The importance of social comparison in perceived justice during the service recovery process

Abstract

Purpose

This study aims to analyse how consumers' perceptions of justice in a service recovery scenario vary, not only due to the company's actions but also due to the comparisons they make with the experiences of other consumers.

Design/methodology/approach

Based on justice theory, social comparison theory and referent cognitions theory, this study describes an eight-scenario experiment with better or worse interactional, procedural and distributive justice (better/worse interactional justice given to other consumers) × 2 (better/worse procedural justice given to other consumers) × 2 (better/worse distributive justice given to other consumers).

Findings

First, consumers' perceptions of interactional, procedural and distributive justice vary based on the comparisons they draw with other consumers' experiences. Second, the results confirmed that interactional justice has a moderating effect on procedural justice, whereas procedural justice does not significantly moderate distributive justice.

Originality/value

First, based on justice theory, social comparison theory and referent cognitions theory, we focus on the influence of the treatment received by other consumers on the consumer's perceived justice in the same service recovery situation. Second, it is proposed that the three justice dimensions follow a defined sequence through the service recovery phases. Third, to the best of the authors' knowledge, this study is the first to propose a multistage model in which some justice dimensions influence other justice dimensions.

研究目的

: 本研究擬探討在服務補救的處境裡, 消費者對公平的看法不但會受公司的行動所影響, 同時也會因他們與其他消費者的經驗作比較而有所改變。

研究設計/方法/理念

: 本研究根據正義理論、社會比較理論和參照認知理論, 描述一個涵蓋八個處境的實驗, 實驗包含更好的或更差的互動的、程序上的和分配性的公平 (給予其他消費者更好的/更差的互動公平) × 2(給予其他消費者更好的/更差的程序上的公平) × 2 (給予其他消費者更好的/更差的分配性的公平)。

研究結果

: 研究結果顯示, 消費者對互動的、程序上的和分配性公平的看法, 是會根據他們與其他消費者的體驗所作的比較而有所改變; 研究結果亦確認了互動的公平對程序上的公平會有調節作用, 而程序上的公平對分配性的公平則沒有顯著的調節作用。

研究的原創性

: 首先, 我們根據正義理論、社會比較理論和參照認知理論, 把研究焦點放在於相同的服務補救情景中, 其他消費者受到的待遇, 如何影響消費者自身的認知公平; 另外, 我們建議, 這三個公平維度, 在各個服務補救階段裡, 均會跟隨一個清晰的次序。最後, 就研究人員所知, 本研究為首個提出一個公平維度互為影響的多階段模型的研究。

Citation

Aguilar-Rojas, Ó., Fandos-Herrera, C. and Pérez-Rueda, A. (2024), "The importance of social comparison in perceived justice during the service recovery process", European Journal of Management and Business Economics, Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/EJMBE-02-2023-0056

Introduction

The fierce competition in the service sector and high rates of customer loss after service failures have increased the attention paid to service recovery as a means of retaining customers (La and Choi, 2019; Sánchez-García and Curras-Perez, 2020). Service failure arises in situations where businesses do not meet their customers' expectations (Simões-Coelho et al., 2023). This will happen, sooner or later, often with very negative results (La and Choi, 2019). When a service failure occurs, the probability of losing the customer is high, and the reputation of the company may be seriously affected (Grégoire et al., 2018). Specifically, 86% of consumers leave brands they were once loyal to after only two to three bad customer service experiences, 63% leave because of poor customer experience and 49% stated that, during the previous 12 months, they had left a company they had been loyal to for that reason (Emplifi, 2022). The cost of poor customer service ranges from $75 billion to $1.6 trillion per year (McCain, 2023). To combat this situation companies have developed service recovery strategies to restore customer satisfaction, mainly through process-related treatments (e.g. explanations) and monetary compensation (Ahmad et al., 2023). Previous studies have shown that, when customers are compensated for service failures by receiving service better than they expected, they usually rate their satisfaction with companies and their services higher than prior to the failure (Cheng et al., 2015). One of the most common service failure research perspectives is the evaluation of customers' responses to failures based on their perceptions of the justice they receive (La and Choi, 2019). Justice theory proposes that customers' satisfaction increases when they experience “fair” recovery (Grégoire et al., 2018). However, some studies have suggested that customers can affect one another in a service recovery scenario (Albrecht et al., 2019), because they are social comparers (Ludwig et al., 2017). Social comparison research has aroused special interest in the social sciences since Sherif (1936) showed that two people facing the same situation develop a point of reference through a process of mutual social influence (Buunk and Gibbons, 2007). However, relatively little research has examined how consumers perceive the outcome of system recovery processes when they compare their experiences with those of other consumers (Bonifield and Cole, 2008; Chen et al., 2023).

