1 Introduction

Research has often addressed negative customer reactions to service failure as expressed in complaints (Mittal et al. 2008), desire for retaliation (Grégoire et al. 2009), negative word of mouth (Grégoire and Fisher 2006), and switching behavior (Keaveney 1995). However, customers might also react to an unsatisfactory service with forgiveness, i.e., a prosocial change toward a perceived transgressor (McCullough 2000). Previous research has overlooked such constructive reactions, so that there is only partial understanding of customer reactions to unsatisfactory service at present (Tsarenko and Rooslani Tojib 2011; Zourrig et al. 2009).

Exploring customer forgiveness is in line with a positive organizational scholarship perspective, which is designed to balance the tendency to focus on negative organizational aspects (Dutton and Glynn 2007) and address a wider set of constructs. Essentially, forgiveness by one party toward a transgressor is considered a prerequisite to restoration of a harmonious relationship (e.g., Finkel et al. 2002). Understanding customer forgiveness might assist organizations in actions designed to restore a positive relationship with customers, beyond the notion of customer retention (Tsarenko and Rooslani Tojib 2011).

Furthermore, most research on customers’ reactions to service failure has addressed customers’ attitudes toward the service organization rather than the individual employee (Grégoire et al. 2009; Mittal et al. 2008). Yet, Forgiveness is most centrally characterized as an “intraindividual, prosocial change toward a perceived transgressor that is situated within a specific interpersonal context” (McCullough 2000, p. 9). Accordingly, we suggest that, following unsatisfactory service, forgiveness toward an employee might be different from attitudes toward an organization because the former might be characterized by factors that are unique to interpersonal relationships. Prominent examples are the relationships of forgiveness with empathy (McCullough et al. 1997), compassion, sympathy (Takaku 2001, 2006), or caring for the other’s well being (Karremans and Van Lange 2004). All these emotions could be experienced in regard to another individual but are less likely to be experienced in regard to an organization. Furthermore, research has demonstrated the importance of interpersonal relationship in customers’ evaluations of a service and loyalty to the organization (Gremler and Gwinner 2000). In discussing the differences between service relationships (repeated customer contact with the same employee) and pseudo-relationships (customer contact with different employees in the same organization), Gutek et al. (1999) suggested that service relationships can develop into friendships via development of mutual positive emotions and attitudes. Understanding customers’ forgiveness can elaborate our understanding of the positive aspects of customer-employee relationships.

The first purpose of this research is to document manifestations of costumer forgiveness in order to gain more comprehensive understanding of customers’ reactions to transgressions. For this purpose, we have conducted a qualitative study exploring experiences of forgiveness following an unsatisfactory service. The second purpose of the research is identification of the antecedents of customer forgiveness. We explored customers’ willingness to forgive employees’ unsatisfactory service as a function of relationship strength, employee blame, and customer gender. The choice of these variables was guided by (a) previous research demonstrating their relevance for customer reactions to service failures (Grégoire and Fisher 2006; Grégoire et al. 2009; Keaveney 1995; Miller et al. 2008; Zourrig et al. 2009); (b) the relevance of the variables to the service employee-customer interpersonal relationship; and (c) the motivation to include both personal (gender) and situational (blame) variables to explore their effects on forgiveness. More specifically, relationship strength was found to affect customer negative reactions to service failures (Grégoire and Fisher 2006; Grégoire et al. 2009). While previous research focused on strength of relationship with the service organization, we explore the effect of strength of relationship with the service employee, because customer-employee relationship is highly influential in affecting customer attitudes and behavior (Gremler and Gwinner 2000; Gutek et al. 1999). Blame represents a process of attribution which affects the reactions to transgression (Aquino et al. 2001; Joireman et al. 2013) and in the service context was found to moderate the effect of relationship quality (Grégoire and Fisher 2006). Moreover, gender was found to significantly affect the tendency to forgive a transgression (Miller et al. 2008) and interacts with relationship quality to affect customer behavior (Melnyk et al. 2009).

Understanding the impact of situational and personal variables on that forgiveness can contribute to creating tailored service recovery models that take into account customers’ gender and type of relationship with service providers.

2 Study 1

2.1 Literature review

In recent years, scholars have begun to coalesce around the conceptualization of forgiveness as a motivational phenomenon (Fehr et al. 2010; Fehr and Gelfand 2012). Forgiveness is often defined in terms of behavioral, affective, or cognitive responses following interpersonal offense. Forgiveness involves acknowledging the seriousness of the offense, but also the process of relinquishing vengeful thoughts and revenge motivations (Tsarenko and Rooslani Tojib 2011). Individuals are described as forgiving if they inhibit retaliatory or destructive responses, responding instead with conciliatory or constructive behaviors, affect, and cognitions (McCullough 2000; McCullough et al. 1997; Rusbult et al. 1982, 1988). As victims of transgressions come to forgive, motivations for revenge and avoidance give way to motivations that are benevolent and prosocial (Fehr and Gelfand 2012). Thus, most theorists agree that an increase in positivity in terms of attitudes and/or behavior toward the offender is a vital part of true forgiveness (McCullough 2000; Riek and Mania 2012). Tendencies toward retaliation are modified when greater value is placed on constructive responses concerning situations and/or relationships (Zechmeister et al. 2004). Forgiveness also involves both an inner dimension—the forgiving party’s mental state, and an interpersonal dimension—the ongoing relationship in which forgiveness occurs (Finkel et al. 2002).

