The Future of Complaint Handling: The Effects of Perceived Justice on Post-Complaint Satisfaction

(1)

1

The Future of Complaint Handling: The Effects of

Perceived Justice on Post-Complaint Satisfaction

Master Dissertation

MSc. Dual Award International Business Management & Marketing

Newcastle University Business School

&

University of Groningen, Faculty of Economics and Business

Renaldo Boom Supervision

140088865 / S2003899 Markus Blut

r.w.boom@newcastle.ac.uk Newcastle University

r.w.boom@student.rug.nl Henk Ritsema

’t Spiek 27, 9411 KX Beilen Rijksuniversiteit Groningen

(2)

2

List of tables

Table 1.1: Online Complaint Handling Literature………....18

Table 4.1: Manipulations Check...29

Table 4.2: Perceived Justice Effects…...30

Table 4.3: Moderator Influence…………...32

Table 4.4: Influence on Consequences...…...33

Table 5.1: Perceived Justice Correlation Comparison.………34

(5)

5

Abstract

With an increasing amount of complaints handled being voiced via social media nowadays, a better understanding of customer preferences regarding the handling of those complaints is necessary. The purpose of this study is to assess the influence and relative importance of distributive, procedural and interactional justice on post-complaint satisfaction after service failure, for social media as a complaint platform in specific. Regarding this relationship, the amount of time a customer spends on social media is tested as a moderator variable. Furthermore, the subsequent relationship with word of mouth intent and overall satisfaction with the firm is addressed in order to analyze the importance of post-complaint satisfaction. A quantitative research design was chosen to collect data by the use of an online questionnaire in which participants were asked questions regarding an experimental scenario about service failure. With a 3x2 scenario manipulation, six scenarios were developed to find differences in reactions to service recovery actions regarding distributive, procedural and interactional justice. The analysis showed that the effect of distributive justice on complaint handling satisfaction was stronger than that of procedural and interactional justice, while all relationships show to be significantly positive. These findings imply that organizations have to take good care of complaints voiced via social media, since the handling greatly influences customers’ satisfaction. In addition, social media usage showed to have a marginally significant, positive effect on this relationship with distributional and interactional justice. This moderating effect implies that the relationship between complaint handling outcomes/interactions and post-complaint satisfaction is stronger for customers with a high social media usage. In return, complaint handling satisfaction was found to have a positive effect on its consequences; word of mouth intent and overall satisfaction. Complaint handling satisfaction with a high degree of fairness for all justice dimensions is therefore important for management, since word of mouth and overall customer satisfactions are important indicators of healthy organizations

(6)

6

Acknowledgements

(7)

7

1. Introduction

1.1 Introduction

The purpose of the first chapter is to provide an overview of the study and to provide the reader with an explanation of the research topic and goals. Background and Study Rationale are presented in section 1.2. Section 1.3: Research Questions, describes the research’s main and sub-questions while section 1.4 describes the Research Objectives. Subsequently, the

Research Methodology is presented in section 1.5 after which section 1.6 concludes the

chapter with an overview of the Dissertation Structure.

1.2 Background and Study Rationale

(8)

8 from the opportunity of customer retention (Huppertz, 2007; Fornell and Wernerfelt, 1988; Strauss and Hill, 2001).

Nowadays an increasing amount of complaints is being voiced via social media, not only by young people, Millenials, but also by an increasing group of 35+ users (Clark, 2013; Kaplan and Haenlein, 2010). Kietzmann et al. (2011) and Kaplan and Haenlein (2010) argue that most firms are generally not comfortable yet in dealing with social media and feel lost in a jungle of possibilities. Meanwhile, Mangold and Faulds (2009) state that organizations dealing with complaints via social media are struggling since there still is little to no literature to guide them regarding the handling of complaints on this relatively new complaint channel. At the same time however, the article argues that organizations are forced to act upon this development, since the balance of power has shifted towards the customer with the public visibility of complaints.

(9)

9

1.3 Research Questions

Related to the brief introduction above, the main research question of the study is the following:

Compared to offline channels, how important is complaint handling satisfaction for organizations dealing with complaints on social media and what role does perceived justice play in this?

This research question allows the researcher to develop a better understanding of social media complaint handling and how justice aspects are perceived by customers in a social media environment. Perceived justice, as used in the research question, points towards the evaluation of fairness in the complaint handling process, a concept that will be further explained in section 2.3. In addition, the answer to this main question will shed light on the unique characteristics of complaint handling and satisfaction drivers in the context of social media as opposed to traditional complaint channels.

In order to be able to fully answer the research question, the following sub-questions need to be addressed:

What is the relationship between perceived justice and satisfaction with complaint handling in the social media context?

Does the intensity of social media usage influence the significance of the relationship between perceived justice and complaint handling?

How does satisfaction with complaint handling influence overall firm satisfaction and word-of-mouth intent in a social media context?

(10)

10 order to test the complexity of the relationship between perceived justice and complaint handling satisfaction in the social media context. This is important since the relationship can have moderators specifically for the social media context that might help explain the answer to the main question. Finally, the importance of complaint satisfaction in a social media context will be analyzed to test the importance of satisfaction with complaint handling in the social media context. This is relevant to investigate since contrary to the offline setting, post-complaint satisfaction could be more or less influential on both word of mouth intent and the overall satisfaction of customers, customer metrics that have high value for management to know.

1.4 Research Objectives

This study aims to fill several research gaps of existing literature by looking into social media as a complaint channel and studying the effects of differences in fairness of the justice dimensions. As identified in the meta-study by Orsingher et al. (2010), the online context for complaint handling is an area of interest since little research has been conducted with regard to perceived justice. The article also suggests researching the impact of perceived justice in the online environment; therefore this dissertation aims for the social media platforms in specific. In addition, by filling this gap in current literature, this study also attempts to provide management with literature and knowledge regarding complaints on social media, as Mangold and Faulds (2009) identified a lack of managerial insights regarding this topic. By looking for differences between traditional complaint channels studied by Mattila and Wirtz (2004) and social media as a complaint channel described by Clark (2013), this study tries to identify unique characteristics that alter perceived justice evaluations and satisfaction. With these insights, knowledge can be developed to help management with social media operations and the development of effective complaint handling strategies for the social media context.

