (C) The extent to which customers are aware of the role they must fulfill to obtain a service (Lovelock, 1983, 2001; Rodie & Kleine, 2000; Siehl et al., 1992). Role clarity (C) The extent to which customers understand the role that they must play (Lengnick- Hall, Claycomb, & Inks, 2000; Meuter et al., 2005). Customer ability (C) The extent to which Customers have, and are convinced that they have, the appropriate knowledge, competence, time, etc., in order to participate (Auh et al., 2007; Chervonnaya, 2003; Etgar, 2008; McKee et al., 2006; Meuter et al., 2005). Customer willingness (C) The extent to which customers are ready to do what the firm expects from them during the process (Anderson, 2006; Auh et al., 2007; Bowen, 1986; Meuter et al., 2005; Rodie & Kleine, 2000). Organizational socialization (F) The process that helps customers to acquire and develop the appropriate skills, knowledge, and attitudes to participate, by using organizational literature (e.g., brochures, websites, etc.), thus emphasizing the advantages of CP, and inciting employees to encourage customers to participate, among other things (Algesheimer et al., 2010; Bove et al., 2009; Claycomb et al., 2001; Kelley et al., 1990; Kelley, Skinner, & Donnelly, 1992; Wu, 2011).

Informational Information the customer gives to the firm and mental efforts customers make to obtain the service (e.g., to understand how to participate) (Fang, 2008; Fang et al., 2008; Mills et al., 1983; Mills & Turk, 1986; Rodie & Kleine, 2000). Emotional Any emotion that the customer may feel during the service encounter (Rodie & Kleine, 2000). Physical The “customer’s own tangible and physical efforts” (Rodie & Kleine, 2000, p. 112), including goods that the customer provides to the firm so that the latter has an action on them (Lovelock, 1983; Rodie & Kleine, 2000; Siehl et al., 1992). Financial The price the customer pays to the firm (Bitner et al., 1997). Temporal The duration of CP, including the time the customer spends trying to understand the way the process works even before participating in it (Lovelock, 2001; Mills & Morris, 1986; Song & Adams, 1993). Behavioral The way the customer behaves during interpersonal interactions that take place with service employees during a service encounter (Kelley et al., 1990; Plé et al., 2010). Relational The customer’s state of mind resulting from past service encounters when entering a new encounter in a similar context (Grönroos, 2001; Plé et al., 2010).

Antecedents and inputs of customer participation

Research field

19 This research was performed in two French retail banks (names changed for confidentiality): FB1, a mutual regional bank, and FB2, a national bank. Like other banks in their sector, both were developing multiple channels, including diverse new channels such as sophisticated call centers (CCs), mobile services, Internet services, etc. However, because my purpose was to explore customers’ influence on relational coordination between frontline employees, I focused exclusively on branches and CCs in FB1 and FB2, which communicate and work with one another. I also excluded self-service channels that involve no customer–employee interaction.

20 Both banks had set up these channels for similar reasons. Their main objective was to outsource many operations that did not generate any direct profit for the bank (e.g., transfers or accounts consultation) to customers. In so doing, they hoped to increase the productivity and efficiency of branch advisors, who would no longer be disturbed by phone calls during their selling time. FB1 and FB2 also wanted to improve service quality; for example, 25 % of the phone calls to FB1 branches were unanswered, compared with only 5 % of phone calls to the CC. To ensure that customers would use the inbound CC, both banks listed the CC phone number, and changed the phone number of the branch. Moreover, both banks communicated these shifts heavily to customers and employees, to help them understand the channels’ complementarity. This move was indispensable, because many customers were unhappy about the change and feared that their branch would be closed and replaced by the CC. However, FB1 and FB2 insisted that its sole purpose was to help the branches and provide customers with better service in the course of their global, multichannel relationship with the bank. Branch advisors and CC advisors also used different means to share information (internal e-mail and telephone at FB1; fax and very occasional phone calls at FB2). Finally, both banks implemented outbound CCs, employing outsiders with no previous banking experience. They had to canvass customers to get appointments with branch advisors.

Data collection

Data triangulation

21 The relative dearth of academic literature on the research topic and the “how” nature of my research question together raise the need for qualitative data that can provide rich descriptions and explanations of processes grounded in the reality of a local context (Miles & Huberman, 1994). It is interesting to note here that most of earlier research on relational coordination has traditionally relied on quantitative methods. This study thus provides qualitative input into a quantitative stream of research. The collection of qualitative data relied on two case studies that lasted approximately 11 months. With a view to triangulation (Miles & Huberman, 1994; Yin, 2003), I used a variety of collection methods, sources of information, and types of data to enhance the reliability of the data (Table 3).