Due to the highly competitive nature of the service industry, customers often have idealized expectations of flawless service (Boulding et al. 1993). Accordingly, most research has focused on customers’ negative reactions to service failures (Grégoire et al. 2009; Mittal et al. 2008). Keaveney (1995) found that core service failure, reflected in mistakes, billing errors, and service catastrophes, was the most frequently mentioned reason for customers’ switching behavior. The second most frequently mentioned events responsible for service switching were service encounter failures (employees who were uncaring, impolite, unresponsive, or unknowledgeable). Nonetheless, it has been shown that customers sometimes choose to avoid such reactions (Tsarenko and Strizhakova 2009; Wieseke et al. 2012; Zourrig et al. 2009), for example if they have a long-term positive relationship with an organization (Hess et al. 2003a, b; Mittal et al. 2008; Tax et al. 1998).

Several definitions of forgiveness emphasize that forgiving individuals may also respond constructively in such a relationship (e.g., McCullough et al. 1997, 1998; Takaku 2001a, b). Zechmeister et al. (2004) found that forgiveness and retaliation are affected by different variables and concluded that “forgiveness requires more than simply returning to a ‘neutral’ point following an offense” (p. 557). Similarly, Brown’s research (2003) showed that the dispositional tendency to forgive is both theoretically and empirically distinct from dispositional vengeance. Thus, customers may be unforgiving and bear a grudge without engaging in retaliation, or may forgive even though they complain about the service.

In the first study, we examined how the goodwill facet of forgiveness is reflected in customers’ reactions to service failures. We asked: What are the manifestations of customer forgiveness of a service employee following an unsatisfactory service experience?

2.2 Method

2.2.1 Respondents

The convenience sample consisted of 52 Israeli customers, 18 men and 34 women. Ages ranged from 20 to 58 years (mean = 32.5). Years of education ranged from 8 to 22 (mean = 14.96). Respondents worked in administration (8), human resources (4), sales (6), management (4), computing (3), finances (3),engineering (2) manufacturing (2) plumbing (2) education (4) law (1), and social work (1). Seven of the respondents were students and five were unemployed at the time of the study. All participants were volunteers. The sampling method was that of snow-ball sampling (Patton 2002), where participants were asked to recommend possible participants who met criteria for participation.

2.2.2 Data collection

We designed a semi-structured, open-ended interview. Participants were first asked to describe an experience of forgiveness: “Please describe a service experience in which something went wrong or you experienced personal injustice, and you decided to forgive even though your emotions were negative.” This was followed by questions about the experience itself and the reasons for the decision. Participants were also asked to describe an experience of deciding not to forgive an unsatisfactory service, followed by questions about making the decision, and the reasons for it.

2.2.3 Data analysis

We analyzed the interviews by coding the main themes and identifying the patterns that emerged in regard to expressions of forgiveness (Miles and Huberman 1994). The analysis was inductive in that the themes emerged from customers’ accounts rather than from a priori hypotheses. We independently identified concepts and dimensions in the data, and grouped them into categories. We then searched for relationships among the categories in order to group conceptually similar codes and relate them to higher-order themes (Glaser and Strauss 1967).

2.3 Findings and discussion

The data indicated that, for customers, forgiveness is reflected in positive reframing of a situation, perspective taking, and demonstrating tolerance. Table 1 summarizes expressions of forgiveness.

Table 1 Expressions of customer forgiveness
Full size table

2.3.1 Positive reframing

Forgiving customers make a conscious decision to focus on positive aspects of a situation instead of on the negative aspects. This was reflected in a positive relational orientation, isolating the event or choosing to “make the best of” the experience regardless of the failure. Positive relational orientation was reflected in the desire to maintain a positive relationship with the service organization despite the service failure. A customer who felt she was given misleading information regarding terms of payment in a gym said: “It is a place I like very much. If I had argued with the employees it would be unpleasant to go there. I’m glad I behaved like that, because if I had fought with them I wouldn’t be able to go back.” These results support the notion that customers who emphasize the importance of interpersonal relationships tend to be more forgiving than individualistic customers (Zourrig et al. 2009).

Similar views were expressed when customers took some responsibility for a failure or tried to prevent such failures in the future. A customer in a pub who was not told that the beer he had ordered was not included in the offered discount said “We go there often and really like the place. I’ve simply learned that I must clarify all details in advance so that this will not happen again.”

Customers were also forgiving when they did not want to hurt an employee by behaving unpleasantly. A woman who was cheated by a cleaning lady said: “I don’t like conflicts. I didn’t want to start an argument, and I didn’t want to hurt her.”

Isolation of a failure by viewing it as a single event was reflected in customers’ emphasis on the good service usually provided by an employee or organization. The customer in the pub said, about the service provided by the waitress—“Before the bill arrived she was really alright with us, and occasionally came by to ask if everything was OK.” Customers also recalled good service provided by organizations in the past. A customer who received misleading information from a salesman of a cell phone company said: “I looked back at my contacts with the company, and saw that they were reliable. There were many times when the employees and the company came to terms with me about prices or service.” Another customer of a cable company, who waited too long for the arrival of a technician, said “I have been their customer for a long time, and over the years several technicians have come and have always arrived on time. Before this happened, I always received good service from them.”