1.5 Research Methodology

(11)

11 quantitative data will be analyzed with IBM SPSS 20 to find the relations between variables used in the study. The results will be compared to the comprised results of 60 studies about perceived justice and satisfaction with recovery from the meta-study of Orsingher et al. (2010), which is focused on perceived justice in the offline context. More details of the research methodology are presented in Chapter 3, where the research strategy and data collection method will be discussed extensively.

1.6 Dissertation Structure

The subsequent parts of the study are structured as follows; Chapter 2; literature review, starts with a deep analysis of the theoretical background of the field of study. It reviews the existing base of knowledge and analysis the theory used for this research. Chapter 3; Research

methodology, extensively describes the research strategy and elaborates on the scenario

choices. Also the measurement of the variables and conceptual model are described. After data collection, findings are analyzed in Chapter 4: Results, in which the relationships between variables are examined and presented to answer the research questions posed in section 1.3. Next, the results are compared to existing literature to provide managerial insights and knowledge in Chapter 5: Discussion. Lastly in Chapter 6: Conclusion and

recommendations, conclusions about the study are drawn, recommendations about future

research are given and limitations of this study will be described.

2. Literature Review

(12)

12 describes the uniqueness of social media as a complaint channel. Finally, Section 2.5:

Summary will provide a short overview of the literature review.

2.2 Customer Complaint Behaviour

In organizations and business, mistakes are inevitable since when working with people, flaws that lead to dissatisfaction of the customer are always existent (Karatepe, 2006). Previous research has linked this dissatisfaction to the expectations and subsequently perceived reality of the customer (Gilly and Gelb, 1982). Thereafter, dissatisfaction causes an increase in post-purchase behaviour, with complaint behaviour in specific (Mattila and Wirtz, 2004). However, not solely the degree of dissatisfaction influences whether or not a customer will complain, also other factors such as perceived barriers to complain, controllability of the problem, prior knowledge, the cost/benefit involved and other customer-related issues are involved in this process (Day, 1984).

(13)

13 behaviour significantly, followed by altered ways of handling complaints by organizations. Consequently, social media as complaint channel seems to fit in between existing complaint channels for customers because of the unique possibilities due to the channel’s technological advancements (Clark, 2013). The possibilities of interactive (phone and face-to-face) and remote (letter and e-mail) channels are combined in a single channel, social media, that will be further described in section 2.4.

2.3 Perceived Justice

Justice theory by Adams (1963) is widely used in complaint management literature and is a generally accepted theory regarding the concept of justice. As the theory is a powerful mechanism in forecasting satisfaction with service recovery (Kim et al., 2009), it is widely used in the existing literature regarding complaint handling (Karande et al., 2007; Smith et al., 1999; Tax et al., 1998; Sparks and Fredline, 2007). The promise of justice is of importance for service recovery since it often is complex for customers to make a proper evaluation of the service recovery beforehand (Maxham and Netemeyer, 2002). In addition, customers in general have expectations regarding the service recovery such as compensation, empathy, acknowledgement and apology that influence whether or not the customer perceives the actual complaint handling as just (Johnston & Mehra, 2002). With the findings of Clemmer and Schneider (1996), studies started to investigate perceived justice in the form of three dimensions: distributive justice, procedural justice and interactional justice (Karatepe, 2006; Kim et al., 2009; Blodgett et al., 1997; Maxham and Netemeyer, 2002). The distributive component reflects the perceived outcomes, the procedural component focuses on the perceived fairness of the procedures and policies employed to produce the outcome, while the interactional component represents the personal treatment people enjoy during the handling of the complaint (Blodgett et al., 1997). The following sessions will discuss each of the justice components in more detail and describe the application of each in the complaint handling context.

2.3.1 Distributive Justice

(14)

14 inputs and the outcomes of the process. An exchange is perceived as fair when the customer sees the outcomes of the recovery process as being in proportion to their inputs in the exchange (Blodgett et al., 1997; Maxham and Netemeyer, 2002). Among possible other outcomes of service recovery processes where customers seek redress, Kim et al (2009) provide monetary refunds, discounts, coupons, free-of-charge adjustments and substitutions as examples. The perceptions of distributive justice therefore differ per individual complainant, as the same absolute redress can produce differences in perceived justice among complainants. Since distributive justice focuses on the final outcome of the service recovery process, distributive justice is more important for complaints that seek actual redress compared to complaints in which anger venting is the customer’s intention. Finally, as the effect of distributive justice on satisfaction with complaint handling has been confirmed by Tax et al. (1998) and Goodwin and Ross (1992), this research attempts to reflect on this relationship in a social media perspective.

2.3.2 Procedural Justice

(15)

15 2.3.3 Interactional Justice

The third component of perceived justice, interactional justice refers to the degree in which customers perceive to have been handled fairly by the employees of the firm handling the complaint with regard to human interactions (Maxham and Netemeyer, 2002; Blodgett et al., 1997). For example, during the process of service recovery, customers can perceive the personal interaction with employees as rude or as respectful and with courtesy. As the service recovery process relies heavily on interpersonal interactions, even in an online environment such as that of a social media platform, interactional justice is found to significantly influence evaluations of service recovery by Smith et al. (1999). In addition, a strong positive relationship is found between interactional justice and satisfaction with complaint handling, which is also expected to be found in the context of social media complaints (Tax et al., 1998). The elements that are brought in association with interactional justice are apology, explanation, interest, friendliness and effort (Kim et al., 2009; Blodgett et al., 1997). In the study towards satisfaction with complaints by e-mail, Strauss and Hill (2001) found that an apology, signing with the employee’s name and dealing with the specific problem were especially important for complaint handling by e-mail since there only is online contact. These findings have been taking into account in the argumentation with respect to the expectations of the relationship between perceived interactional justice and complaint handling satisfaction.