Table 3

Data collection methods and sources

Documentary analysis	Direct unstructured observation	Semi-structured interviews
- Internal communication (e.g., notes, memos) - External communication (e.g., leaflets, brochures, letters sent to customers) - Articles from banking and general business journals	- Note taking when in the branches or CC (use of waiting times to observe, listen, note useful information – Yin, 2003) - One half-day observation in a CC of each bank	- 42 interviews (recorded) – 9 branches, 3 call-centers, and the banks’ headquarters - FB1: 24 interviews (5 branches, 2 CCs, and the headquarters) - FB2: 18 interviews (4 branches, 1 CC, and the headquarters)

Data collection methods and sources

22 The interviews were the main source of data. Extant literature suggests that customers may influence relational coordination through service employees’ perception of CP, so I decided not to interview customers. Instead, I met branch advisors, CC advisors, and site managers (i.e., managers of a branch or CC), who interact daily with customers and whose jobs – which involve delivering banking services to customers – are interdependent. The sites were chosen in such a way that, in each of these, I could meet diversified respondents who shared many characteristics with respondents from other sites (e.g., age, seniority, structure of the clientele). In addition, I ensured that branches sampled from FB1 and FB2 were from the same cities or geographically very close. Thus, I favored heterogeneity among the respondents within each site and homogeneity in the respondents’ characteristics from one site to another. All the recorded interviews were structured in four parts: daily work, introduction of new channels, channel coordination, and consequences of customers’ perceptions of the development and use of new channels. A short synthesis (no longer than two pages) was then sent to each interviewee for validation. On-site interviewing had three important implications for the data collection. First, the interviews were time-constrained; I was granted approximately the same amount of time as a customer (1–1.5 hours). Usually, the interview ended when a customer arrived for an appointment, though many interviewees granted me additional time (even sometimes to the detriment of their customers) to ensure we had covered all the topics. Second, most of the interviewees either relied on events that occurred during the interview (e.g., they received an e-mail from the CC, or a customer call) or gave me some documents to illustrate their statements. Third, I was able to observe the interviewees and their colleagues in their working environment, which provided worthwhile additional information and gave me a better understanding of their perspective.

Data analysis

25 As soon as I became aware of this role stress, I began my analysis of the first set of data. Following prior literature, I developed a thematic coding list that included both dimensions of relational coordination (communication and relationships) and their sub-dimensions, as well as the inputs and antecedents of CP. I progressively refined and enriched this list in the coding process to reflect the subjective perspective of the actors, in accordance with my interpretive approach of the field and the data.

26 First, a new sub-dimension of relational coordination related to relationships rapidly emerged. I called it “mutual leniency,” because it describes the capacity of frontline service employees to forgive their colleagues who work in other channels for making mistakes in their work. Mutual leniency differs from mutual respect, insofar as the latter indicates situations in which service employees value the work contribution of their colleagues from other channels. In contrast, mutual leniency refers explicitly to the tolerance and understanding that some employees show towards colleagues who work in other channels and make mistakes because of the customer’s participation in the process. For instance, some branch advisors noted that the lack of accurate information from CC advisors could be due to the fact this information was not provided by the customer. Because they encountered this kind of situation with their own customers, some branch advisors were more tolerant and ready to forgive their CC colleagues.

27 A second complementary important theme that emerged from the data is “customer–employee interaction.” This encompasses all the information related to the nature of the interactions between the customer and frontline employees, as well as how those interactions occurred over time. This theme progressively appeared to play a role both in the way that customers influenced relational coordination, and the scope of this influence.

28 The data were coded with Nvivo 7 software; all the documents, observational notes, and interviews had been transcribed. This was helpful to create relationships among the codes and exemplify them with quotes. Thus, I gradually built a helpful corpus of relationships that gave a better insight into the data and provided more complex, detailed explanations of the links between CP and relational coordination. I also relied heavily on three other Nvivo tools to analyze each case and compare them. In particular, I created many matrices using the “Queries” tool. For example, I determined the number of interviews in which a code appeared, or the occurrences of this code in all the interviews. I occasionally used the matrices for quantitative analysis of the data, for two reasons: firstly, to ensure the process was not distorted by my own perception, due to the aforementioned role stress, and secondly, to remain as close as possible to the interviewees’ perspective and subjective frame of reference. For example, the content analysis of one FB1 branch led me to believe that customers’ influence was higher (in terms of frequency) than in the other FB1 branches. However, a cross-branch comparison of the number of words that illustrated this phenomenon resulted in low differences. The “Sets” tool also enabled me to regroup the information according to the functions of the site, case, relationship, and other factors. For instance, a set called “FB1–CP influences RC” regrouped relationships to illustrate the influence of CP inputs and antecedents on relational coordination (RC) sub-dimensions in FB1. I was thus able to quickly access the data grouped by sites, cases, and relationships. Finally, I used the “Models” tool to illustrate schemes of the relationships among many different codes. These schemes were useful as they provided more detailed explanations of the links between CP and relational coordination.