This study makes three contributions. Based on justice theory (Rawls, 1971), social comparison theory (Festinger, 1954) and referent cognitions theory (Folger, 1986), we examine the influence of the treatment received by other consumers on the consumer's perceived justice in the same service recovery situation. Second, it is proposed that the three justice dimensions follow a defined sequence during the service recovery phases (Murphy et al., 2015). Third, to the best of the authors' knowledge, this study is the first to propose a multistage model in which some justice dimensions influence other justice dimensions.

Theoretical background and hypotheses development

Perceived justice

Kelley and Davis (1994) defined service recovery as the process by which firms attempt to rectify a service delivery failure. Service recovery includes all the activities/responses that service providers perform/make to repair losses experienced by customers (Grönroos, 1998). Service research has adopted justice theory as the dominant theoretical framework (Huang, 2011). Justice has been said to be related to evaluations, based on moral criteria, of how the individual is treated by others (persons and entities) (Furby, 1986). Tax et al. (1998) proposed that perceived justice is a complex, tri-dimensional concept (interactional, procedural and distributive justice). Interactional justice relates to how the consumer is treated during a complaints process and includes elements such as the courtesy and kindness exhibited by company staff, empathy perceived, efforts made to resolve and willingness to provide reasons for the failure, for example, by an airline when a flight is cancelled (Schoefer and Ennew, 2005). Procedural justice, as the term suggests, relates to the perceived fairness of the processes applied by the company to recover the failure. It includes aspects such as delays in the processing of the complaint, response time to the complaint and the company's flexibility in adapting to the consumer's needs (Blodgett et al., 1997). Distributive justice is the degree to which consumers feel they have been treated fairly, specifically, what economic compensation the company offers for the failure. Distributive justice may result in refunds, discounts or other forms of compensation (Maxham and Netemeyer, 2002).

The previous literature has found that perceived justice has a critical influence on the development of consumers' evaluative judgements (Schoefer and Ennew, 2005), influences behavioural reactions (Colquitt et al., 2006), creates trust and evokes positive emotions (La and Choi, 2012) and satisfaction (Sánchez-García and Curras-Perez, 2020). Specifically, consumers' satisfaction with recovery service is significantly affected by procedural and interactional justice (Mohd-Any et al., 2019). Mathew et al. (2020) showed that perceived justice had a significant moderating effect on the relationship between e-service recovery quality and e-service recovery satisfaction. In a novel approach we suggest that the three perceived justice dimensions unfold in a particular order. Our sequential model is consistent with suggestions made by other authors in different research fields, such as organisational management, who have proposed that interactional justice is a precursor of procedural and distributive justice (Cohen-Charash and Spector, 2001; Tran et al., 2021). In the present study, it is expected that individuals affected by a service failure will primarily attribute any associated (in)justice to the person in the company responsible for the service at that moment. In fact, previous literature has affirmed that interactional justice relates to how individuals treat and communicate with, each other in the place where the problem occurred (Bies and Moag, 1986). Thus, the recovery process starts with the consumer's first contact with the company's customer service department. This initial contact, which is directly connected to the interpersonal treatment people receive during recovery procedures, is encompassed within the interactional dimension. Second, social psychology research has gradually shifted its emphasis from focussing solely on the outcomes of reward allocation (distributive justice) to a focus on an earlier stage in the process, that is, the company's flexibility in adapting to consumers' needs (Blodgett et al., 1997; Wood et al., 2020), which has been described as an important dimension of their perceptions of justice (Thibaut and Walker, 1975). This process, related to the procedure through which the complaint is handled, is likely to unfold after the consumer has filed the complaint and before the company has resolved it. Finally, the consumer focuses on compensation, that is, the distributive justice dimension. As previously noted, distributive justice relates to the consumer's perception of justice in the outcome of the process, so it seems logical to place it at the end of the sequence.