2.4 Social Media

(16)

16 communication between customer and company has changed according to Clark (2013). A shift in the balance of power has put the customer in charge of the communication. This shift is mainly caused by the transparency of social media, making it possible that negative experiences go viral, a possibility that brings risks, such as harming the reputation of the company (DiStaso & Bortree, 2012; Einwiller and Steilen, 2013; Hogreve et al., 2013). Due to this shift in the balance of power, organizations are forced to respond to customers choosing social media as channel to voice their complaint. In addition, the technology used by social media has enabled fast and easy two-way communication that has increased the possibility for online articulations (Kalsbeek, 2012). However, not all forms of online

consumer-generated media by customers can be classified as social media activity as is

suggested by some literature (Mangold and Faulds, 2009). The demarcation of social media has been given multiple efforts that might not all be complete since some literature focuses too much on applications, while others come with a very broad definition (Treem and Leonardi, 2012). Therefore several explanations of social media are taken together in an attempt to make clear what will be studied in this dissertation in the best possible way.

To do so, a distinction needs to be made between social media and related terms as Web 2.0 and User Generated Content, which are often used in conjunction (Kaplan and Haenlein, 2010). Web 2.0 refers to the possibility of collaboratively and continuously modifying and creating content and online applications. User Generated Content, on the other hand, consists of all forms of online media created by end-users that are publicly available, including the use of social media. Social media fits in with these terms in such a way that they allow the creation of User Generated Content by the use of internet-based application that build on the foundation of Web 2.0 (Kaplan and Haenlein, 2010). This specification, together with the use of the term social media by Treem and Leonardi (2012) and Kalsbeek (2012) brings us to the use of the definition of Jones (2009), who states that social media are “a category of online

media where people are talking, participating, sharing, networking and bookmarking online. Most social media services encourage discussion, feedback, voting, comments and sharing of information from all interested parties”.

(17)

17 hand, receive complaints from customers on a day to day basis (KLM, 2014a). These social media platforms have established themselves as part of our daily lives and physical or mental costs to interact with firms via these platforms are close to zero (Hogreve et al., 2013). Furthermore, an important characteristic of social media is the visibility and publicly availability of the content that is put on the platform. This visibility probably is the most important difference with the other existing complaint channels (face-to-face, phone, letter and e-mail) as identified by Mattila and Wirtz (2004). It is effortless for users to locate information and knowledge others share on a platform as Facebook due to the technologies involved. Treem and Leonardi (2012) suggest that this differentiates social media from traditional complaint channels and creates unique consequences. Following the argumentation of Goetzinger (2007) organizations should encourage the use of social media as a complaint channel, since barriers to complain are lower as customers already use the platform on a daily basis. Since a lower barrier to complaint causes more customers to voice complaints (Fornell and Wernerfelt, 1987), more information will be received by the organization and more customers can be retained. Among very few researchers that look at social media as a complaint channel, Pinto and Mansfield (2011) made use of Facebook in their research as Facebook is the most widely used social media platform as well as the platform most likely to be used for complaining behaviour because of the interaction possibilities (eBIZMBA, 2014).

(18)

18

Author

Year of

publication

Research Focus

Hong and Lee 2005 Internet and its possibilities have created a new complaint channel for communication of negative experiences. This brings with it a large set of possibilities and challenges for organizations. Mangold and

Faulds

2009 Social media conversations between customers occur beyond the control of managers. Organizations must learn how to shape these conversations and examples of methods are given. Park and Lee 2009 The effects of electronic word of mouth are found

to be greater for negative word of mouth, versus positive word of mouth. A greater effect is also found for established websites versus un-established websites and for experience goods rather than search goods

Kaplan and Haenlein

2010 In what way can businesses make profitable use of social media channels and recognize the most important channels. 10 advices are provided for organizations dealing with social media

Bajic and Lyons 2011 Software companies make use of customers to improve their products, while company size, transparency, software deployment and the number of social media tools employed influence the effectiveness of the companies.

Kietzmann et al. 2011 By many businesses, social media is not understood correctly and left aside. Social media can be defined by seven building blocks as presented with their implications.

Pinto and Mansfield

2011 Millenials are more likely to use social media as a complaint channel, especially when using social media a lot in their daily lives. Service providers are more likely to receive complaints on social media.

Clark 2013 Contrary to traditional complaint channels as described by Mattila and Wirtz (2004), social media are semi-interactive as a complaint channel, allowing fast anger venting and putting the customer into a powerful position.

(19)

19

2.5 Summary

Through this literature review, it was made clear that customer complaint behaviour is a broad concept, while this study only focuses on a narrow selection of consequences of customer complaint behaviour; public complaints, specifically voiced via social media channels. In addition, satisfaction with complaint handling is linked to perceived justice, consisting of three elements. Comprised of distributive justice, procedural justice and interactional justice, this concept is widely studied in the field of complaint satisfaction and describes the several aspects about which customers decide upon their satisfaction outcomes. Furthermore, the differences between social media as a complaint channel and other traditional complaint channels has been emphasized on, after a general description and lay-out of the possibilities and challenges of social media were given. These constructs help this study by shaping the barriers of the concepts it operates with and thereby the aims of the research have been narrowed down. With the boundaries of the dissertation discussed, the next chapter will present and explain the research methodology in detail. In this chapter, the literature review will be taken into account by the use of a data collection method that suits the concepts involved, while the theory will provide guidance in developing the measurement instrument.

3. Research Methodology

After clarifying the concepts used in this study in chapter 2, this chapter focuses on the explanation of the research methodology that comes back to the research aims and objectives of this study. The quantitative design of this study will be described in section 3.2: Research

Strategy, and will explain the need of a quantitative approach. Next, the conceptual model and

(20)

20

3.2 Research Strategy

This deductive study derives its ideas from existing theory and applies these in a new context, namely that of social media. Considering the research philosophies involved in the research and the strategy to come to findings, the epistemological orientation is that of objectivism. Although there are expectations and previous findings in slightly different contexts, this study aims for an objective view to come to the best possible answers without searching for the expected outcomes. However, subjectivism is taken into account regarding the research itself, since the concept of perceived justice is a subjective one as participants in the study can see the same as different due to their own perspective. This is also linked to what Bryman and Bell (2007) call constructionism, which refers to the process in which people constantly change and thus alter their perceptions of reality. In this study, people will be asked to their perceptions of reality regarding complaint handling on social media and most likely different perceptions will be found because of the before mentioned subjectivism.

