Service recovery:
How to get the maximum return on service
recovery efforts.
Master Thesis: “Service recovery: How to get the maximum return on service recovery efforts”
Author: Dave Westerhof Date: november 30th 2009
2
Management Summary
This study, examines whether a service recovery paradox exists on the temporary labour market and how to achieve such a paradox. A service recovery paradox is a situation in which post recovery satisfaction is greater than the satisfaction prior to the service failure when customers receive recovery from the service failure. In such a situation chances may occur in which a company can strengthen its relationships with customers. This study finds support in situations where process failures, a mishap in the way the service is delivered, lead to higher satisfaction after the
failure/recovery.
3
Index
Management Summary ... 2 Index ... 3 1. Introduction ... 4 2. Literature Review ... 8 2.1 Overview of literature ... 8 2.2 Service Failure... 9 2.3 Service recovery ...10 2.4 Customer satisfaction ...122.5 The Service Recovery Paradox ...14
3. Hypothesis development ...17
3.1 Effectively dealing with service recovery ...17
3.2 How the transaction-specific satisfaction influences the overall satisfaction ...19
3.3 Service recovery paradox ...19
3.4 The Conceptual Model ...20
4. Research design ...21
4.1 The research design ...21
4.2 Experimental design ...22 4.3 Composition of Questionnaire ...23 4.4 Parametric Testing ...26 5. Results...27 5.1 Overview of data ...27 5.2 Manipulation checks ...27 5.3 Hypothesis Testing ...28 6. Discussion ...33 6.1 Discussion ...33
6.2 Limitations and future research indications ...35
6.3 Conclusion ...36
7. References ...37
Appendix I, Constructs and measurement items ...41
4
1. Introduction
A healthy customer base is the true asset of a firm’s profitability. By acquiring new customers, retaining customers and cross- and up selling, the firm has a solid base to grow on profitability. Berger et al. (2006) show how the customer lifetime value of a customer base directly influences the shareholder value. As firms seem to know how to acquire new customers and how to make them buy more by cross- and up selling, increasing interest has grown on how to retain customers by setting up retention campaigns and specific programs to fight against defection of customers. The importance of this last component of a customer base, retention, is mentioned by Hart et al (1990) and further illustrated by Luo & Homburg (2008) with the stock value concept. Luo & Homburg (2008) state that focussing on satisfaction, and more specifically on dissatisfaction and complaints, helps lowering the stock value gap. Here they show a direct relationship between customer satisfaction/dissatisfaction and the stock value gap, whereas the stock value gap represents the difference of a firm’s actual market value from its optimal market value, as measured by a best-performance benchmark.
Indirectly, literature from the customer satisfaction field shows that customer loyalty due to satisfied customers, ultimately causes higher profitability (Anderson, E.W. and Fornell, 1994; Holowell, 1996; Yeung and Ennew, 2000). Thus, customer satisfaction is extremely important to a firm in maintaining a healthy customer base and to retain customers (Hart et al, 1990).
As illustrated above, it is extremely important to keep the satisfaction of the customers at its optimum and the complaints at the minimum (Luo & Homburg, 2008). However service failures occur, and it is usually the service recovery that makes a difference and turns the dissatisfied customers into satisfied ones (Hart et al, 1990). Homburg and Fürst (2007) and Buttle and Burton (2001) even state that, in case of service failures, loyalty depends essentially on complaint
satisfaction and not as much on satisfaction that has accumulated over time. “Additionally it seems that immediately following a complaint, customers’ perceptions are so dominated by the way their complaint was treated, that complaint handling becomes the main driver of loyalty”, (Homburg and Fürst, 2007). Alternatively, what seems to be a threat to the firm can be turned around and seen as an opportunity when dealt with effectively. Of course, some authors state that “doing it right the first time”(Andreassen, 2001; Michel and Meuter (2007); McCoullough, 2000) is better than recovering on a perpetuate basis. But, when failures in service occur, the company has to make sure there are sufficient resources available to address to the recovery of the failure and come up with a solution. It is this effective complaint handling that can have an enormous impact on customer retention rates, deflect the spread of damaging word of mouth, and improve bottom-line performance (Tax et al, 1998).