Emotion enhancement was reflected in customers’ determination to nurture positive emotions and suppress or deny negative emotions for their own sake. A customer who was misinformed about the timetable of courses she wished to take described her thoughts after the event: “I decided to make a fresh start so as to have a pleasant feeling when I arrive to the classes, and to prevent anger from interfering with focusing on my studies.” Emotional enhancement was sometimes achieved by focusing on positive aspects of the service experience. Customers expressed a resolution to ‘let go’ and enjoy the service despite the failure, and thus maintain their positive emotions. A hotel guest who had some personal belongings damaged by fresh paint said: “I decided to see the positive side and be happy about this opportunity to get out of our routine and have fun.” Another customer who felt that the dish she ordered in a restaurant was overpriced said: “I decided not to complain and just enjoy the meal with my mother.” Thus, forgiveness in this respect may be viewed as a coping strategy designed to reduce stress and maintain the customer’s mental wellbeing (Tsarenko and Gabbott 2006; Tsarenko and Rooslani Tojib 2011; Tsarenko and Strizhakova 2009).

2.3.2 Perspective taking

Forgiveness involves relating to a situation from the employee’s point of view. Identification with the employee was stimulated by similarities between the customer and the service provider. When a customer was told that the printing of her wedding invitations would be delayed, she said: “The employee explained to me that printing would be delayed due to the holidays. I was more forgiving because I’m religious and don’t work on holidays, and could understand the reason.” Several customers who had worked in service jobs understood what the employee was going through. A customer of a waitress who got the orders wrong said: “I know how stressful and unpleasant it is to make a mistake. I could see how she felt and remembered how it felt”. Customers have also acknowledged employees’ difficulties and viewed them as possible reasons for service failures. A customer who received an impatient and rude response when asking for information said: “I told myself that she must be exhausted from all the customers who call and having to give personal service at the same time. It was obvious to me that I wasn’t the first person to call her about this matter”.

Understanding the service provider’s perspective also involved awareness of organizational constraints that might affect service, which often resulted in accepting service failures as inevitable. A customer who received misinformation regarding payment from a travel agent said: “I understood that some things happen by mistake, and you don’t have to make a big deal out of them. I understood that failures are not intentional, which is why I spoke to her nicely and quietly.” Perspective taking can also increase customers’ appreciation of efforts made by employees to provide good service. A customer who had argued with an employee regarding replacement of broken apparatus said: “I can understand his side. He’s committed to the organization and its interests, and because he put this aside and responded personally I appreciated his behavior even more.” These results support previous findings showing that empathic customers are more likely to respond to a dissatisfying encounter with forgiveness reflected in loyalty (Wieseke et al. 2012).

2.3.3 Tolerance

Forgiveness does not involve passive acceptance of a service failure. Customers often required correction of the failure, but their reaction was polite and pleasant. Because the anticipated reaction to service failure is negative, customers see themselves as forgiving when they engage in positive communication and refrain from harmful reactions. In addition to positive communication, forgiveness is also reflected in avoiding behaviors such as complaining, shouting, demanding an apology or compensation, or engaging in negative word of mouth which are often taken for granted following a service failure. A customer who received misleading information regarding the price of a laptop said “I didn’t fight with the employee or ask to talk to his manager, but solved the problem together with him.” However, customers’ intention of exhibiting patience and calm often required mental effort. A customer who did not receive the newspaper he had ordered for a long time said: “I decided to be patient because I know it’s unpleasant for the service representatives to receive angry calls, so I made an effort to stay calm.” In the long run, tolerant behavior was reflected in customers’ decisions to remain loyal to the organizations in question. A customer who found an error made by the bank said: “I forgave her. I’m still at the same bank, and I didn’t tell anyone.”

The results suggest that customer forgiveness does not merely comprise refraining from negative acts. It involves considerable cognitive and behavioral effort to enhance the positive aspects of a service and mitigate the negative aspects of a failure. Following such efforts, which are frequently motivated by goodwill, customers are able to maintain a friendly and empathic attitude toward the service provider.

In summary, the findings of Study 1 indicate that, for customers, forgiveness is manifested in both reasoning and behavior of adopting a more benevolent stance toward the transgression (Bright and Exline 2011a, b). Forgiveness is reflected in regarding the transgression in a more positive light (positive reframing), considering the other’s point of view (perspective taking) or restraining adverse spontaneous reactions and behaving kindly.

The results of Study 1, as well as previous research on forgiveness and customers’ reactions to unsatisfactory service, have guided us in determining the exploration of antecedents of customer forgiveness in Study 2. The results suggest that some themes are associated with cultivating interpersonal relationships (e.g., positive relational orientation, positive communication). Other themes indicate the impact of attributions made by customers (e.g., identification with the service employees, awareness of organizational constraints, isolation of a failure). Research in non-work contexts suggests that the willingness to view a transgression in a more positive light (positive reframing) or to consider the other’s viewpoint (perspective taking) are affected by the quality of the victim relationship with the transgressor (Finkel et al. 2002). Research conducted in the service context, suggests that customers’ negative reactions to unsatisfactory service is affected by the length of relationship with the organization as well as the attribution of blame (Goodman et al. 1995; Grégoire et al. 2009; Hess et al. 2003a, b; Mittal et al. 2008; Tax et al. 1998).

Accordingly, in Study 2, we explored the effects of relationship strength and blame on forgiveness. We also explored the impact of gender, which has been found to play a significant role in forgiveness (Miller et al. 2008).