In order to study the concepts described, a questionnaire has been developed to quantitatively test the relationships as will be presented in the conceptual model in the next section. With the concepts having been widely studied already, there is no immediate need to question the existence of relationships between the variables in a social media context. However, the significance and relative importance of the various variables are expected to change because of the context alteration. The aim of the questionnaire is to find answers to the main and sub-questions, test expectations and in the end come to conclusions and suggestions for further research. The use of an online quantitative focused questionnaire gives the opportunity to study a relatively large amount of participants in a small amount of time; with time being one of the major limitations of this study.

(21)

21 of both justice and services (Hocutt et al., 2006; Goodwin and Ross, 1992). The results of the questionnaire will be analyzed by the use of various tests in SPSS that best fit the aim of the study and will provide the reader with answers to the posed questions.

3.3 Conceptual Model

The basis of the conceptual model is derived from Orsingher et al. (2010), a meta-study taking together the findings of many studies towards the effect of perceived justice on satisfaction with complaint handling. As has been found by combining many studies regarding this effect, it appears that the strongest relationship lies between distributive justice and satisfaction with complaint handling. Interactional justice has been found to have the second strongest relationship with complaint handling satisfaction. While procedural justice scores the lowest positive correlation with complaint handling, the correlation is still significant enough for the relationship to be true. As can be seen in the first part of figure 1.1, the relationship between the three dimensions of perceived justice and satisfaction with complaint handling (presented as SATCOM in figure 1.1) are all expected to be positive in the context of social media complaints; the same as they have been found to be in the offline context. The significance of the relationships however, is expected to be different than for the offline context, as will be described more extensively below. For distributive justice a correlation with a lower significance is expected, while the relationship for procedural justice and interactional justice is expected to be stronger than the offline context. The relative importance of the justice dimensions however is expected to remain the same as for the offline context, as is described in hypotheses 4 and 5. Furthermore, as will be explained with hypothesis 6, social media usage is expected to have an effect on the relationship between perceived justice and satisfaction with complaint handling since more experience with social media is expected to bring more expectations regarding the complaint handling on the social media platforms. This moderating effect is shown in part 2 of the conceptual model in figure 1.1, indicating that social media usage is expected to influence all three relationships.

(22)

22 several important concepts for organizations. The meta-study by Orsingher et al. (2010) finds mixed results for the relationship between satisfaction with complaint handling and overall satisfaction with the organization, but a positive relationship between satisfaction with complaint handling and word of mouth intentions. This study will have another look at these relationships for the social media environment to test whether these influences are in place for complaints in such a different context. As in the line of Orsingher et al. (2010) it is expected that a positive relationship between satisfaction with complaint handling and word of mouth intention will be found. In addition, the relationship between complaint handling satisfaction and overall satisfaction is also expected to be positive. These last two relationships are shown in the third part of the conceptual model in figure 1.1.

Figure 1.1: Conceptual Model

(23)

23 3.3.1 Conceptual Model Part 1

H1: Distributive justice is positively related to satisfaction with complaint handling on social media.

H2: Procedural justice is positively related to satisfaction with complaint handling on social media.

H3: Interactional justice is positively related to satisfaction with complaint handling on social media.

As these relationships have found to be positive for traditional complaint channels (Orsingher et al., 2010; Maxham and Netemeyer, 2002), it is expected that these relationships will also be found to be positive for the social media environment. For the specific context of social media complaint handling, customers are not expected to express a totally different reaction to the perception of fairness than for the offline context, resulting in the expectation of a positive effect of distributive, procedural and interactional justice with complaint handling satisfaction. However, differences in the significance, and therefore importance of the three justice aspects with the offline context are expected. First of all, distributive justice is expected to be less important for complainers, because of the informal nature of social media as a complaint channel, which therefore most likely lowers customers’ expectations regarding the complaint resolution. Secondly, the relationship of procedural justice with complaint handling satisfaction is expected to be more significant for the social media context than offline. This expectation is based on the high interactive nature of social media as a complaint channel and therefore, a higher focus on speed of resolution is expected from customers. Lastly, as the social media has a high interactive nature, interactional justice is also expected to have a higher significance since personal interaction is an important aspect.

H4: In a social media context distributive justice is more important for satisfaction with complaint handling relative to the other justice perceptions.

H5: In a social media context interactional justice is more important for satisfaction with complaint handling relative to procedural justice.

(24)

24 satisfaction with complaint handling, as reflected in hypothesis 4. Furthermore, as hypothesis 5 states, interactional justice is expected to be more important for satisfaction with complaint handling than procedural justice. Although this has also been concluded by Orsingher et al. (2010), it is expected that there will be less of a difference between the effects of these dimensions for the social media context.

3.3.2 Conceptual Model Part 2

H6a: Social media usage negatively moderates the effect of distributive justice on satisfaction with complaint handling.

H6b: Social media usage negatively moderates the effect of procedural justice on satisfaction with complaint handling.

H6c: Social media usage negatively moderates the effect of interactional justice on satisfaction with complaint handling.

By the use of hypotheses 6a, 6b and 6c, a possible moderator variable will be tested. This is done to test whether the amount of time a person spends on social media every day alters the relationships as presented in respectively hypothesis 1, 2 and 3. It is expected that the results will show that participants with a higher social media usage pattern will have higher expectations regarding the fairness of complaint handling. These higher expectations will most likely be found back in the complaint satisfaction results with lower scores for the complaint satisfaction variable. Therefore, social media usage is expected to have a negative moderating effect on the relationship between perceived justice and satisfaction with complaint handling.