In addition, like many of the personal evaluations that humans make, perceived justice can be strongly influenced by the individual's way of thinking, perceptions and personal experiences (LaFave, 2008). In this regard, humans assess the experiences of their peers to evaluate their own experiences. Regardless of whether consumers have had much prior experience of any particular event/incident, the experiences of their peers will help them understand what has happened. However, little research has delved into the influence of other consumers on the consumer's experience of the same system failure (Albrecht et al., 2019).

Social comparisons

Previous studies have shown that the presence of other consumers affects the individual's behaviours (Albrecht et al., 2019). For example, Viglia and Abrate (2014) found that consumers are more influenced by social comparisons, for instance, price information given to them by friends, than they are when the information source is anonymous; in the latter case they are likely to lower their reference price (to be closer to average past prices). Social comparison theory argues that individuals evaluate their opinions and abilities by comparing them with those of other, similar individuals (Festinger, 1954). In the justice context, Greenberg (1982) argued that people perceive injustice when they receive dissimilar treatment, procedures or economic benefits to those received by others. Thus, individuals use social comparisons to associate with others, learn from others, self-assess against others (Taylor and Lobel, 1989) and to make sense of their own outcomes (Moore, 2007). This process, as it helps to reduce uncertainty, is a fundamental aspect of human experience (Suls and Wheeler, 2000) and has been explored in service recovery research. Indeed, social comparisons are an inevitable part of social intercourse (Brown et al., 2007) because, when people interact with others, consciously or unconsciously they compare themselves with these other people (Wheeler and Miyake, 1992). Steinhoff and Palmatier (2016) confirmed that comparing oneself to someone “worse” produces positive feelings and comparing oneself with someone “better” produces negative feelings, for example, in the context of hotels and flying.

Referent cognitions theory (Folger, 1986) recognises the role of comparisons in perceived justice and proposes that procedures that affect oneself and others, are taken into account. Comparisons are important for establishing justice perceptions because they allow consumers to evaluate whether they received what they deserved (Chen et al., 2023). Furthermore, in the consumer's evaluation of whether a deal is fair, knowing what others obtained is often more important than the procedural justice (s)he himself/herself received (Bonifield and Cole, 2008). Consumers use this information to assess justice and satisfaction (Chen et al., 2023). Therefore, we propose the following hypotheses:

H1.

The consumer's perception of the interactional justice received by other consumers inversely influences his/her perceptions of the interactional justice s(he) has received.

H2.

The consumer's perception of the procedural justice received by other consumers inversely influences his/her perceptions of the procedural justice s(he) has received.

H3.

The consumer's perception of the distributive justice received by other consumers inversely influences his/her perceptions of the distributive justice s(he) has received.

Many service encounters occur on what is known as the organisational frontline. Unlike other frontline interactions (e.g. in the sales/purchase process, which may develop over many interactions), on the service failure recovery frontline employees play a critical role in the provision of quality service (Carlzon, 1987; Lindsey-Hall et al., 2023). The first few moments of the interaction are very critical and have a great impact on how the customer perceives the whole service (Lin et al., 2016). Previous studies have concluded that, during customer-company face-to-face interactions, the customer's initial impressions influence subsequent interactions and can, ultimately, influence customer outcomes (Anwar, 2023). Thus, on the basis that justice perceptions are based on the consumer's perceptions of the gains and losses (s)he experiences in a relationship with a provider (Kwon and Jang, 2012) and that equity theory (Adams, 1965) proposes that his/her perceptions during a recovery process take into account the company's previous efforts to recover the situation, it is proposed that the consumer's perceptions of the justice (s)he receives may be formed by his/her perceptions of the justice (s)he received in previous justice dimensions. Therefore, the following hypotheses are proposed:

H4.

The consumer's perceptions of the interactional justice received by other consumers moderates the relationship between his/her perceptions of the procedural justice given to those consumers and his/her perceptions of the procedural justice (s)he has received, such that:

The consumer's perceptions of the procedural justice received by other consumers will have a greater influence on his/her perceptions of the procedural justice s(he) has received when the interactional justice received by others is worse (H4a) than when it is better (H4b).

H5.

The consumer's perceptions of the interactional justice received by other consumers moderates the relationship between his/her perceptions of the distributive justice given to those consumers and his/her perceptions of the distributive justice (s)he has received, such that:

The consumer's perception of the distributive justice received by other consumers will have a greater influence on his/her perceptions of the distributive justice s(he) has received when the interactional justice received by others is worse (H5a) than when it is better (H5b).