3 Study 2

3.1 Literature review

3.1.1 Strength of relationship and forgiveness in response to unsatisfactory service

The quality of customer-organization relationships significantly affects customers’ responses to transgressions (Berry 1995; Goodman et al. 1995; Kelley and Davis 1994). Furthermore, the nature and the process of forgiveness are complex and occur differently in different types of relationships (Tsarenko and Strizhakova 2009). However, results of the effect of strength of relationship on customers’ negative responses to a service failure are inconsistent. Some studies show that customers in long-term relationships with an organization react less negatively to a service failure (Hess et al. 2003a, b; Mittal et al. 2008; Tax et al. 1998), while others show the opposite effect (Goodman et al. 1995; Grégoire et al. 2009). Retaliation is motivated by a wish to punish (an organization) for violation of the central promise of the exchange relationship between service employee and customer (Bies and Tripp 2005; Grégoire and Fisher 2006). This might be stronger in a close relationship, in which violation generates a sense of betrayal. Conversely, forgiveness has an intrapersonal dimension that transcends calculations of exchange and reflects a sense of goodwill toward the other (McCullough et al. 1997). Such goodwill, expressed in forgiveness, is likely to be more pronounced in a strong relationship in which the customer develops positive emotions toward the employee. Commitment between the offender and victim is likely to influence one’s reaction to an offense. When people have invested a great deal in a relationship and see it as beneficial, they may be motivated to “explain away” the transgressor’s behavior in order to allow the relationship to continue. Indeed, as the level of commitment between two people increases, the likelihood of forgiveness also increases (Finkel et al. 2002; Riek and Mania 2012). Worthington (2005) found that when strangers or people in poor relationships offend, the focus of forgiveness is on reducing negative emotions. In valued relationships, the focus is on both reducing the negative and increasing the positive emotions toward the offender. Gutek (1995) suggests that, in strong relationships, the customer believes that the service provider is interested in him/her as an individual and that the service encounter transcends commercial transaction parameters. Strong relationships are characterized by social benefits for customers (e.g., understanding, familiarity), the creation of a dual bond between parties, and higher levels of customer trust and satisfaction.

The results of the qualitative study indicate that forgiveness consists of considerations regarding the personal and interpersonal impact of forgiving and is related to loyalty. Thus, in order to gain better understanding of customer forgiveness, we have operationalized forgiveness as reflected in the willingness to forgive a service employee, the expected outcomes of forgiveness and loyalty intentions. Expected outcomes of forgiveness (Exline et al. 2004) address the anticipated impact of forgiveness on the customer’s self-image and his/her relationship with an organization. Loyalty intentions following a service failure reflect the customer’s long-term willingness to forgive (Grégoire and Fisher 2006). Loyalty indicates patience, i.e., taking no action and believing that a situation will improve (Geyskens and Steenkamp 2000). Rusbult et al. (1982, 1988) describe loyalty as a passive-constructive behavior because it is directed at improving a relationship by being supportive. Loyalty following a service failure thus reflects the goodwill that underlies forgiveness.

Because a strong relationship involves trust and belief that the relationship with the service provider is beyond commercial considerations (Gutek 1995), we suggest that customers in strong relationships expect forgiveness to be associated with positive outcomes in terms of future relationship with the service provider. Based on the evidence that high-quality relationships inhibit negative reactions and enhance positive reactions to a transgression (Finkel et al. 2002; Grégoire and Fisher 2006), we also expect the strength to the customer–service employee relationship to have a positive effect on customers’ willingness to forgive. Lastly, customers in strong relationships who experience positive emotions and commitment toward the service provider (Gutek 1995; Mattila 2001) are expected to be more loyal, even after a transgression. Hence, following an experience of unsatisfactory service,

Hypothesis 1a

A customer in strong relationship with a service provider will expect more favorable outcomes of forgiveness than a customer in a weak relationship.

Hypothesis 1b

A customer in a strong relationship with a service provider will be more willing to forgive than a customer in a weak relationship.

Hypothesis 1c

A customer in a strong relationship with a service provider will express stronger loyalty than a customer in a weak relationship.

3.1.2 The impact of blame and gender on forgiveness

When something goes wrong, customers attempt to figure out who is to blame for a service failure. They might believe that the failure was due to the employee’s lack of caring and negligence, or that it was not the employee’s fault (Grégoire and Fisher 2006; Zourrig et al. 2009). Customers blame the employee when they believe that s/he has violated the norms for appropriate conduct, and this is perceived as intentional, controllable, and avoidable (Joireman et al. 2013; Zechmeister et al. 2004). Research suggests that a high level of blame attribution for a transgression increases the tendency to behave negatively and reduces the willingness to forgive. For example, Joireman et al. (2013) found that following a service failure, the attribution of blame to negative motives (e.g., prioritizing sales over service) increased the wish for retaliation, while the attribution of positive motives (e.g., prioritizing the customer’s needs) increased the wish to engage in reconciliation. A customer who blames the service employee for the service failure, is less likely to expect positive outcomes as a result of forgiving, because blame involves attribution of negative motives to the service provider (Joireman et al. 2013) and thus reduces the prospect of enhancing positive relationship through forgiveness. Blame is also expected to have a negative effect on the willingness to forgive because it generates anger, thereby increasing a drive that is opposite to forgiveness, i.e., the motivation to retaliate (Aquino et al. 2001; Joireman et al. 2013). Customers who blame the service provider are expected to be less loyal following a transgression because blame involves lack of trust in the service provider’s goodwill and might thus impair the expectation of high-quality service in the future (Joireman et al. 2013). Accordingly, we present the following hypotheses:

Hypothesis 2a

A customer who attributes a low level of blame to a service employee will expect more favorable outcomes of forgiveness than a customer who attributes a high level of blame to the employee.

Hypothesis 2b

A customer who attributes a low level of blame to a service employee will be more willing to forgive than a customer who attributes a high level of blame to the employee.

Hypothesis 2c

A customer who attributes a low level of blame to a service employee will express stronger loyalty than a customer who attributes a high level of blame to the employee.