3.3.3 Conceptual Model Part 3

H7: Satisfaction with complaint handling on social media is positively related to word of mouth intentions

H8: Satisfaction with complaint handling on social media is positively related to overall firm satisfaction

(25)

25 respect to the offline environment, the relationship between post-complaint satisfaction and word of mouth intentions is expected to be more significant. These expectations are based on the fact that social media are also extensively used for spreading word of mouth (Mangold and Faulds, 2009). By handling a complaint on the same platform where spreading word of mouth is a mouse-click away, customers are expected to have higher intentions of spreading word of mouth when satisfied.

3.4 Data Collection Method

As stated before, to collect the needed data for the study, a questionnaire regarding complaint handling on social media will be employed. The questionnaire records general data, such as age, gender and nationality, to be able to identify the general characteristics of what group of people has participated in the study. Furthermore, control variables are added in the questionnaire in the form of attitude towards complaining and past flying experience. Attitude towards complaining is important since some people have a higher tendency to voice complaints when facing dissatisfaction with organizations than others (Yuksel et al., 2006). Therefore, participants with a low complaining tendency might influence the outcomes, since the questionnaire uses scenarios instead of real-life experiences and the participants will not be able to reflect on the scenario correctly. Furthermore, past flying experiences is added as a control variable to verify the ability of participants to be able to reflect properly on a scenario that regards a complaint about a flying service. In addition, before the scenario, participants will also be asked about their social media usage for the moderator variable and about their general satisfaction with KLM as a company to give the option for comparison with results after the complaint scenario.

(26)

26 dissatisfaction is more likely to occur due to service-failure than due to product-failure, because of the people involved. Therefore, people are more likely to identify the failure of the service and understand the dissatisfaction as described in the scenario. Furthermore, Pinto and Mansfield (2011) have studied the likelihood of complaints on Facebook and find that service organizations are more likely to receive complaints. Moreover, larger organizations are also more likely to receive complaints via social media since people expect those organizations to be present on Facebook (Pinto and Mansfield, 2011). Since KLM is a large multinational organization, and has a well-established presence on social media, the organization is a good example to use in the scenario (KLM, 2014a; KLM, 2014b). Furthermore, following the argumentation by Park and Lee (2009), the likelihood of spreading of word of mouth about a product is larger for organizations selling experience goods, rather than search goods. The reason for this is that experience goods have to be owned and used before feedback can be given, while search goods can be evaluated before purchase. Since flights are an experience good, it is likely that customers are resolute to spread word of mouth about the product, something that will have to be taken into account in this research. The social media platform used in the scenario is Facebook, firstly because of the way Facebook is organized. Facebook’s design gives the possibility to have a “wall” as an organization on which customers can post complaints, after which a conversation can be started between the complainer and organization to resolve the complaint. Secondly, Facebook is also the social media platform with most active users, far exceeding the number of users of the platform in second place; Twitter (eBIZMBA, 2014).

After reading the scenario, participants are asked to answer questions regarding the scenario and what they think about the complaint handling process and their satisfaction regarding the scenario. Participants will firstly be asked about their perceptions of one of the three justice dimensions, depending on their scenario. The choice for one justice dimension per participant is made based on the fact that they will then judge one scenario only and no bias will be formed due to experience with another scenario. After answering the questions regarding perceived justice, participants will receive questions about their satisfaction with the complaint handling, word of mouth intentions and overall satisfaction with KLM as an organization. Each of these constructs is measured by the use of multiple statements and a 7 point Likert-scale for participants to rate their perceptions from strongly disagree to strongly agree for the statements presented. The questions involved in the study can be found in

(27)

27 the operationalization of the different concepts involved in the study. The operationalization of the concepts is derived and adapted from the articles by Blodgett et al. (1997), Maxham and Netemeyer (2002), Kim et al. (2009) and Pinto and Mansfield (2011). These articles have studied the same concepts in different complaint contexts and therefore they have been adapted to fit the social media scenarios used in this study.

Sampling of the study has been done by posting a link to the questionnaire on Facebook and asking people to fill in the questionnaire and share the link in order to get as much response as possible. In this way, snowball sampling was encouraged, to not only spread the link to first relationships but spread the link further on Facebook.

3.5 Validity and Reliability

First of all, to maximize validity, control variables in the form of flying experience and attitude towards complaining are added to be able to see whether participants are sufficiently able to judge the scenario. If participants would have no experience with flying for example, their judgment of the scenario cannot be seen as valid, which is not good for the results of the study. Secondly, the questionnaire is spread via Facebook and therefore it is most likely that only people with experience on social media participate; in that way a sample of people from the right population are studied to maximize validation. In addition, to maximize reliability of the study and to let the questionnaire measure what it is supposed to measure, several aspects of the study have been optimized. First of all, a pre-test of the different scenarios regarding the justice dimensions has been done to check whether the scenarios reflected a high/low situation of that dimension. After the pre-test several characteristics of the scenarios were altered to optimize the credibility. Furthermore, the measurements regarding perceived justice, satisfaction with complaint handling, word of mouth intention and overall satisfaction are conducted with at least 3 statements, to minimize distortion due to the use of specific words. All these statements of the different measurements were adapted from two or three existing studies regarding these concepts and have therefore proven to measure the intended.

3.7 Ethical Issues

(28)

28 in the study are fully aware of their participation and do so on a participatory basis, there are no ethical issues disregarded. Their consent with participating and acceptance for the use of their information is given by participating in the study. The introduction of the questionnaire states that by participating, people accept and have understood the consent they give to the researcher. Furthermore, since the study does not incorporate sensitive information, there are no risks involved with the participation in the study. Participants are also informed about the fact that their data will only be used for this study and are given the opportunity to send an e-mail regarding questions. Lastly, at the end of the study, the possibility to leave an e-e-mail address is given, to provide the opportunity to receive the dissertation once it is finished. The introduction and closing of the questionnaire can be found in Appendix III; Introduction and

Closing.