Finally, companies overemphasise distributive justice (the customer received the promised result) whilst neglecting procedural justice (Michel et al., 2009). Thus, companies tend to assume that the most important aspect of service failure recovery is monetary compensation, a form of distributive justice. However, the majority of the reasons given by consumers for their low levels of satisfaction after a service failure relate to procedural justice, overly complicated toll-free numbers, user-unfriendly websites and outsourced customer care contact centres (NCRS, 2020). This leads us to suggest that consumers' perceptions of the procedural justice they receive may influence their subsequent justice perceptions.

H6.

The consumer's perceptions of the procedural justice received by other consumers moderates the relationship between his/her perceptions of the distributive justice given to those consumers and his/her perceptions of the distributive justice (s)he has received, such that:

The consumer's perceptions of the distributive justice received by other consumers will have a greater influence on his/her perceptions of the distributive justice s(he) has received when the procedural justice received by others is worse (H6a) than when it is better (H6b).

Figure 1 depicts the proposed conceptual model.

Research methodology

To guarantee the validity of the data and the representativeness of the sample, the specialised market research company Netquest was hired. The company, at the end of 2019, used a consumer panel to randomly assign the participants to the different scenarios. The participants were remunerated. The vast majority of the panellists had taken part in previous studies and their prior participation had been considered satisfactory by the company.

Pre-test study

Following Harris et al. (2006), a pre-test was conducted to assess the realism of the experimental setting and scenarios. Some 51 respondents participated in the pre-test, 56% women, 44% men, from 18 to 62 years old. Following receipt of the experimental instructions, the participants were randomly assigned to one of the eight experimental conditions. Subsequently, the participants were thanked, debriefed and asked to answer a short survey. We measured the realism of the scenarios (see Appendix 2) through four items, with 7-point bipolar scales, adapted from Collie et al. (2002). An example item is: “I believe that situations like this happen in real life” (α = 0.73***). The participants reported that they perceived the scenarios as being realistic (Mean = 5.93, Standard Deviation = 1.05).

Main study

The airline sector has been growing. In 2022, it gained 64% in turnover over the previous year and is forecast to grow by 28.3% in 2023 (Statista, 2023). The experiment examined a recovery process after a service failure, that is, a baggage loss incident. This scenario was selected because baggage loss is one of the main service failures in the sector (Mohd-Any et al., 2019).

To ensure the subjects could identify with the proposed scenario, a condition of participation was that they must have taken at least one flight in the previous six months. To test the research hypotheses, Netquest recruited 259 Spain-based panellists. Table 1 shows the socio-demographic characteristics of the sample.

The participants were first told that the questionnaire was an academic-focused opinion survey about service recovery, and they were then asked to answer questions about the research framework's variables. First, the survey described a baggage loss incident. Thereafter, the participants were randomly assigned to a condition in a 2 (better/worse interactional justice given to other consumers) × 2 (better/worse procedural justice given to other consumers) × 2 (better/worse distributive justice given to other consumers) design. At least 30 participants were used for each condition. As Table 2 shows, the researchers were particularly interested in ensuring that the groups consisted of similar numbers.

The experiment described the following situation: a passenger arrives by plane at an airport, but his/her check-in luggage did not appear on the carousel. After submitting his/her complaint, (s)he sees that another passenger on the same flight has had the same problem and is also making a complaint. At that point the participant is randomly assigned to one of the eight possible scenarios (interpersonal, procedural and distributive justice), outlined in Appendix 1. As the central proposition of social comparison theory (Festinger, 1954) is the “similarity hypothesis”, which argues that individuals tend to compare themselves with similar people in similar situations, the traveller/participant then had to compare himself/herself with someone who was travelling on the same flight, has the same problem and is even staying in the same hotel. The participants were asked to rate, on a scale of 1–7, their perceptions of interpersonal, procedural and distributive justice associated with the way the airline resolved the failure. Finally, they were asked to provide socio-demographic information.

Measurement

The measurement scales for the questionnaire were adopted from previous literature (see Appendix 2). We measured interactional justice using four items on 7-point bipolar scales, adapted from Karatepe (2006), for example, “The hotel employee was courteous” (α = 0.89). Procedural justice was measured using four items on 7-point bipolar scales, based on DeWitt et al. (2008), for example, “The policies and procedures the firm had in place were adequate for addressing my concerns” (α = 0.91). Distributive justice was measured using three items on 7-point bipolar scales, also adapted from DeWitt et al. (2008), for example, “The outcome I received was fair” (α = 0.92).