The attribution of blame interacts with relationship strength to affect customers’ reactions to service failure. Grégoire and Fisher (2006) found that attribution of a low level of controllability triggers the wish to retaliate among customers in weak relationship, but not among customers in a strong relationship with the service provider. Accordingly, we propose that following an experience of an unsatisfactory service, forgiveness will be affected by an interactive effect of blame and relationship strength. Specifically, blame is expected to have a weaker negative effect on the expectation of forgiveness outcomes when relationship is strong because customers in strong relationships are motivated to maintain their positive view of the organization (Grégoire and Fisher 2006) and are therefore more likely to be optimistic about the outcomes of forgiveness. Blame is also expected to have a weaker inhibiting effect on the willingness to forgive when the relationship is strong because customers in such relationships might downplay the importance of the blame in order to ensure consistency between their general opinion of the organization and the failure (Grégoire and Fisher 2006). Relying on the study of Hess et al. 2003a, b, who found that expectations of relationship continuity and lower attributions of failure to stable causes were associated with higher satisfaction following service recovery, we suggest that the effect of blame on loyalty intentions will be weaker in strong customer-service provider relationships. We thus hypothesize

Hypothesis 3a

The effect of blame on customers’ expectations regarding the outcomes of forgiveness will be stronger in a weak customer-employee relationship than in a strong relationship.

Hypothesis 3b

The effect of blame on customers’ willingness to forgive will be stronger in a weak customer-employee relationship than in a strong relationship.

Hypothesis 3c

The effect of blame on customers’ loyalty intentions will be stronger in a weak customer-employee relationship than in a strong relationship.

Research indicates that forgiveness is affected by gender. A meta-analysis of the relationship of gender and forgiveness (Miller et al. 2008) indicates that women are more forgiving than men. The authors suggested that women’s stronger inclination to forgive might be related to gender differences in dispositional qualities as women tend to be more agreeable and empathic. In addition, men may be more drawn to justice-based morality and to responses to transgressions emphasizing vengeance or justice while women may be more drawn to warmth-based virtues (Miller et al. 2008). Thus, we expect a higher willingness to forgive among women. Additionally, women view themselves as more interdependent and focus more on maintaining relationships (Melnyk et al. 2009). We thus suggest that women will expect more positive outcomes of forgiveness in terms of relationship maintenance with the service provider. Because women exhibit higher levels of intrinsic interpersonal commitment in the relationship with their service providers (Bhagat and Williams 2008), we expect women to express stronger loyalty intentions than men following a service failure. We accordingly hypothesize the following gender differences in customer forgiveness:

Hypothesis 4a

Women will expect more favorable outcomes of forgiveness than men.

Hypothesis 4b

Women will be more willing to forgive than men.

Hypothesis 4c

Women will express stronger loyalty intentions than men.

Gender differences in the willingness to forgive are also affected by relationship strength. Baumeister and Sommer (1997) suggested that women tend to focus on establishing and maintaining a few close relationships with specific individuals (relational interdependence), whereas men tend to focus on establishing and maintaining relationships with more anonymous and larger groups of people (collective interdependence). Regarding gender differences in the service context, the indications are that service relationships are more significant to women. Noble et al. (2006) found that women’s loyalty to service organizations derives more from motives of social interaction, whereas men’s loyalty derives more from the wish to acquire information. We therefore suggest that women in a strong relationship with a service provider will expect forgiveness to have more positive effects on the relationship and will be more willing to forgive, compared to women in a weak relationship, while men’s expectations will be less affected by relationship strength. Melnyk et al. (2009) found that female consumers tend to be more loyal than males to individual service providers, whereas the opposite is the case in regard to service organizations. Melnyk and van Osselaer (2012) found that women responded more positively to loyalty programs emphasizing personalization. Thus, women in a strong relationship with a service provider are expected to be more loyal following service failures than women in a weak relationship, while relationship strength is expected to have a weaker effect on men’s loyalty following a service failure.

In sum, women not only tend to be more forgiving than men, but also ascribe greater significance to relationship quality in a variety of situations including service interactions. Therefore, the quality of relationship, which was found to affect reaction to service failures (Mattila 2001), can be expected to have a stronger influence on women’s forgiveness. When the relationship with a service employee is strong, women’s forgiveness can be expected to be higher than men’s forgiveness because women attribute more importance to such strong relationships (Bhagat and Williams 2008), and will be more motivated to maintain the relationship by experiencing and expressing forgiveness. Accordingly, we propose that following an experience of an unsatisfactory service:

Hypothesis 5a

The effect of relationship strength on customer expectations regarding the outcomes of forgiveness will be stronger for women than for men.

Hypothesis 5b

The effect of relationship strength on customer willingness to forgive will be stronger for women than for men.

Hypothesis 5c

The effect of relationship strength on customer loyalty intentions will be stronger for women than for men.

3.2 Method

3.2.1 Participants, design, and procedure

We used a convenience sample of customers. Of the participants (N = 286), 55.4 % were female, average age of 35.63 (SD = 11.92) years, average years of education 14.96 (SD = 2.24).

We used a 2 × 2 factorial design (relationship strength: weak, strong; blame: low, high) with both factors varied between respondents.