5
customers complain about the company’s failures. The solutions to these failures are a way to
increase customer satisfaction and ultimately customer retention (Hart et al 1990). This phenomenon is known as the service recovery paradox. Thus, other than seeing complaints as a huge liability to the firm, these failures can be seen as huge opportunities to strengthen relationships with customers and ultimately increase customer satisfaction in the bigger picture.
However, in the service recovery literature there much discussion about the generizability of the service recovery paradox. Concerning the service recovery paradox, or SRP, only a few different industries have been researched of which hotel and restaurant industries the most, while amongst others Ok et al. (2007), McCollough et al (2000) suggest that conclusions of these studies,
researching the SRP paradigm, must be exercised with extreme caution. Conclusions remain very context and industry specific and therefore its presence cannot be assumed in all markets.
Furthermore, it seems that mixed results have been found on the SRP, making it extremely difficult to rely on. On one hand, Magnini (2007) found a huge set of moderators that influence the
relationship between service recovery and complaint satisfaction. These moderators are among others, degree of severity of service failure, stability of the cause of the failure and level of control over the failure. On the other hand, Davidow (2003) found six dimensions of organizational responses to customer complaints, in which he examined the direct relationship between service recovery and complaint satisfaction. The author created a theory, although not empirically tested, of what works and what does not. Additionally, Smith et al. (1999) found out that different
organizational responses relate to different failures. They categorize two types of failures, outcome failures and process failures. The authors found that the different dimensions of organizational response relate differently to the types of failures. Thus, this study aims at clarifying and empirically testing these dimensions that influence the service recovery process and ultimately the SRP. Eventually firms know when to deal with service failures and in which manner they should resolve the problem effectively, to such a level that even a SRP should arise. Since much has been written about the moderators that influence the relationship between service recovery and complaint satisfaction, this thesis does not go in detail on this particular area. However, this study is the first one, which links the organizational responses and different service failure types to the SRP and examines the relationship between the variables.
The following problem statement is put central in this thesis.
6
From here we distinguish three research questions. The first research question examines whether the SRP is a “paradox or peril” (Smith and Bolton, 1998), “a justifiable theory or a smoldering myth” (Magnini et al, 2007), and whether the service recovery is “true and/or overrated” (Michel and Meuter, 2007). Many questions have risen about the true existence of the SRP, but improvements have been made the last years. Multiple authors mention the necessity of researching the SRP in different industries and contexts. By researching the SRP in a new setting, we contribute to the knowledge of it.
Secondly, ensuring that a SRP will arise it is necessary to execute a highly qualitative service recovery. The literature on the field of service recovery provides us with different mediating and moderating variables that influence the SRP (Magnini, 2007), but has never been differentiating between the two types of service failures, outcome and process. Smith et al (1999) argue that service recovery needs to be in sync with the occurred service failure in order to maximize complaint satisfaction. Evidently, in order to maximize customer satisfaction a firm needs to differentiate between the different types of failures. Research question 2 will give us insight in how to match all the different organizational responses with the two types of service failures.
Finally, Homburg and Fürst (2007) and Buttle and Burton (2001), as previously mentioned, say the last moments of interaction dominate the way in which customers evaluate their service provider. Overall satisfaction is build up over time and is subject to all kinds of influences, such as failures, recoveries and plain services. This overall satisfaction should not be confused with a transaction specific satisfaction, since this type of satisfaction represents the attitude towards one specific action, say one specific delivery of service. However, these transaction specific satisfactions together build up an overall satisfaction and disposition towards the company or brand as a whole, which is the overall satisfaction. Later in the literature review it shows that this relationship of the last moments of interaction on overall satisfaction is a necessary stage for a SRP to arise.
In summary, the research questions:
RQ1) Are customers in their post-failure state more satisfied compared to their pre-failure state (SRP)?