Convergent validity was verified as the factor loading of each indicator was found to be above 0.5 and significant at the 0.01 level (Steenkamp and Van Trijp, 1991), and the statistical values of the AVEs were greater than 0.5 (Fornell and Larcker, 1981). Similarly, composite reliability exceeded the minimum recommended value of 0.65 (Bagozzi and Yi, 1988). Finally, to determine discriminant validity, we compared the square roots of the AVEs (the values on the diagonal, in bold) with the inter-construct correlations (values below the diagonal); to ensure discriminant validity, the on-diagonal values should be higher (Fornell and Larcker, 1981). The results from these analyses were satisfactory, as shown in Table 3.

Results

To test the effects proposed in the hypotheses we conducted three 2 × 2 analyses of variance (ANOVA), using IBM SPSS Statistics v.26 software. The results showed that the interactional justice given to other consumers during the recovery process inversely influenced the participants' perceptions of interactional justice they received (F (1, 257) = 6.59, p = < 0.05), supporting H1. More specifically, the results showed that consumers perceived higher levels of interactional justice if others had been treated worse (M_{Other’sWorseInteractionalJustice} = 4.36; M_{Other’sBetterInteractionalJustice} = 3.93). Similarly, the procedural justice given to other consumers inversely influenced the respondents' procedural justice perceptions (F (1, 257) = 8.44, p = < 0.01), supporting H2. Again, the participants perceived higher levels of procedural justice if others had been treated worse (M_{Other’sWorseProceduralJustice} = 3.43; M_{Other’sBetterProceduralJustice} = 2.94). Finally, supporting H3, the distributive justice given to other consumers inversely influenced the participants' distributive justice perceptions (F (1, 257) = 25.43, p = < 0.01). In line with the previous results, the participants perceived higher levels of distributive justice if others had been treated worse (M_{Other’sWorseDistributiveJustice} = 3.50; M_{Other’sDistributiveJustice} = 2.58). As Table 4 shows, we checked for the presence of heteroscedasticity. First, we performed Levene's test; this tests the null hypothesis that the error variance of the dependent variable is equal between groups. The results were not significant for the interactional justice and procedural justice variables, so it was concluded that the variance of the groups was equal, and thus, an analysis of variance (ANOVA) could be performed. As for the distributive justice variable, although a statistically significant p-value appeared in the analysis of variance, Levene's test showed that heteroscedasticity is present. To remedy this heteroscedasticity problem, Welch's test was applied; this test is more robust in these cases (Norusis, 2011). The levels of statistical significance observed for distributive justice using Welch's tests were less than 0.05, therefore, the means of all groups are equal, allowing an analysis of variance (ANOVA) to be performed.

The overall interaction effects show that consumers' perceptions of procedural justice vary when they believe that other consumers have received better, or worse, interpersonal justice (confirming H4), (F (1, 255) = 11.01, p = < 0.01). Contrary to our expectations, when other consumers received worse interactional justice, the participants' procedural justice perceptions increased, but not significantly, supporting H4a (M_{Other'sWorseInteractional−Other'sWorseProceduralJustice} = 3.18; M_{Other'sWorseInteractional}−M_{Other'sBetterProceduralJustice} = 3.22; t(132) = −0.169, p > 0.10). On the other hand, supporting H4b, when other consumers received better interactional justice, the participants' procedural justice perceptions decreased (M_{Other'sBetterInteractional−Other'sWorseProceduralJustice} = 3.67; M_{Other'sBetterInteractional−Other'sBetterBetterProceduralJustice} = 2.60; t(123) = 4.48, p < 0.01); see Figure 2.

With respect to H5, it was found that the moderating effects of the interactional justice received by other consumers on distributive justice perceptions was not significant (F (1, 255) = 1.33, p = > 0.10). However, the results indicated that when other consumers received better interactional justice, the procedural justice perceived by the participants increased (M_{Other'sBetterInteractionalJustice−Other'sWorseDistributiveJustice} = 3.70, M_{Other'sBetterInteractionalJustice−Other'sBetterDistributiveJsutice} = 2.56; t(123) = 4.19, p < 0.01); (M_{Other'sWorseInteractionalJustice−Other'sWorseDistributiveJustice} = 3.31, M_{Other'sWorseInteractionalJustice−Other'sBetterDistributiveJustice} = 2.60; t(132) = 2.92, p < 0.05); see Figure 3.