3.2.2 Manipulation and procedure

Relationship strength and blame were manipulated with scenarios describing service interactions in a restaurant, based on Mattila (2001) and Mittal et al. (2008). Scenarios, frequently used in service research, allow for more convincing evidence of causal relationships than other designs, and enhance internal validity because they control for respondents’ perception of independent variables (Cooper and Emery 1995; McQuilken and Robertson 2011). Moreover, scenarios are appropriate tools for measuring effects of unsatisfactory service, because creating service failures to simulate real experiences poses ethical problems (McQuilken and Robertson 2011). The restauant scenario represents a familiar experience, thus enhancing participants’ sense of involvement in the service event. Respondents were asked to imagine that they were the customers depicted in the scenarios, and to think about how they would have felt and reacted.

Relationship strength was manipulated as follows:

Strong relationship You and your friend have decided to go out for dinner at an Italian restaurant called Toni’s. You have returned to the same restaurant for many months, and the same waiter has always waited on you. As you enter the restaurant, the waiter recognizes you, escorts you to a table, and after an exchange of ‘news,’ he takes your order.

Weak relationship You and your friend have decided to go out for dinner to an Italian restaurant called Toni’s. A waiter escorts you to a table and takes your order.

The effect of the manipulation was tested with the following questions (Mittal et al. 2008): “How close, would you say, are you to this waiter?” (1 = not close at all, 7 = very close); “How friendly do you feel towards this waiter?” (1 = not friendly at all, 7 = very friendly). The results of a t test of a combined index (r = 0.83, P < 0.01) show that sense of closeness and familiarity were significantly higher in the strong relationship condition (M = 4.76, SD = 1.48) than in the weak relationship condition (M = 2.82, SD = 1.57) (t = 10.64, P < 0.01) .

Blame was manipulated as follows:

High blame After 30 min you are still waiting for your food. All that time you see the waiter standing in the corner, talking on his cellular phone and occasionally laughing. Finally, after an hour, the waiter brings your food.

Low blame After 30 min you are still waiting for your food. Due to the bad weather conditions there are disruptions in the electricity supply of the entire area, including the restaurant, and the cooks are having a hard time preparing the dishes. Finally, after an hour, the waiter brings your food.

The effect of the manipulation was tested with the following questions (Russell 1982): “In your opinion, to what extent could the waiter control the waiting time for the food?” (1 = to a very low extent; 7 = to a very high extent); and “In your opinion, to what extent was the waiting time the waiter’s fault?” (1 = to a very low extent; 7 = to a very high extent). The results of a t test of a combined index (r = 0.83, P < 0.01) show that blame was significantly higher in the high blame condition (M = 4.72, SD = 2.07) than in the low blame condition (M = 2.82, SD = 1.57) (t = 17.67, P < 0.01) .

3.3 Measures

Expected outcomes of forgiveness were measured with items from a scale developed by Exline et al. (2004). After reading the prompt, “If I forgive the waiter in this situation…,” respondents rated responses on a scale from 1 (strongly disagree) to 7 (strongly agree) for the following items: “I would feel I was getting less than I deserved”; “I would feel weak”; “I would have less respect for myself afterwards”; “I would lose power within the situation (originally “within the relationship”);” I would feel better about myself”; “I would feel peaceful.” Reliability = 0.81.

Willingness to forgive was measured with a scale used by Xie and Peng (2009). The items were adjusted to the context, i.e., instead of “company” as in the original scale, respondents were asked to refer to the waiter. Respondents rated willingness to forgive the waiter on a scale from 1 (not at all true for me) to 7(very true for me) for the following items: “I would think favorably of this waiter,” “I would condemn the waiter,” “I would forgive the waiter,” “I would disapprove of this waiter.” Reliability = 0.80.

Loyalty intentions were measured with two items adapted from Hennig-Thurau et al. (2006): “I will return to this restaurant” and “I plan to receive service from the waiter in the future” (r = 0.58, P < 0.01). Response scale = 1 (strongly disagree) to 7(strongly agree).

Control variable -customer entitlement, recognized as generating extremely high expectations regarding service quality (Boyd and Helms 2005), was the co-variate in all analyses. The scale, developed by Boyd and Helms (2005), consists of nine items (e.g., “I absolutely believe in the saying ‘the customer is always right.’”). Response scale = 1 (strongly disagree) to 7(strongly agree). Reliability = 0.84.

3.4 Results and discussion

Data were analyzed with three ANOVA analyses testing effects of the independent variables (gender, blame, relationship strength) on expected outcomes, willingness to forgive, and loyalty intentions. Consistent with hypothesis 1, a significant effect of relationship strength was found regarding willingness to forgive (F (1, 281) = 16.84, P < 0.01), with higher willingness to forgive in strong relationships (M = 4.80, SD = 1.41) than in weak relationships (M = 4.28; SD = 1.36). The effect of relationship strength (F (1, 284) = 71.09, P < 0.01) on loyalty intentions was also significant. As expected, loyalty was higher under strong relationship (M = 4.44, SD = 1.70) than under weak relationship (M = 3.04, SD = 1.36). However, no significant main effect of relationship strength on expected outcomes was found.

Hypothesis 2 posited that higher blame would result in less forgiveness. The results show significant main effects of blame on expected outcomes of forgiveness (F (1, 285) = 15.10, P < 0.01), with more negative outcomes expected under high (M = 2.74, SD = 1.32) than under low blame (M = 2.20, SD = 1.10). The effect of blame on willingness to forgive was also significant (F (1, 281) = 146.53, P < 0.01), with higher willingness to forgive under low blame (M = 5.37, SD = 1.03) than under high blame (M = 3.76, SD = 1.28). Additionally, blame significantly affected loyalty intentions (F (1, 284) = 78.45, P < 0.01), which were stronger under low blame (M = 4.46, SD = 1.65) than under high blame (M = 3.02, SD = 1.39).