RQ2) Does the service recovery effort contribute to the complaint satisfaction of a customer and how do the organizational responses of companies relate to the complaint satisfaction when concerning the different types of service failures?
7
The research will be held under customers from a leading company in the temporary labour market which is the leading temporary labour agency in the Netherlands. This study uses a pre-test post-test experimental design whereas customers will be asked to evaluate their satisfaction levels. The satisfaction levels will be influenced by a specific service failure and organizational responses. Scenarios will be presented to customers. Questionnaires are used to gather results from actual customers.
8
2. Literature Review
In this section, the previous studies and literature concerning the SRP and service recovery will be outlined. All necessary variables will be discussed and defined.
2.1 Overview of literature
To show which variables this study deems important, a model has been depicted below. This model comes forth from the literature discussed in this review and represents the stages of satisfaction the customers follow, hence the path from dissatisfaction to satisfaction in case of a SRP. In between the dots (satisfaction levels), certain actions and/or variables influence these elements. Those are the variables that are important for this study. The reason for the arrow not declining after a service failure, is due to the fact that this is a necessary step to take for customer to be ultimately more satisfied in the end. So instead of seeing the dissatisfaction as a threat, this theory interprets the dissatisfaction as a chance to reach the ultimate goal of increased customer satisfaction. Hence, the service recovery paradox.
In the first phase, overall satisfaction lowers due to the service failure. Then a dissatisfied customer is turned into an overall satisfied customer again, but not before going through the stage of having a very high transaction-specific
satisfaction from the service recovery efforts of the firm. It is therefore that the service
failure and the service recovery are discussed,
because these actions influence the satisfaction levels of customer. This study also discusses the different sets of satisfaction, because these can be quite complicated. Finally, literature on the Service Recovery Paradox itself will be outlined
extensively. Overall customer satisfaction Dissatisfaction, due to service failure. Recovery satisfaction (transaction specific). SRP, increased overall customer satisfaction.
9
2.2 Service Failure
Service failures are defined as any service-related mishaps or problems (real and/or perceived) that occur during a consumer’s experience with a firm (Maxham, 2001). Even though it is impossible for a firm to constantly perform at an excellent level of service quality, firms can thrive to prevent from complaints or service failures to happen. More importantly because the complainers represent the minority of customers that are dissatisfied. Of customers that are dissatisfied with the service, 5-10 percent takes the effort to complain (Bamford and Xystouri, 2005, Tax et al., 1998). After a service failure, customers have several post purchase options : complaining (1), third-party action(2), exiting(3) and continue patronage(4) (Tax et al, 1998). However by reducing defections by just 5 percent , profit can boost up by 25 percent to 85 percent (Harari, 1992), hence the impact of the service recovery.
There is a distinction between two types of failures (Smith et al, 1999; Ok et al. (2007). On one hand, there is an outcome failure which is directly related to the core service offerings. Smith et al (1999) simply put is as the outcome that the customer actually receive from the service and can be recognised as the utilitarian benefit the customer receives. An example is a reserved hotel room which is unavailable because of overbooking. On the other hand there is the process failure, which relates to how customers receive the service. An example of this type of failure could be the rude treatment of a hotel clerk. The authors state that these two categories of failures differently affect the customer complaint satisfaction. However they find weak and mixed support across industries, but do find support that redressing has more effect on the re-establishment of satisfaction after an outcome failure than in case of a process failure. In return, other organizational responses affect the process failure. No other previous studies have examined the relationship between the different service failures and the service recovery (paradox). This thesis does an effort to examine this relationship.
10
2.3 Service recovery
The response by the firm is often referred to as the service recovery (Kelley and Davis, 1994). This response is subsequent to a service failure and the firms aims its efforts at re-establishing the satisfaction of the customer. Davidow (2003) has done extensive research on the field of service recovery. He suggests six dimensions of service recovery that are important in making the customer satisfied about the complaint handling itself. These elements are timeliness, facilitation, redress, apology, credibility and attentiveness. For building effective complaint handling, a company needs to know how to react to different complaints, outcome or process, and how to use the different
dimensions that are of the company’s disposal. We do however exclude the variable “facilitation” from this study, because facilitation does not represent a true response by an organization, but more the allocation of resources dedicated to stand-by for service recovery. It has a certain level of
proactivity and that is beyond the scope of this study. This leaves five dimensions of organizational response.