The influence of CP on the communication dimension of relational coordination

30 The influence of CP on the communication dimension is identical for FB1 and FB2 (see Table 4).

Communication frequency

31 My results indicate that the branch and CC advisors’ perception of CP influences the frequency of communication between actors. First, customers do not always accept the new rules that result from the existence of the CC (willingness). Many of them refuse to give CC advisors the information needed to handle their demand, thus forcing CC advisors to send a message to the branch: “When the customer definitely refuses us to take care of her demand, then we send the

Table 4

The influence of CP on the communication dimension at FB1 and FB2

	FB1 and FB2
Relational coordination – Communication	Frequency	Accuracy	Timeliness	Problem-Solving
Antecedents of CP	Willingness	Willingness	No influence	No influence
Inputs of CP	Behavioral Informational Emotional	Informational	No influence	No influence

The influence of CP on the communication dimension at FB1 and FB2

32 message” (CC, FB2). Second, specific customer behaviors such as impatience (behavioral inputs) may demand more frequent communications: “Sometimes, even though we are meeting a customer, we receive a second message from the call-center. The customer has phoned again, and told them, ‘She did not call me back, is it normal? Can you please give her the message again?” (Branch, FB1). Third, the nature or quality of information the customer provides (informational inputs) may oblige CC advisors to contact the branch: “According to the nature of the information the customer gives us, we may need to contact the branch” (CC, FB2). Lastly, the emotional state of the customer (emotional inputs) greatly influences the intensity of communication between the channels: “Sometimes, if they have someone who gets worked up, they call and ask us to deal with the customer, because the situation is tense” (Branch, FB1).

Influence of CP on the relationship dimension of relational coordination

35 In contrast to the communication dimension, the manner in which CP influences the relationship dimension of relational coordination greatly differs between FB1 to FB2 (see Table 5).

Table 5

The influence of CP on the relationships dimension at FB1 and FB2

Relational coordination – Relationship	Shared knowledge	Shared goals	Mutual leniency	Mutual respect
	FB1
Antecedents of CP	No influence	No influence	Willingness	Willingness
Inputs of CP	No influence	No influence	Behavioral	Behavioral Emotional Informational Relationa
	FB2
Antecedents of CP	Ability Willingness	Willingness	No influence	Willingness
Inputs of CP	Behavioral	Informational	No influence	Behavioral Emotional Informational Relational

The influence of CP on the relationships dimension at FB1 and FB2

Emergining result: differential influence of CP on relational coordination

44 An unexpected result emerged from this analysis: customers’ influence over relational coordination seemed different when they interacted with branch advisors, versus CC advisors. The influence of CP on branch advisors seems to depend on the level of shared knowledge between the channels: “Some customers often told me they had to wait at least ten minutes before they could talk to somebody, and I thought, well, it is not good, not good at all… But now, I know how it works, I went there and I saw them, so I tell the customers, ‘Listen, it is impossible, they have these technical things that prevent customers from waiting more than, say, 3 minutes. Before that, we did not really know, and we had to be the devil’s advocate, in a way, between the call-center and the customers” (Branch, FB2). Because customers and branch advisors gained repeated contacts over time and knew each other personally, the latter were prone to give more credit to their customers when the level of shared knowledge between the channels was low, but they tended to take some distance from what the customers told them when they learned how their colleagues worked in other channels. This finding suggests that the influence of CP on relational coordination could evolve over time in function of the firm’s decisions and actions (e.g. by increasing the level of shared knowledge). Long-term relationships between customers and branch advisors also enabled the latter to recognize the usual behavior of many of their customers, which made them more or less inclined to believe what they were told: “You know, we have customers who are very rude. So from time to time, when one we know not to be very polite tells us ‘they answered me like that, it’s not normal’… We think, well, he must behave the same on the phone, so even though it is the customer’s [word] versus the call-center advisor’s word, it is not necessarily the call-center that did something wrong” (Branch, FB1).