4. Results

This chapter provides the analysis of the data obtained with the questionnaire as discussed in chapter 3. First, section 4.2: Participants, gives a general overview of the sample of people that have participated in the study. Then, section 4.3: Preliminary Analysis generally describes the variable outcomes and presents the manipulation check done for the scenarios. Afterwards, section 4.4: Perceived Justice and Complaint Satisfaction will discuss the findings regarding the relationship between perceived justice and satisfaction with complaint handling. The analysis of Social Media Usage as moderating variable is discussed in section 4.5, where the differences in high and low social media usage are displayed. Then, Section 4.6:

Complaint Satisfaction Effects elaborates on the relationships of complaint satisfaction with

word of mouth intention and overall satisfaction. A Summary is given in the last section of this chapter, section 4.7.

4.2 Participants

(29)

29 sample population. With respect to the control variables, only 4 cases had never travelled by plane, while 10 cases reported the likelihood of complaining very low. Lastly, participants’ social media usage was recorded to test the moderator variable, and shows an average of 108 minutes social media usage a day, with the lowest value being 0 and the highest 420 minutes. This social media usage falls above the average of the study by Pinto and Mansfield (2011), being 75 minutes on average. However, this difference can probably be predicted by the further development of social media over time and the recruitment of participants via Facebook.

4.3 Preliminary analysis

With the use of SPSS, the questionnaire data was transformed into different variables. First the scores were calculated by taking together the scores for the separate statements per variable and comprising them into a mean score to use in the further analysis. This has been done to increase reliability of the study by the use of multiple statements to lower the chance of mistakes by the use of specific words. Then a manipulation check was conducted to check whether the different scenarios represent a different degree of fairness due to the manipulation. In table 4.1, below, the mean value of justice of the high and low scenario can be found for the different justice dimensions. As can be seen in the table, the high scenario for all three dimensions report higher perceived justice on average than the low scenario; showing that the manipulations have succeeded in the goal of receiving different reactions to the different scenarios.

Justice Dimension: High Low

Distributive Justice 4,42 2,44

Procedural Justice 5,45 4,85

Interactional Justice 5,34 5,17

Table 4.1: Manipulations Check

(30)

30 confidence level. The manipulation of Interactional Justice however is not significant, as it shows a low difference between mean values.

4.4 Perceived Justice and Complaint Satisfaction

With the calculated means of the statements regarding the perceived justice dimensions and satisfaction with complaint handling, the relationship between these variables will be calculated. The data is analyzed with the use of a bivariate correlation analysis in order to find an answer to the first sub-question as stated in the first chapter. This analysis will be done separately for all justice dimensions of which the full results are shown in Appendix V;

Perceived Justice Correlations, while the correlations are shown in table 4.2 below. The

reason for this separate analysis mainly is the nature of the data, not allowing for a regression including all variables, since participants have only answered questions regarding one justice dimension; as explained in section 3.4. Therefore, no correlations between justice dimensions can be calculated and issues with multicollinearity shall not be controlled for. The analysis shows that a significant positive relationship exists between distributional justice and satisfaction with complaint handling on a 5% confidence level. This result supports H1, and the findings confirm the expectations as derived from existing literature. In addition, the relationship between procedural justice and satisfaction with complaint handling also shows to be positive on a 5% confidence level as expected, finding support for H2. Lastly, the effect of interactional justice on complaint handling satisfaction is found to be significant on a 5% confidence level. Therefore, H3 is supported by the results and a significant relationship between all justice dimensions and satisfaction with complaint handling has been found in the social media context.

Justice Dimension: Complaint Handling Satisfaction Including Control Variables Distributive Justice 0,943 0,945 Procedural Justice 0,821 0,822 Interactional Justice 0,670 0,668

Table 4.2: Perceived Justice Effects

(31)

31

Appendix V; Perceived Justice and table 4.2, these control variables only alter the results for

the distributive, procedural and interactional justice dimensions by a fraction. For distributive and procedural justice the significance of the relationships slightly increases, while for the interactional justice dimension, the relationship with complaint handling satisfaction becomes a little less significant. Therefore, the control variables flight experience and complaint attitude do not significantly alter the findings with regard to the first three hypotheses. They also do not clarify much of the relationships existing between the discussed variables. Because of this low effect of the control variables, they have not been taken into account in the further analyses of this study.

In addition, H4 and H5 regard the relative importance of the different justice dimensions in their relationship with complaint handling satisfaction. As table 4.2 shows, distributive justice has a more significant relationship with complaint handling satisfaction than the other dimensions and therefore H4 is supported. H5 on the other hand is not supported by the findings, since interactional justice has a less significant relationship with complaint handling satisfaction than procedural justice does. Since procedural justice has a correlation with complaint handling satisfaction of 0,822, while interactional justice has a correlation of 0,668, it should be noted that procedural justice appears to be more influential on complaint handling satisfaction than interactional justice; findings that prove against the expectations expressed in H5.

Further discussion of the results, and a comparison with the findings from an offline context as reported in Orsingher et al. (2010), will follow in chapter 5: Discussion.

4.5 Social Media Usage

(32)

32 variable and social media usage variable had to be centralized, subtracting the mean of the variable from the value for each case. Then, these centralized variables were multiplied to compute the interaction variable, which will be used as the moderator in the analysis.

As the SPSS output, that can be found in Appendix VI; Social Media Usage, suggests; the results of the linear regression analysis, summarized in table 4.3, do not support the H6 hypotheses. With a two-tailed significance of respectively 0,231 and 0,212 for distributive justice and interactional justice, the moderating effect of social media usage is only marginally significant on a one-tailed 10% confidence level (taking half of the two-tailed significance). The marginally significant moderating effect however, appears to be positive, rather than the expected negative effect of the moderator variable. The two-tailed significance of the moderating effect for procedural justice is 0,961, resulting in a one-tailed significance level of 0,481. Therefore, the moderating effect of social media usage is clearly not supported for procedural justice. It can be said that H6a, b and c are not supported, while marginal support has been found for a reversed H6a and H6c. Discussion about the rejection of the H6 hypotheses will take place in the next chapter, where the implications of these findings will be elaborated on.