Table 2.1, Davidow's dimensions of organizational responses Dimensions Definition
Timeliness The perceived speed with which an organization responds to or handles a complaint.
Facilitation The policies, procedures, and structure that a company has in place to support customers engaging in complaints and communication.
Redress The benefit or response outcome that a customer receives from the organization in response to the complaint.
Apology An acknowledgement by the organization of the complainant’s distress.
Credibility The organization’s willingness to present an explanation or account for the problem.
Attentiveness The interpersonal communication and interaction between the organizational
representative and the customer. (Davidow, 2003)
Attentiveness
Some studies report strong effects from attentiveness on complaint satisfaction (Davidow, 2000). This relationship seems to be even stronger than redress, which is remarkable at first glance.
11
variable causes for the delight in the service recovery process. In making sure a SRP arises in this study, this might be an interesting variable.
Attentiveness can be broken down into respect, effort, empathy and willingness to listen to the customer. Attentiveness merely refers to the way the service is delivered.
Redress
Redress is mentioned in many studies as one of the largest drivers of complaint satisfaction
(Estelami, 2000; Davidow, 2003; McCollough, 2000), but needs to be used with caution and delicacy. Examples of redress are, amongst others, financial compensation, discount, reimbursement, free merchandise, coupons and so forth. Davidow (2003) suggests that partial redress is better than no redress at all, whereas more redress appears to be better than partial redress, but up to a certain limit. Redress is more appropriate for monetary complaints than for non-monetary complaints (Davidow, 2003). Smith et al. (1999) elaborate on this subject. They state that redress will have a greater positive effect on complaint satisfaction when outcome failures occur. They state, in the case of process failures, that redress could become superfluous, because other organizational responses may serve to restore customer satisfaction. An apology, credibility or both could be enough in this case, although Smith et al. (1999).
Apology
Davidow (2003) suggests that apology has a positive effect on all types of failures and has
intermediating effects on other organizational responses. Moreover, apology itself without other organizational responses has no effect on complaint satisfaction (Boshoff, 1997). It does have impact on complaint satisfaction in combination with other responses, such as redress (Boshoff, 1997; Boshoff and Leong, 1998; Goodwin and Ross, 1989). We can state that apology is not a main variable in causing complaint satisfaction, but it does improve the complaint satisfaction when combined with other variables. Hoffman and Chung (1999) agree in suggesting to combine apology with redress.
Timeliness
12
However, Smith et al. (1999) showed in their study that a speedy recovery has greater effect on complaint satisfaction when outcome failures occur. Since outcome failures represent an economic loss (Smith et al, 1999), one can see that results on this variable are mixed.
Credibility
Credibility is the second most important dimension that has impact on repurchase intentions, but also has strong relations with satisfaction (Davidow, 2003). Customers need to know that the failure will not happen again and that the company takes effort in preventing from that. In this case the customers are willing to repurchase and perceive greater complaint satisfaction. The explanation is twofold. Boshoff and Leong (1998) state that when the company takes the blame for the service failure higher satisfaction levels will be reached. Additionally, Conlon and Murray (1996) state that it is important to explain why the failure occurred. This will give insights into the process and may create some understanding. However, Davidow (2003) adds that credibility is very context specific and can therefore be perceived differently. Furthermore, credibility is even stronger in combination with redress. It shows that the company is serious with their explanation.