Similarly, it was shown, as proposed in H6, that the procedural justice received by other consumers moderated the participants' distributive justice perceptions, but the differences were not significant (F (1, 255) = 0.17, p = > 0.10). (M_{Other'sWorseProceduralJustice−Other'sWorseDistributiveJustice} = 3.57, M_{Other'sWorseProceduralJsutice−Other'sBetterDistributiveJustice} = 2.57; t(125) = 3.53, p < 0.01). (M_{Other'sBetterProceduralJsutice−Other'sWorseDistributiveJustice} = 3.45, M_{Other'sBetterProceduralJustice−Other'sBetterDistributiveJustice} = 2.60; t(130) = 3.60, p < 0.01); see Figure 4.

Discussion and implications

There is a need for an in-depth study of the different strategies companies employ for customer recovery after service failures, how they are implemented and how they are experienced by the consumer (Ahmad et al., 2023; Chen et al., 2023). In this sense, the theory of justice has been widely examined and has emerged as one of the main theoretical service recovery frameworks (Peinkofer et al., 2022). The present study proposes that the three justice dimensions follow a particular sequence during the recovery process after a service failure. This research is based on the fact that humans are social beings by nature and assess the experiences of others in identical/similar situations to evaluate their own. Specifically, this study is based on the idea that, when faced with a service failure, the customer uses peer comparison to analyse and evaluate the treatment provided to him/her. Thus, when a customer observes that, for the same service failure, (s)he is being treated worse than other customers, (s)he may perceive that (s)he is being treated unfairly. Similarly, if the customer perceives that (s)he is being treated better than other customers, (s)he may perceive greater fairness in the service recovery process. The results showed that the interactional, procedural and distributive justice provided to other consumers during the recovery process inversely influenced the participants' perceptions of the interactional, procedural and distributive justice they received. Interestingly, the results confirmed that consumers perceive higher levels of justice when they believe that others have been treated worse than they have and, conversely, they perceive lower levels of justice when they believe that others have been treated better. Regarding moderation effects, the results suggest that the interactional justice given to one consumer influences other consumers' perceptions of procedural justice. However, the results did not show that interactional justice significantly influenced distributive justice, or that procedural justice influenced distributive justice.

Theoretical implications

Taking as bases social comparison theory (Festinger, 1954) and referent cognitions theory (Folger, 1986), this study analyses how the treatment given to some consumers during service recovery incidents influences other consumers' perceptions of justice. Social influence has been widely examined in social psychology (Gerber et al., 2018); however, few studies has analysed that influence in a service recovery context (Bonifield and Cole, 2008; Ludwig et al., 2017). Previous studies have focused on spontaneous and relatively automatic, comparisons; for example, the social comparisons that some consumers might draw based on the information posted on other consumers' Facebook pages (Morry et al., 2018). In contrast, our research is based on the social comparisons that some consumers may draw based on actions taken by companies, that is, we examine whether how companies behave towards some consumers is used by other consumers as material through which to make comparisons.

First, in line with Ludwig et al. (2017), this study demonstrated that, after a service failure, consumers' perceptions of the justice they receive varies based on how the company acts towards other consumers. Specifically, consumers perceive higher levels of justice when they believe that other consumers have been treated worse than them and, conversely, they perceive lower levels of justice when they believe that other consumers have been treated better. These findings confirm the importance of social comparison in service recovery (Bonifield and Cole, 2008) and are consistent with the results of previous studies that found that comparisons with “worse” individuals evoke positive feelings and with “better” individuals evoke negative feelings (Steinhoff and Palmatier, 2016).