45 In contrast, the absence of long-term relationships between customers and CC advisors had two consequences. Firstly, “because they do not know [CC advisors]” personally (CC, FB1) and do not have repeated contacts with them, customers are reluctant to give information to CC advisors. This diminishes the communication accuracy between the channels. Secondly, customers have far less power over CC advisors, whose job is highly constrained by the organizational procedures that they must follow, and who are less impressionable because of the absence of direct and repeated contacts with the same customer: “I often have customers who insist [on getting] in touch with their branch. The pressure is very high, sometimes, but the rule is the rule. I only have to take them through if they have a particular need” (CC, FB1).

Theoritical implications

47 This research offers two key contributions to the study of relational coordination. First, it illustrates how customers can influence relational coordination between frontline service employees. Second, it suggests that the nature and history of the interactions between customers and service employees are potential moderators of this influence. In addition, this study suggests a new sub-dimension of relational coordination: mutual leniency. Figure 1 builds on these two contributions to propose a conceptual framework of how the customer may influence relational coordination between service employees.

Figure 1

A framework proposal of how the customer may influence relational coordination

Contact employees’ perception of CP
inputs and determinants
++orCustomer–employee
or • Natuirneteorfatchteioinnteraction
• History of the interaction
Contact Employee A Relational coordination Contact Employee B

A framework proposal of how the customer may influence relational coordination

48 The main contribution of this study pertains to the influence of CP on relational coordination. Previous research on relational coordination has been limited, thus far, to the study of interactions among employees. Customers were either not taken into account (Ahuja, 2003; Anderson, 2006; Gittell, 2002b) or taken into account only indirectly, through the uncertainty that they brought to the process (Gittell, 2002a, 2008). Yet, scholars in services marketing have shown the importance and influence of the customer on service processes (Bendapudi & Leone, 2003; Bolton & Saxena-Iyer, 2009; Wiertz et al., 2004). By combining these two literature streams, my research views the customer as a participant in relational coordination and reveals that the interactions between a customer and frontline service employees may influence relational coordination among the latter. This influence occurs through the service employees’ perceptions of the inputs and antecedents of CP and applies to the two dimensions of relational coordination: communication and relationships. It seems to occur in three different, complementary ways.

49 First, customers can filter the information between employees, whether consciously, by providing no information at all or different information to interacting employees, or unconsciously, by explaining their issues in a different manner or forgetting to provide the same information. At this level, customers influence the accuracy of the communication between service employees, insofar as they do not deliver the same quality or quantity of information to the diverse employees with whom they interact. Given that relational coordination is coordination between roles, this means that customers endorse a role of “informational filter” between the employees with whom they interact.

50 Second, customers also influence the frequency of communication between employees. For instance, they may refuse the conditions of their participation with some employees or behave in a manner that initiates or discourages communication and interaction among employees. In line with the role-based nature of relational coordination, the customer could thus be called an “interactional catalyst,” insofar as the way in which some of the inputs and antecedents of CP are perceived by service employees provokes the presence, or absence, of interactions between employees.

51 Third, customers may influence relational coordination when they positively or negatively modify the perception that employees have of one another. This aspect involves the relationship dimension of relational coordination. The inputs and antecedents of CP seem to influence how service employees regard one another’s work and capacity to fulfill their function, manifested in the level of mutual respect or mutual leniency. In addition, this perception might be altered by the influence of the customer on the level of shared knowledge and shared goals between employees. Thus, customers can trigger either the improvement or weakening of the opinion that service employees have of one another. Accordingly, I suggest that customers may act as perceptual catalysts in their relational coordination between service employees.

52 To reflect the insights that come with this first contribution, I offer the following propositions:

53 Proposition 1: Customer participation influences relational coordination among frontline employees through the perception that these employees have of (a) the inputs and (b) the antecedents of customer participation.

54 Proposition 2: The customer may play the role (s) of (a) “informational filter,” (b) “interactional catalyst,” and/or (c) “perceptual catalyst” in relational coordination among frontline employees who must coordinate with one another.