2.4 Customer satisfaction
Satisfaction has been widely researched, as it has been an important mediator in customer loyalty. Satisfaction is best explained through the expectancy-disconfirmation paradigm, whereas customer satisfaction represents an evaluation process where the customer develops an expectation for the service, based on e.a. benchmarks, previous encounters, advertisements and word-of-mouth, and measures it against the actual experience of the service (Oliver, 1993; Magnini et al, 2007). From there 3 types of outcome arise, negative disconfirmation, confirmation (service experience equals expectation) and positive disconfirmation. Positive disconfirmation occurs when the service level experience is better evaluated than the expectation. The opposite occurs in case of negative
disconfirmation. Satisfaction is derived from this paradigm. When service encounters meet or exceed expectations, satisfaction arises.
Furthermore, this part of the text tries to bring understanding to all the different types of
satisfaction. This is important, because this study measures satisfaction at three different moments and they all represent a different state of satisfaction. First of all, Andreassen (2001) makes a general distinction between two types of satisfaction that finds its usefulness throughout this whole study.
13
relation to the anticipated consequences. Cumulative satisfaction reflects the customer’s satisfaction with multiple interactions with the company over time” (Andreassen (2001)). This means that
customers have a disposition towards one single service encounter, transaction-specific, and they have an overall satisfaction towards the product, brand or company as a whole. This last overall satisfaction, or cumulative satisfaction according to Andreassen (2001), is merely based on all the single evaluations of previous service encounters, in which overall satisfaction is accumulated by all the transaction-specific evaluations. This distinction is important, because when measuring the service recovery paradox, one needs to make sure that the overall satisfaction levels are measured. In this study there is a pre-failure satisfaction and a post-failure satisfaction which are both overall satisfaction levels. These are the two entities that are measured when testing the service recovery paradox. Additionally there is the complaint satisfaction, which is a result of one service encounter namely the satisfaction that comes forth from the service recovery effort of the company. This transaction-specific satisfaction level is an evaluation of just one encounter, but contributes to the overall satisfaction level. Smith and Bolton (1998) even state that the last transaction dominates and therefore the overall satisfaction is most strongly influenced by the last transaction specific
satisfaction. So, previous encounters that have occurred in the far past, diminish over time. The figure below shows how specific service encounters are evaluated by transaction-specific satisfaction levels and ultimately form a disposition about one entity, an overall satisfaction level.
Figure 2.2. The influence of transaction-specific satisfaction on overall satisfaction. Derived from Ok et al. (2007)
14
information sources (Ok et al., 2007). When a service failure occurs, the customers’ overall satisfaction after the failure is lower than the initial overall satisfaction. This customers’ overall satisfaction is influenced by a transaction specific satisfaction, which is the failure encounter evaluation.
Building up an overall assessment (or satisfaction) is an ongoing and accumulative process and is, amongst other things, influenced by transaction specific evaluations. In case of exceptional service recovery, the overall assessment of satisfaction of customers can be higher than customers who have not experienced any problems (Smith and Bolton, 1998).
2.5 The Service Recovery Paradox
The service recovery paradox (SRP) is defined as the situation in which post recovery satisfaction is greater than the satisfaction prior to the service failure when customers receive high recovery performance (Maxham 2001, Smith and Bolton 1998). There are other definitions which are not consistent with the statistical meaning of the service recovery paradox. Ok et al. (2007) argue that in some studies (Andreassen, 2001; Michel and Meuter, 2007) the SRP was tested as to whether the satisfaction levels of customers are greater than those of customers who have not experienced any service failure (control group). Since the SRP is only applicable to those customers that have experienced a service failure, this study uses the definition from Maxham (2001).
15
Table 2.2. An overview of results on the service recovery paradox
Author(s) Industry Methodology Main Results SRP
Ok et al. (2007) Restaurant Scenario based experiment
The SRP applies when the recovery is valued as very high.
Yes
Maxham (2001) Internet
Provider / Haircut
Scenario based exp. Design / cross-sectional survey
The author finds the result reasonable given the high
involvement regarding the industry No
Maxham and Netemeyer (2002)
Banking Longitudinal study SRP holds up for one service failure, not two or more. SRP applicable to satisfaction, WOM and repurchase intent.
Yes
Magnini et al (2007) Hotel Role playing experiment
The SRP is proven and has many moderators to take into account.