Justice Dimension: Two-tailed Moderating Significance

Distributive Justice 0,231

Procedural Justice 0,961

Interactional Justice 0,212

Table 2.3: Moderator Influence

4.6 Complaint Satisfaction Consequences

In order to test the importance of complaint handling satisfaction for variables that are more influential in organizations and answer sub-question three; a correlation analysis of complaint handling satisfaction, word of mouth intention and overall satisfaction has been done. The one-sided correlation analysis reveals that the positive effects of complaint handling satisfaction with both word of mouth intentions and overall satisfaction are significant on a 5% confidence level. The results are summarized in table 4.4 and fully shown in Appendix VII;

(33)

33 of these findings and a discussion about the importance of complaint handling satisfaction for its consequences, relative to the offline environment will be provided in the next chapter.

Word of Mouth Intent Overall Satisfaction Satisfaction with complaint Handling 0,728 0,644

Table 4.4: Influence on Consequences

4.7 Summary

In this chapter, several expectations have been tested by the use of hypotheses and SPSS analyses of the data collected with the questionnaire. The results have shown support for hypotheses H1 to H4, H7 and H8, while H5, H6a, H6b and H6c have been rejected. These results imply that all three justice dimensions have a significant positive relationship with complaint handling satisfaction and that interactional justice is the least important dimension, after distributional and procedural justice. In addition, social media usage marginally moderates the effects of both distributive and interactional justice in a positive direction, while no evidence is found for moderation of the procedural dimension. Lastly, complaint handling satisfaction is found to have a significant positive relationship with word of mouth intention and overall satisfaction with the company. In chapter 5: Discussion, the findings and implications of the findings will be further discussed.

5. Discussion

In this chapter, the results of the study will be discussed with regard to the expectations and actual outcomes, followed by a comparison of the findings with the research from Orsingher et al. (2010). Starting, section 5.2 will discuss the hypotheses regarding Perceived Justice, with explanations about the correlations and the relative importance of the three justice perceptions. Next, section 5.3; Moderating Effects, will discuss the findings regarding the moderating effect of social media usage. Furthermore, the Word of Mouth Intent and Overall

Satisfaction as consequences of complaint handling satisfaction will be covered in section 5.4.

(34)

34

5.2 Perceived Justice

An experimental study was conducted to find out the effects of the three justice dimensions and their relationship with complaint handling satisfaction in a social media context. The hypotheses predicted that all three dimensions would positively relate to complaint handling satisfaction and that procedural justice would be least important in the social media context, after respectively interactional justice and distributive justice. The findings show that the expectations regarding the positive relations between the perceived fairness of complaint handling and complaint handling satisfaction are confirmed. When comparing these results with findings from an offline setting, as Orsingher et al. (2010) report, it should be noted that Orsingher et al. have combined 60 independent studies to come to composed results. The composed results of the meta-study most likely make the findings of the article more accurate than a single study. The comparison of the correlation scores is presented in table 5.1, showing the differences between the offline and social media context.

Context: Offline (Orsingher et al., 2010) Social Media

Distributional Justice 0,450 0,943

Procedural Justice 0,090 0,821

Interactional Justice 0,250 0,670

Table 5.1: Perceived Justice Correlation Comparison

(35)

35 Among the different justice dimensions, distributional justice remains to have the strongest correlation with complaint handling satisfaction. This suggests that the number one priority of customers is getting a fair redress from the company when complaining. The ability of the firm to restore service failure therefore proves to have the most influence on post-complaint satisfaction. Most likely, customers are most influenced in their post-complaint satisfaction by direct outcomes of the complaint, since this is the ultimate goal of redress seeking complaints. Contrary to the offline environment however, is the second strongest correlation with complaint handling satisfaction; procedural justice. While procedural justice has a relatively small correlation with complaint handling satisfaction in the offline context, the results of this study show quite a high correlation for social media complaints. Since speed, flexibility and employee empowerment majorly make up the procedural justice dimension, it can be noted that the perceived fairness accounted to these elements are more important in the social media context than offline. As Asur and Huberman (2010) state, ease of use and speed of social media are important characteristics that most likely also shape the expectations and influence the perceived fairness of the process. Therefore, it is highly advisable for organizations to provide timely responses to complaints, especially for complaints voiced via social media. As people carry their social media with them all day, expectations regarding timeliness are most likely higher as the results show. The difference between findings of offline and social media context is the smallest for the interactional justice dimension, with still quite some difference between studies to be explained. It is likely that on social media, the interactional aspect is more important than complaints via letter or e-mail for example, which are regarded as remote complaint platforms (Mattila and Wirtz, 2004). In addition, Strauss and Hill (2001) note that fairness in the interaction is already more important for e-mail than for older-fashioned complaining mechanisms such as letter or phone. Following their line of reasoning, personal interaction has become even a more important aspect on social media platforms, with Facebook used as example in this study. Therefore, the friendliness of the message, the willingness to help and putting a single contact person on a complaint are important managerial implications that the findings indicate.

(36)

36 interactional aspects. As the findings contradict the expectations stated earlier in this study; speed of resolution, flexibility and employee empowerment appear to be more influential than friendliness, interpersonal treatment and apology. While Clark (2013) identifies social media as a complaint platform that is semi-interactive, it would be reasonable to expect interactional aspects to be more important as opposed to remote complaint platforms. Strauss and Hill for example (2001) note that for e-mail as a complaint platform, interactional aspects are already important, it would be reasonable to find such outcomes in this study as well. Arguably, procedural aspects have also seen a significant increase in correlation with complaint satisfaction due to the nature of social media as a complaint channel; apparently a much larger increase than the rise in correlation for interactional fairness.