Hypothesis 3 addressed the interactive effect of relationship strength and blame on forgiveness. The results show significant interaction effect on expected outcomes of forgiveness (F (1, 285) = 3.86, P < 0.05). To interpret the form of the interaction, we conducted simple effects tests, assessing the impact of blame manipulations at each level of relationship strength. Consistent with the hypothesis, tests of simple effects showed that, under weak relationships, there were no significant differences in expected outcomes of forgiveness between high and low blame (F (1, 276) = 1.99, P = 0.16) but under strong relationships, expected outcomes differed significantly across levels of blame (F (1, 276 = 17.47, P < 0.01). Additionally, a significant interaction effect of relationship strength and blame on loyalty was found (F (1, 284) = 5.20, P < 0.05). Simple effects tests show that the difference between high and low blame under strong relationships (F (1, 276) = 63.06, P < 0.01) is considerably larger than under weak relationships (F1, 276) = 21.40, P < 0.01) (see: Table 2; and Figs. 1, 2). No significant effect was found with regard to the willingness to forgive.

Table 2 Means and standard deviations—expected negative outcomes of forgiveness and loyalty intentions in low and high levels of relationship strength and blame
Full size table
Fig. 1
figure 1

Effects of relationship strength and blame on expected outcomes of forgiveness

Full size image
Fig. 2
figure 2

Effects of relationship strength and blame on loyalty intentions

Full size image

Hypothesis 4 addressed the effect of gender on forgiveness. As predicted, the results show a significant effect of gender on expected outcomes of forgiveness (F (1, 276) = 6.48, P = 0.011), with men expecting more negative outcomes than women (M = 2.68, SD = 1.31 and M = 2.30, SD = 1.16, respectively). However, no significant effects of gender were found in regard to willingness to forgive or loyalty intentions.

Hypothesis 5 predicted an interactive effect of gender and relationship strength. As expected, the results show a significant interaction effect of relationship strength and gender on loyalty intentions (F (1, 276) = 3.96, P < 0.05). Simple effects tests show that for women the difference between weak and strong relationships (F (1, 276) = 60.44, P < 0.01) is stronger than for men (F (1, 276) = 19.00, P < 0.01). The results are presented in Fig. 3.

Fig. 3
figure 3

The effect of relationship strength and gender on loyalty intentions

Full size image

While no significant interaction effect of gender and relationship strength was found regarding expected outcomes of forgiveness or willingness to forgive, a significant triple interaction effect was found (F (1, 284) = 12.85, P < 0.01) in regard to willingness to forgive. Simple test effects show significant gender difference under strong relationships and high level of blame (F (1, 276) = 15.10, P < 0.01). The results, presented in Table 3 and Fig. 4, show that under weak relationship the negative effect of blame on willingness to forgive is stronger for women, while under strong relationship the effect is reversed, and is stronger for men.

Table 3 Means and standard deviations of willingness to forgive by gender at low and high levels of relationship strength and blame
Full size table
Fig. 4
figure 4

Effects of blame and gender on willingness to forgive in weak and strong relationships

Full size image

The results thus demonstrate the interactive effect of relationship strength, blame, and gender on customer forgiveness. We found that strong relationships only contribute to customer forgiveness when blame is low. Conversely, the combination of strong relationship and high blame decreases forgiveness. Women were found to be more forgiving than men under strong relationships, even when they attributed high blame to the service employee.

4 General discussion

Customers’ reactions to unsatisfactory service experiences are important for determining satisfaction and loyalty (DeWitt and Brady 2003). However, most studies have focused on negative reactions (e.g., retaliation, Grégoire et al. 2009), without examining possible positive reaction such as forgiveness, which embodies a benevolent approach as well as the absence of negative emotions. The present study addresses this shortcoming by exploring manifestations of customer forgiveness as well as documenting conditions that foster such forgiveness after an unsatisfactory service experience. Considering the effect of the interpersonal connection between customers and employees on customers’ satisfaction and loyalty (Gremler and Gwinner 2000; Gutek et al. 1999) we have focused on interpersonal forgiveness, i.e., toward a service employee. The first contribution of the research is in promoting the understanding of the construct of customer forgiveness. The results of Study 1 show, for the first time, the composition of forgiveness in the service context. Forgiveness directed toward a service employee involves reframing events, viewing situations from the employees’ viewpoint, and engaging in tolerant behavior toward the employee. The results support previous research indicating that empathy, by drawing customers to consider the perspective of the service employee enhances the inclination to forgive (Exline and Zell 2009). Our findings also support the premise that forgiveness requires more than returning to a neutral point in a relationship (Zechmeister et al. 2004). Forgiveness is an effortful, transformative process that occurs despite the recognition that an offense occurred and was wrong (Fehr and Gelfand 2012) involving deliberation and effort on the part of customers. In the service context, in which organizations often encourage customers’ demanding behavior, forgiveness requires customers’ emotional and behavioral self-regulation, because entitlement often inhibits forgiveness (Exline et al. 2004).