Yes
Matos et al (2007) Various Meta Analysis SRP for satisfaction, but not for WOM and repurchase intentions. Choice of surveys or experiments does not alter result of SRP. Studies conducted in hotels have stronger results than other categories.
Yes
McCollough (2000) Airline industry
Scenario based experiment
The magnitude of the failure was great (severe failure) and no recovery paradox was found.
No
McCollough et al (2000) Hospitality Scenario based experiment
When a failure causes low harm and the recovery can completely mitigate the harm, then SRP is possible.
Yes
Smith and Bolton (1998) Hotel / Restaurant
Experimental-generated scenario
A customer’s cumulative
satisfaction increases when he/she is very satisfied with the
transaction-specific handling of the service failure
16
17
3. Hypothesis development
The previous literature preview helped in gaining deeper insights in the problem statement and research questions. From the literature, we now derive clear hypotheses that provide us with new knowledge and insights on the field of this study. The hypotheses will serve as input for the empirical research.
3.1 Effectively dealing with service recovery
This hypothesis contributes to the knowledge of service recovery literature by showing how certain organizational responses differently effect complaint satisfaction under different service failure situations, process or outcome. This way, service recovery can be more effective and evaluated better
by the customer, making a SRP more likely.
Davidow (2003) proposed dimensions of organizational response. Of these responses, attentiveness and redress are highlighted as the most important responses in service recovery and result in achieving customer complaint satisfaction (Estelami, 2000, Davidow 2003, Smith et al., 1999). Other elements of Davidow’s dimensions of organizational response have a certain influence on the complaint satisfaction, but to a lesser extent and usually effect the complaint satisfaction better, when combined with other responses.
Timeliness is considered to play a small role in the complaint satisfaction of customers, when
reviewing the literature. This is especially the case when outcome failures occur. Redress is supposed to play such a dominant role, that the effect of timeliness diminishes. However, Boshoff (1997) does mention the significance of long delays on complaint satisfaction. Additionally, Davidow (2003) argues, based on previous empirical research, that timeliness is very context and mode specific. In the restaurant, were people see waiting as a big hazard, time plays a more important role than in situations where time is less essential, for instance the temporary labour market. It is assumed that timeliness can be seen as a dissatisfier in the temporary labour market. Therefore no measures are taken on behalf of the variable timeliness, but this variable is included in the research and is presumed to be fixed under the condition high.
An apology itself has no important meaning (Boshoff, 1997). It does however influence the complaint
satisfaction when combined with other organizational responses, such as redress (Boshoff, 1997; Hoffman and Chung, 1999). Shortly, an apology is meaningless without the use of other
18
Credibility is seen as a key element in service recovery. There are strong linkages with redress, but
does not have the proper impact on complaint satisfaction to stand alone (Davidow, 2003). Literature provides us with a rather clear-cut relationship between credibility and complaint satisfaction. A well executed service recovery needs an (1) explanation, (2) a plan to prevent from making the same mistake again and (3) to take blame when the organization is accountable for it. These variables have been tested and all proven to be very powerful in service recovery (Davidow, 2003). Additionally, it is proven to be more useful with redress. No further measures are taken on behalf of the variable credibility, but this variable is included in the research and is presumed to be fixed under the condition high.
Estelami (2000) mentioned that redress was the far most dominant variable in service recovery. It does not matter how hard a company tries. If redress is lacking in a service recovery, customer satisfaction will not be as high as it could have been compared to a situation whereas redress was included. However, the relationship is not so straightforward as it seems. Some redress is better than no redress, but up to a certain limit, because overcompensation can be seen as a bribe. Blodget, Hill and Tax (1997) even state that a service recovery with low redress and high attentiveness was more satisfying than high redress and low attentiveness. Davidow (2003) even says: “attentiveness is the single most important dimension, having the largest effect of any dimension on satisfaction and repurchase.” So who do we believe?