55 Next, the unexpected result that emerged from the analysis may lead to a second theoretical contribution: the scope of the customer’s influence may depend on the nature of the interactions between the customer and frontline employees. These interactions can be of two kinds: a service relationship or a service encounter (Gutek, 2000). In the former, the customer and service employee expect to have repeated contacts in the future and to get to know each other on a personal level, such that “they develop a history of shared interaction that they can draw on whenever they interact to complete some transaction” (Gutek, 2000, p. 371). In this paper’s empirical study, this refers to the nature of the interactions between customers and branch advisors. Conversely, in service encounters, the customer interacts with different service employees over time – such as when contacting a CC – and these are expected to be functionally equivalent. Hence, the customer does not expect to be served by the same employee again, and no history of interactions develops (Gutek, Groth, & Cherry, 2002). My results illustrate that customers may have a stronger influence on relational coordination when they have relationships, rather than encounters, with service employees. Indeed, in the case of service encounters, employees seemed less influenced by customers, and instead seemed to favor procedures over customers. This tendency also arose with customers’ decreased willingness to provide information to employees whom they did not know or necessarily trust, which in turn lowered communication accuracy among employees. Contrariwise, the influence of customers seems greater in service relationships, because employees are more likely to favor their demands. However, a high level of shared knowledge between employees appears to counterbalance the influence of CP. A potential explanation for this is that prior interactions with service employees influence subsequent customer reactions (Bolton, Smith, & Wagner, 2003). Thus, I offer the following additional propositions:

56 Proposition 3: The scope of the customer’s influence on relational coordination among frontline employees depends on (a) the nature and (b) the history of interactions between the customer and frontline employees.

57 Proposition 4: The customer has a stronger influence on relational coordination among frontline employees when having a relationship, rather than an encounter, with these employees.

58 Finally, an emerging result also suggests a new sub-dimension of relational coordination, which I call mutual leniency. Although not directly related to the main aim of the study (i.e., to explore the influence of the customer on relational coordination), it may enrich the concept of relational coordination. Previous research has identified shared knowledge, shared goals, and mutual respect as the three sub-dimensions of the relationship dimension of relational coordination (Gittell, 2002b; Gittell et al., 2010). This study suggests that mutual leniency could be a fourth. Mutual leniency refers to the capacity of service employees to forgive colleagues with whom they coordinate for making mistakes in their work. It differs from mutual respect, which pertains to situations where service employees value the contributions of their colleagues. Because relational coordination focuses on relationships between roles (Carmeli & Gittell, 2009; Okhuysen & Bechky, 2009), mutual leniency means forgiving colleagues who did not fulfill their role, whereas mutual respect means acknowledging colleagues’ high-level role fulfillment. Mutual leniency thus requires a certain amount of role empathy among employees, since they must show tolerance and understanding with regard to the mistakes made by their frontline colleagues. In this study, I focused on such mistakes as results of customer participation, but other contexts could be considered. Mutual leniency could occur even in the absence of a customer, depending on work circumstances (e.g., when an employee makes a mistake because they are new to the job or when an employee makes an uncharacteristic mistake) or individual characteristics (e.g., a personal tendency to be tolerant or show understanding). Thus:

59 Proposition 5: Mutual leniency, defined as the capacity of service employees to forgive colleagues with whom they coordinate for making mistakes in their work, is a fourth complementary sub-dimension – in addition to shared knowledge, shared goals, and mutual respect – of the relationship dimension of relational coordination.

Managerial implications

60 At a managerial level, this research underscores the potential importance of managing CP to improve coordination processes in service firms. The inappropriate management of CP is likely to lead to at least three detrimental consequences. First, relational coordination is supposed to increase organizational efficiency (Gittell, 2002a; Shah et al., 2008), but the influence of CP on relational coordination may make this relationship more complex to determine. Second, relational coordination plays a significant role in the level of service processes’ quality outcomes (Gittell, 2002b; Gittell et al., 2010) and, accordingly, in the level of customer satisfaction (Anderson, 2006), such that customer satisfaction may be negatively altered by the influence of CP on relational coordination. Third, firms may overestimate their customers’ profitability, because the coordination costs that may result from the detrimental influence of CP on relational coordination are usually either unforeseen or hidden.

61 Firms could avoid or limit these possible damages by enhancing customer socialization, increasing shared knowledge among employees, and developing new indicators of customer profitability. First, firms should rely on socialization techniques in a traditional manner (that is, integrate customers in their processes more efficiently) and in a way that can transform CP into a positive organizational leverage. This could improve many dimensions of relational coordination, and would prevent customers from raising relational barriers between employees. Put another way, firms should learn how to use their customers, because their positive feedback may facilitate organizational change (here, the introduction and development of a new channel).