In Study 2, we explored factors that might affect customers’ motivation to engage in such efforts of forgiveness. The second contribution of the research concerns the effect of relationship strength on customer forgiveness. Our results suggest that strong relationships enhance benevolent reactions. Customers in strong relationships tend to be more committed, patient, and trusting (Gutek et al. 1999), and consequently more inclined to be positive in their reactions to transgressions (Ahluwalia 2000; Ahluwalia et al. 2000). However, blame relationship strength interactively affect forgiveness: blame has a stronger effect on customers’ expectations of negative outcomes of forgiveness and reduced loyalty intentions under strong relationships. This, as in previous research, suggests that service failure, interpreted as reflecting insufficient motivation, engenders more severe reactions by customers who expect high-quality service on the basis of strong relationship (Grégoire and Fisher 2006). Moreover, such customers might also feel that they are entitled to special treatment due to their loyalty (Grégoire et al. 2009). Thus, forgiveness based on relationship strength is limited when a relationship is felt to be unilateral because a service employee does not do “his/her share” in the relationship, the level of forgiveness drops considerably. However, this is mainly applicable in regard to male customers.

Thirdly, the study shows the effect of gender on customer forgiveness to individual service providers. In line with previous research (Miller et al. 2008), we found that women expect less negative outcomes of forgiveness following unsatisfactory service experiences, and that their willingness to forgive is more strongly affected by relationship strength than that of men. Women are more inclined to form close relationships with a small number of people, and to attribute importance to such relationships (Baumeister and Sommer 1997). Thus, in cases of unsatisfactory service, women may react more benevolently if the service provider is viewed as being close. Conversely, our results suggest that, for men, unsatisfactory service experiences have the same meaning regardless of their relationship with the employee. Their major relationship is with the organization Melnyk et al. (2009), so that a relationship with a specific employee is irrelevant.

A fourth and unexpected finding is that forgiveness is a function of the combined effect of blame, gender, and relationship strength. We found that women are not only more forgiving under strong relationship, but also under a high level of blame, i.e., that in such cases, women are less affected by the level of blame than men. These results may be explained by Gilligan’s premise (1994) of gender differences in moral reasoning, according to which women are oriented toward a motivation to preserve relationships, while men are oriented toward a need for justice done through considerations of fairness and equity. Thus, in terms of customer profiles, these results suggest that the most forgiving customers are women in strong relationship with the employee, who do not blame the employee for the unsatisfactory service; while the least forgiving customers are men in strong relationship who blame the employee for the service.

4.1 Managerial implications

Service failures and unsatisfactory service experiences are inevitable in organizations. Since “the customer is always right,” organizations often expect problems after a service failure—customers leaving, retaliation, negative word of mouth. However, customers may decide to forgive and maintain their loyalty. Our results show that engaging in forgiveness is often a deliberate decision that involves cognitive processing of the failure event. Management might be able to affect this process by providing information to customers (e.g., explaining the failure) or affecting customer’s emotions (e.g., apology) immediately after the failure, thereby increasing customers’ inclination to engage in forgiveness rather than a negative reaction (e.g., negative word of mouth).

The results indicate that from the customers’ viewpoint, forgiveness is not the default reaction to a service failure but rather an exceptional act reflecting kindness and tolerance. Accordingly, following a service failure it is desirable to communicate to forgiving customers that their reaction is appreciated by the organization. Our results also suggest that under strong relationships and when the employee is not considered as responsible for the problem, customers react with forgiveness and remain loyal; but when the relationship is weak, customers react less positively even if the service problem is not viewed as the employee’s fault. These results suggest that, while explaining the reasons for a service failure to customers is always important, such information could be especially effective in maintaining the goodwill of customers in strong relationships. Organizations should therefore invest efforts in explaining the reasons for service failure to loyal customers, especially if the failure is not under the organization’s control.

Lastly, the results show gender differences in regard to forgiveness. Our results suggest that different strategies for service recovery might be developed for men and for women. For example, for men, service recovery activities could focus on the restoration of justice (e.g., through compensation). For women, service recovery could be personalized and emphasize the service provider’s awareness of the customer’s specific requirements. Such adapted strategies might increase the likelihood that men, as well as women, will decide to react with forgiveness to a service failure.

4.2 Limitations and future research

The current study has several limitations. First, the research addressed service transgressions representing low-level seriousness. In Study 1, respondents have mainly reported mild transgressions (e.g., receiving inaccurate information from a service employee) while the manipulation in Study 2 also addressed a transgression that lacks any grave consequences (having to wait in a restaurant). Yet, the service context (e.g., hospitality vs. medical services), and the resulting seriousness of consequences of transgression, might play a role not only in forgiveness level but also in the content of forgiveness and its antecedents. Future research should explore customer forgiveness in various service contexts as well as in regard to seriousness of consequences. The role playing in Study 2, although essential for experimental manipulation, elicits less involvement from participants than a real service encounter. Thus, participants’ responses to our scenarios may be weaker than their reactions to actual problems and recoveries (Hess et al. 2003a, b). Furthermore, because we have focused on customers’ reactions to service employees, we do not know if these found effects are also applicable to customer forgiveness of organizations. This should also be explored in other types of customer–organization relationships (e.g., pseudo-relationships where the customer has repeated contact with the organization but with different service providers). The results of Study 1 indicate that forgiveness evokes recall of high-quality services previously provided by an organization. It would be desirable to explore the organizational attributes (e.g., caring, personalization, lack of failure) that precede forgiveness of unsatisfactory service. Lastly, previous studies have shown that service recovery efforts have a significant effect on customers’ reactions (e.g., Hess et al. 2003a, b). Future research should investigate which aspects of such efforts can generate benevolence and forgiveness.

5 Conclusion

Customer forgiveness is a multidimensional reaction, involving cognitive processing of a service failure event, as well as the development of a tolerant attitude toward the service provider. The inclination to engage in the process of forgiveness is affected by combination of the customer’s gender, strength of the customer–service provider relationship, and the blame attributed to the service provider.