Again, concerning interaction effects, Blodgett, Hill and Tax (1997) found an interaction effect between redress and attentiveness, whereas redress can only be significant if attentiveness is high. McCollough et al. (2000) also found an interaction effect, but they additionally found that both variables need to be high. On the contrary, Estelami (2000) mentioned that redress was the dominant variable in organizational response, while not mentioning the need for attentiveness altogether.
Furthermore, all the previous studies did not make a difference between the different types of failures. Except for Smith et al. (1999) which looked at the two different types of failures, mention that customers will place greater value on exchanges involving similar resources than on those involving dissimilar resources. Smith et al. (1999) say that in service failure encounters and service recovery encounters, customers will prefer to receive, in exchange for the loss suffered, resources that match the type of loss incurred. Shortly, organizational responses need to be in sync with the type of service failure. Could it therefore be that both Davidow (2003) and Estelami (2000) were right, by saying that attentiveness or redress is most important respectively? Maybe they just did not differ between outcome and process failures.
This previous theory possibly implies that attentiveness has a greater impact on complaint
19
hand, redress might be evaluated more strongly when recovering from an outcome failure as
opposed to a process failure. From this rationale we come up with the following hypotheses. Giving a recovery situation where apology, timeliness and credibility are held constant:
H1a: Attentiveness will have a greater impact (positively) on complaint satisfaction when a process
failure occurs than when an outcome failure occurs.
H1b: Redress will have a greater impact (positively) on complaint satisfaction when an outcome
failure occurs than when a process failure occurs.
3.2 How the transaction-specific satisfaction influences the overall
satisfaction
With this hypothesis we hope to prove that by performing an excellent service recovery, it contributes to the overall image and evaluating of the organization, from a customer’s perspective. Showing how
important it is to execute a well performed service recovery.
Since the satisfaction with complaint handling is a transaction specific state of mind, it does not mean that it simply carries over to the overall satisfaction, because one transaction is fairly diminishable in creating a good overall satisfaction level. Building up an overall assessment (or satisfaction) is an ongoing and accumulative process. However, Smith and Bolton (1998) state that the last transaction dominates and therefore the overall satisfaction is most strongly influenced by the last transaction specific satisfaction. Buttle and Burton (2001) confirm this in their study. So when service recovery is evaluated extremely well, complaint satisfaction will be higher as well. By showing the customer the company is up for the test of performing at an optimum and delivering an excellent service recovery, the complaint satisfaction will affect the overall satisfaction. Eventually the overall satisfaction is expected to be directly influenced by the complaint satisfaction. From this rationale we state the following hypothesis:
H2: Satisfaction with complaint handling is related positively with overall post recovery
satisfaction.
3.3 Service recovery paradox
20
Matos et al. (2007) already found that SRP has been found in different markets. However, as mentioned in the literature review, one has to be extremely careful in generalizing the results from SRP to different markets. Matos et al (2007) found evidence but not specifically for the temporary labour market. Following the statements from Ok et al. (2007), McCollough et al (2000), further research is necessary in making sure the SRP applies to certain industries.
The service recovery paradox is expected in this industry because of the matching core product, differentiating between manufacturing and service and the rather high switching costs of customers hiring temporary labour. One can imagine hiring 10 people and then coming across a service failure. When the customer decides to withdraw from the temporary labour agency, the consequences are extreme, because of losing the (partial) human capital asset of the company. Matos et al (2007) mentioned as a future research indication that the switching costs would play a role in coming across a SRP. Since the temporary labour market is characterized by the relative high switching costs and the other mentioned conditions, it is assumed that the SRP will arise in this particular industry. H3: Overall satisfaction (post recovery) is greater than overall satisfaction prior to the service
failure, in the temporary labour market setting.
3.4 The Conceptual Model
This conceptual model summarizes the hypotheses stated in the section. Unfamiliar is the
relationship between the service recovery dimensions and the post-failure overall satisfaction. Since this study examines the service recovery paradox, and therefore the variety between pre-failure and post-failure satisfaction, the conceptual model shows how the service recovery has effect on the overall satisfaction and ultimately on the SRP.