5.3 Moderating effects

With the analysis to find out the moderating effect of social media usage on the three perceived justice dimensions, the results show marginally significant effects on the relationship of the distributive and interactional dimension with complaint handling satisfaction. These effects however, do not confirm with the expectations of being negative, since the t-value is positive. No support was found for a moderating effect on the relationship between procedural justice and complaint handling satisfaction. These findings imply that more social media usage results in a higher correlation for both perceived distributive justice and perceived interactional justice with post-complaint satisfaction. This higher correlation is possibly explained by a significant correlation between age and social media usage. This correlation shows that younger participants have a significantly higher social media usage. Since there is no existing research towards the effect of age on perceived justice and satisfaction with complaint handling, no theoretical causes for the findings can be stated. For future research however, a suggestion would be to look at this relationship and see what implications age differences give for complaint handling satisfaction.

(37)

37 media, this study regarded social media usage as possibly influencing people’s perception of fairness, an expectation that has not been fully confirmed.

The fact that the positive moderating effect has no significance for procedural justice as opposed to the other two dimensions means that there is no difference in expectations regarding speed and employee empowerment between participants with a high or low social media usage pattern. This conclusion in combination with the higher correlation of procedural justice with complaint handling satisfaction as found before would imply that for all participants, either high or low social media usage, speed and employee empowerment are more important on social media than when complaining offline.

For distributive justice and interactional justice, it is marginally significant that fairness characteristics of these dimensions are relatively more important for people with high social media usage compared to people with low social media usage. This would mean for companies that customers with a pattern of high social media usage, meeting expectations of the complaint is more important, since a higher influence on satisfaction is present.

5.4 Word of Mouth Intent and Overall Satisfaction

(38)

38 explained by the sort of good used in the analysis. Park and Lee (2009) describe that the product type moderates the likelihood of spreading word of mouth. Following their argumentation, the use of KLM has increased the likelihood of spreading word of mouth, compared to other, non-experience focused goods. Since Orsingher et al. (2010) make use of 60+ studies; a part of the difference might be attributed to the type of product used in the scenario presented to the participant.

The second consequence of complaint handling satisfaction, overall satisfaction also shows a positive correlation with complaint handling satisfaction. This suggests that well-managed complaints result in good attitudes towards the company in general. A comparison of the findings of this study with Orsingher et al. (2010) with regard to this correlation shows that a much more significant correlation has been found in the context of social media. With a correlation of 0,644 found in this study, the non-significant correlation of -0,04 in Orsingher et al. (2010) can be said to be really different since it basically states that the relationship is non-existent in the offline context. A part of this difference might be explained by the type of company chosen in this research, with a service selling product, well known by its good service and product quality, a bias can possibly have played a role in the answers of participants. Furthermore, before reading the scenarios, participants also ranked their satisfaction with KLM in general, with which the overall satisfaction has a correlation. The possibility that good or bad complaint handling matters less than the correlation suggests is present. This is possible since participants have not actually experienced the relevant complaint, due to the experimental nature of the research. This bias due to the nature of the study can have altered the correlation in a positive way. However, the correlation with overall satisfaction can also be higher for the social media context because of the online nature of social media, due to which the complaint handling satisfaction has a higher correlation with overall satisfaction. This, nonetheless, should be further researched with qualitative measures in order to be able to properly state how and why this would be the case.

Context: Offline (Orsingher et al., 2010) Social Media

Word of Mouth Intent 0,530 0,728

Overall Satisfaction -0,04 0,644

(39)

39

5.5 Summary

The discussion of the results has revealed that the relationships between perceived justice and complaint handling satisfaction and subsequently its consequences are complex and change in different research settings. While most assumptions are supported, others were rejected as well as reversely supported. Perceived justice has been found to have greater correlations for the social media context than for previously researched contexts. Different was also the relative importance of the perceived justice dimensions, with distributive justice being most important for complaint handling satisfaction, then procedural justice and lastly interactional justice. At the same time, social media usage has been found to have a marginally positive influence the correlation between both distributive and interactional perceived justice and satisfaction with complaint handling. Consequences of post-complaint satisfaction were also found to have higher correlations when complaining via social media than what Orsingher et al. (2010) have found. Word of mouth intention, like in the offline setting had the highest correlation, while overall satisfaction was also found to have a significant correlation with complaint handling satisfaction. It can be said that due to the findings and discussion it has become clear in this dissertation that handling complaints on social media is definitely different than handling complaints via other channels. The findings show that organizations must adapt their complaint handling services to this relatively new complaint channel to maximize the effectiveness of their efforts. To elaborate on these findings, Chapter 6;

Conclusions and Recommendations, will provide more detail about the managerial

implications and consequences of the findings and will give guidance for future research.

6. Conclusions and Recommendations

(40)

40

6.2 Summary of the research

In order to explore the perceived justice – complaint handling satisfaction relationship in a social media context and to test the possible moderating effect of social media usage, a quantitative research design was chosen. By the use of an experimental questionnaire, the researcher was able to test assumptions for the social media context that are already known for traditional complaint platforms. Six different scenarios were used to in order to find correlations for distributive, procedural and interactional justice with complaint handling satisfaction. Next, the influence of complaint handling satisfaction on word of mouth intentions and overall satisfaction with the company were analyzed. The analyses and discussion unveiled that, compared to findings for the offline context; a much stronger correlation exists for all perceived justice dimensions with complaint handling satisfaction. In addition, social media usage as a moderator for this relationship has not received full support, and is only marginally significant in its positive influence on the relationship of distributive and interactional justice with complaint handling satisfaction. Furthermore, relative to the offline context, complaint handling satisfaction has been found to have higher correlations with its consequences; word of mouth intent and overall satisfaction. Of the two, word of mouth intent is found to have the highest correlation.

6.3 Conclusions

At the end of the dissertation, it is now time to refer back to the research questions stated in section 1.3. Firstly, by answering the following three sub-questions;

What is the relationship between perceived justice and satisfaction with complaint handling in the social media context?

Does the intensity of social media usage influence the significance of the relationship between perceived justice and complaint handling?

How does satisfaction with complaint handling influence overall firm satisfaction and word-of-mouth intent in a social media context?

The Future of Complaint Handling: The Effects of Perceived Justice on Post-Complaint Satisfaction