Second, the present study proposes that perceived justice is a multistage model in which one justice type influences others. Taking a novel approach, this study posits that the three justice dimensions follow a specific sequence, that is, first the interactional, next the procedural and, finally, the distributive. The results suggest that the interactional justice given to one consumer influences other consumers' perceptions of procedural justice. This conclusion is consistent with previous research that has indicated that, if customers attribute employees' behaviours to organisations, interactional justice might influence procedural justice evaluations (Tyler and Bies, 1990). Therefore, in line with Anwar (2023), it is proposed that, during the initial stage of the recovery process, how a company treats some consumers affects other consumers' perceptions of procedural justice. As the results show, the procedural and distributive justice mean values were low. This could be because these are the most difficult justice dimensions to address (La and Choi, 2019). Consumers have their own vision of how complaints should be handled and are never fully satisfied with companies' protocols; similarly, they are rarely satisfied with the compensation they are offered and may believe that they deserved more. The results did not show that interactional justice significantly influenced distributive justice, or that procedural justice influenced distributive justice. These results are in line with previous research that suggested that the compensation obtained after a service failure is the most important issue in service recovery (Ahmad et al., 2023). Thus, the compensation obtained by consumers seems to be decisive in their perceptions of justice in the service recovery process. Consistent with Ahmad et al. (2023), customers are more satisfied with the recovery process if they perceive that distributive recovery is fair; thus, they should be compensated fairly, or at least compensated in a way that will cover their losses.

Managerial implications

Identifying the main factors that lead consumers to abandon or switch service providers can help companies design more effective strategies to prevent them from leaving and to win back those who have already left (Anwar, 2023; Sánchez-García and Curras-Perez, 2020). Brun et al. (2017) emphasised that, following service failures, providers should bear in mind that the recovery process must resolve the important issues as quickly and as efficiently as possible. Managers must understand that the treatment given to some consumers during service recovery influences the justice perceptions of other clients (Chen et al., 2023). While, sometimes, the consumer is unaware of the justice received by others for similar service failures, this information is easily accessible from anywhere, and at any time, via the Internet. For example, due to the proliferation of internet-connected devices, consumers have access to information about the attention paid to, the processes used with and the compensation obtained by others who have suffered similar service failures. For example, massive flight cancellations can occur, and companies such as Ryanair, British Airways and Iberia have faced thousands of customer complaints requesting the refund of the cost of flight tickets. The Ryanair Twitter account features users' posts about their service failure experiences, for example: “I've been waiting since March for the return of my flights cancelled due to the pandemic and I still haven't received anything”; “The link does not work or when it works it does not recognize the reservation code”; “customer service ask me to fill out an application to reject a voucher (which I already rejected at the time of cancellation)” (Knowles, 2020). However, users have also posted positive comments about the management of refunds; “the company has handled returns very quickly, much more than Vueling or other airlines”; “So far I have never had problems with Ryanair”; “100% refund in less than 24 h: I have to say something positive, due to the passing of a close family member, they refunded 100% of the tickets with no charge and in less than 24 h. Deep down, they have a heart” (Trustpilot, 2020, 2023). Each of these experiences is related to one of the three dimensions of justice theory and are clear examples of how users compare their experiences with those of others. Su et al. (2021) suggested that companies must manage how consumers communicate their dissatisfaction with service failures via social networks. To do so, companies need to design transparent customer recovery plans that address the different situations that can arise. Frontline customer services should master these plans and follow action protocols designed to make customers feel they are being treated fairly. Transparency could lead companies to strengthen their commitment to quality and provide a strategic advantage. Consumers should be told in their initial contacts with companies what process they will need to go through and what compensation they are likely to receive. This strategy would increase the customer's peace of mind (Siqueira et al., 2020). Thus, empathy, as it influences trust, should be an important characteristic possessed by frontline employees (Flavian et al., 2019): but companies should be cautious about the promises they make, as it has been shown that they influence consumers' expectations and, if the company does not live up to them, this can increase the consumer's dissatisfaction (Simões-Coelho et al., 2023). Our findings are also consistent with Honora et al. (2023), who highlighted the fundamental importance of service employees in any recovery strategy. Service companies should carefully select professionals for frontline positions and provide them with continuous training to improve their behavioural skills and with coping strategies, particularly for handling service recovery interactions (Honora et al., 2023). The airline industry might follow the example of the financial sector, which uses personal managers in its online banking. These frontline employees are available every day, at almost any time, to address any questions and solve problems, both by phone and through the online banking channel.