62 Second, firms should encourage good communication and high levels of shared knowledge between service employees, especially if these employees are geographically distant. In order to limit the detrimental consequences of negative customer feedback on relational coordination, companies need to ensure that customers’ positive feedback is diffused and known throughout the organization. This is likely to reinforce mutual respect and increase perceptions of shared goals.

63 Third, because of the potential coordination costs resulting from the detrimental influence of CP, classic indicators of customers’ profitability could become less reliable—a risk that has most certainly increased in a world where both firms and customers use multiple channels (Kumar, Lemon, & Parasuraman, 2006), and therefore where customers are likely to interact with many different frontline employees. It thus seems necessary to forecast the potential negative impact of customers on service organizations. Firms should draw detailed blueprints that would allow them to identify the manner and moment in which a customer may hinder relational coordination between employees. They can also rely on their own customer relationship management data to statistically evaluate the manner and the moment in which customers are most likely to negatively influence such coordination.

Limitations

64 Although the current study provides an important step in clarifying how customers fit into, and affect, relational coordination, it has a number of limitations. The first is that although relying on qualitative data led to noteworthy results, it also rendered these results less amenable to replication. This study setting is restricted to a specific sector of activity that does not support generalization of the results. Second, I was not able to evaluate the impact of each bank’s different status or structures on the influence of the customer. As a mutual bank, FB1 aims to be closer to its customers and to implicate them more than FB2, but my data did not enable me to determine whether these statutory differences altered the influence of CP. Organizational differences between the two banks occasionally seemed to explain some variation in the scope of customer influence on relational coordination, yet these differences were not studied in depth here.

65 Third, I collected the data only from employees, which was appropriate considering the overall research aim, namely, to investigate how service employees’ perception of CP could influence relational coordination between these employees. However, collecting additional data directly from customers or, even better, from customer–employees dyads would likely have provided complementary insights. For example, I could have put into perspective the actual desire of customers to influence employees and relational coordination, as well as the importance that employees grant such participation. This would have enabled me to determine whether the customer’s influence was voluntary or involuntary. Finally, customers’ individual characteristics (e.g., age, income, educational level, lifestyle) were not taken into account, and might help to explain some of the customers’ relative influence on relational coordination.

Further research avenues

66 These limitations can be seen as fruitful avenues for further research. In addition, scholars may wish to pursue at least three other directions. First, the use of quantitative studies to test the propositions and the conceptual framework formulated in this article would provide worthwhile, generalizable insights for the study of relational coordination. Additional qualitative research about relational coordination and the potential influence of the customer might enable a deeper exploration of the phenomena pointed to in this study. Although research on relational coordination has partially involved some qualitative research methods (Gittell, 2001, 2002a), most work relies essentially on quantitative approaches (Carmeli & Gittell, 2009; Gittell, 2000, 2008; Gittell et al., 2010). A qualitative stance, now that the concept is theoretically well-established, could result in a more profound understanding of its underlying processes and enable “analytic (link to theory) and naturalistic (link to experience) generalization” (Buchanan & Bryman, 2007, p. 494). This would usefully complement, and be complemented by, statistical generalizations from quantitative research.

67 Second, this study has underscored how customers can harm coordination among employees. High levels of relational coordination improve employees’ performance and customers’ outcomes (Gittell et al., 2008; Shah et al., 2008), yet this study shows that these can potentially be limited by CP. In other words, CP generates coordination costs, as manifested in losses to the firm’s efficiency or customer satisfaction. However, researchers and practitioners alike rarely take these costs into account (Chan et al., 2010), or else do not recognize the potential positive influence of customers. In a world where co-creating with customers is becoming the norm (Cova & Salle, 2008; Dong et al., 2008; Vargo & Lusch, 2004), the potential harm CP can bring to coordination raises the following worthwhile research avenues: To what extent can customers generate coordination costs? To what extent do these costs limit the firm’s efficiency? How can a firm combat hidden coordination costs?

68 Third, it would be of interest to explore the influence of CP on coordination mechanisms. The study of coordination comprises research on the organizational design and coordination mechanisms on the one hand, and on the coordination process (relational coordination) on the other (Okhuysen & Bechky, 2009). Some scholars suggest that customers influence coordination mechanisms through the uncertainty they introduce in service processes (Argote, 1982; Bowen & Jones, 1986; Larsson & Bowen, 1989; Rathnam et al., 1995). Yet, the exploration of whether the antecedents and inputs of CP actually influence these coordination mechanisms might lead to a more fine-grained management of CP and the firm’s coordination mechanisms, which could help limit the detrimental effects or increase the positive consequences of CP.