Limitations and suggestions for future research

This research proposes a multistage model of perceived justice and examines the role of social comparison in perceptions of the three dimensions of justice. This issue has been very little explored by marketing scholars and managers. However, our approach has several limitations that suggest other interesting research avenues. First, only one study was undertaken, and the data were collected four years ago, in Spain. Although a single study design is commonly accepted in service recovery research (Bagherzadeh et al., 2020) and justice perceptions research (Blodgett et al., 1997; La and Choi, 2019), other studies proposing cross-cultural and cross-country differences should be tested, and the study might be replicated in another service context to confirm its results (de Juana-Espinosa and Rakowska, 2018). However, this is an exploratory study, and further research is needed to confirm the results. Second, further research is needed to better understand the sequence which the justice dimensions follow and the relationships proposed in this study. Although we believe that the suggested sequence is the most common, there may be situations where the sequence may differ. For example, where a company detects a service failure before its customers detect it and decides to refund them part of the amount charged without contacting them. In this case, distributive justice would precede the previous dimensions, thus altering the sequence and relationships. In fact, if distributive justice has the greatest weight in consumers' perceptions of justice, when it is manifested before the other dimensions it will surely have a great influence on their subsequent perceptions of interactional and procedural justice. Third, consumers' personal traits could affect the degree of influence that a company's attitude towards other customers has on them. For example, an individual's patience level could affect his/her justice perceptions during a service recovery process. Furthermore, future analyses might contrast the influence of the internal (personality) and the external (environment, familiarity with the other consumer(s) and importance of the service) motivations of consumers to compare themselves with other consumers. Fourth, although 49% of the participants said that they had previously experienced a service failure, to safeguard their privacy, the data were presented in an aggregated form. In addition, Netquest subjected the data to an anonymisation process that eliminated values that could be used to identify any individual. However, it would be very interesting to analyse the previous experience variable and identify whether there are differences between customers who have suffered a service failure and those who have not. Fifth, some sectors commonly use robots or chatbots as their first customer service contact. Given that our results have shown that interactional justice is crucially important, it would be very interesting to examine how new technologies affect justice perceptions. In conclusion, to generalise our results this research could be replicated in other service sectors, such as hotels, car rental and retail stores.

Figures

Figure 1

The proposed conceptual model

Figure 2

Moderating effect of the interactional justice received by other consumers on the relationship between the consumer's perception of the procedural justice given to those consumers and the consumer's perception of the procedural justice (s)he has received

Figure 3

Moderating effect of the interactional justice received by other consumers on the relationship between the consumer's perception of the distributive justice given to those consumers and the consumer's perception of the distributive justice (s)he has received

Figure 4

Moderating effect of the procedural justice received by other consumers on the relationship between the consumer's perception of the distributive justice given to those consumers and the consumer's perception of the distributive justice (s)he has received

Table 1

Sample demographic characteristics

Variable		N
Gender	Men	137
	Female	122
Marital status	Married/coupled	148
	Single	105
	Divorced/separated	6
Occupation	Housewife	23
	Unemployed	15
	Employed	113
	Student	106
	Retired	2
Studies	Primary	21
	High School	58
	College	180
Age	≥18, <22	59
	≥22, <30	63
	≥30, <49	66
	≤49	71

Source(s): Authors' elaboration

Table 2

Sample distribution (by scenarios)

Interpersonal	Procedural	Distributive	N
Worse	Worse	Worse	31
Worse	Worse	Better	32
Worse	Better	Worse	36
Worse	Better	Better	35
Better	Worse	Worse	32
Better	Worse	Better	32
Better	Better	Worse	31
Better	Better	Better	30

Source(s): Authors' elaboration

Table 3

Composite reliability and convergent and discriminant validity

	CR	AVE	Interactional justice	Procedural justice	Distributive justice
Interactional justice	0.908	0.713	0.869
Procedural justice	0.903	0.702	0.440	0.885
Distributive justice	0.925	0.804	0.262	0.475	0.931

Note(s): The diagonal elements (in italic) are the square roots of the AVEs (variance shared between the constructs and their measures). Off-diagonal elements are the inter-construct correlations

Source(s): Authors' elaboration

Table 4

Homoscedasticity and heteroscedasticity test

	Interactional justice	Procedural justice	Distributive justice
Levene's test	0.66	0.24	0.07
Welch's test	–	–	0.00

Source(s): Authors' elaboration

Declaration of interest statement: The authors certify that this article is the authors' original work. The work is submitted only to this journal and has not been previously published. The authors also warrant that the paper contains no harmful statements, does not infringe on the rights or privacy of others, or contain material that might cause harm or injury.

Appendix

The supplementary material for this article can be found online.