Figure 3.1 Conceptual model
21
4. Research design
This chapter describes how the actual research is performed and how hypotheses are tested and evaluated. Furthermore it is shown how the questionnaire, the main tool for gathering data, is established.
4.1 The research design
Already mentioned briefly, this study uses a pre-test-post-test between-subject scenario based experiment. This design has been used throughout many studies that researched the SRP
(McCollough 2000, Magnini et al. 2007, Smith and Bolton 1998, Maxham 2001, Ok et al., 2007). It has multiple advantages as opposed to other methods.
First of all, this method makes it possible to test the true definition of the service recovery paradox, which measures one group that is subject to different manipulations. It is possible to test the
different satisfaction levels, but without the use of an expensive longitudinal design. Secondly, due to extremely low incident rates of the service recovery paradox, it would be hard to find sufficient incidents when using a different model, the sample would have been enormous. Additionally, the research would not be disturbed by a response bias due to memory lapses. Finally, this method allows for greater variability in customer responses to service recovery.
On the other side, the risk of having respondents that are unable to identify themselves with the service failure situation is present. To encounter this problem, we have chosen to only hand out questionnaires to active customers.
Data collection and sampling plan
The experiment will be done through personally delivered questionnaires to actual customers. This way of spreading the documents will ensure a higher response rate than questionnaires that are sent through mail of email. Due to the use of actual customers, the chance of having respondents not being able to identify themselves with the subject of the experiment, will be reduced. This will increase higher validity.
22 Figure 4.1 Research flow
The primary interest of this study is to see how the satisfaction levels react to the service failures and the different organizational responses, questions 1 to 10 with an exception for question 4.
Additionally, the severity of complaint will be tested in question 4. The main purpose for this is to help the respondent in evaluating the scenarios better and has a minor contribution to the actual data used for insights.
4.2 Experimental design
Respondents were randomly assigned to a condition of a 2 (Service failure: outcome vs. process) x 2 (Redress: low vs. high) x 2 (Attentiveness: low vs. high) between-subjects design, see table 4.1. The independent variables are the organizational responses, which represent the dimensions of Davidow (2003), and the service failures, which represent an outcome failure and a process failure. The dependent variables are complaint satisfaction and the overall satisfaction towards a company.
Table 4.1. Experimental design
Organizational response
23
4.3 Composition of Questionnaire
The questionnaire is composed of validated scales and items used in previous research (Ok et al., 2007; Maxham and Netemeyer, 2002). Therefore no further validation is needed for the questions asked.
The scenarios that describe the service recovery are build upon clear descriptions of the organizational responses from Davidow (2003). Additionally Ok et al. (2007), Magnini (2007), Maxham (2001), McCollough et al. (2000) and Smith and Bolton (1998) serve as the main basis and inspiration for generating the recovery scenarios. These recovery scenarios can be found in the appendix II.
Furthermore, the scenarios of the service failures are a reflection of actual complaints following from the complaint management from a random company which is active in the temporary labour market. This department has been monitoring the complaints and its types for several years. Therefore, in order to stay as close to reality as possible, these scenarios represent mistakes from the top five of customer complaints. Due to these copies of reality, it is expected that customers will have no problem identifying the scenarios, which causes high validation. These failure scenarios can be found in appendix II.
Table 4.2 shows how the questions in the questionnaire are composed and which items relate to which variables. In this study we use only pre-tested scales, which are highly reliable and have been used on previous experiments.
To test whether the items used in the questionnaire are consistent and valid, a calculation was made on behalf of their consistency. All satisfaction variables show a Cronbach’s Alpha higher than 0,9. The other variables, manipulation checks on redress and attentiveness and the failure severity, cause a Cronbach’s Alpha 0,8 or higher. This provides evidence that all items used in the questionnaire, are valid and sufficiently consistent. Exact numbers can be found in table 4.2.
Table 4.2 Scales Questionnaire
Variable Author Items (Dutch translation)
Cronbach’s Alpha (α) *
Pre-failure satisfaction Ok et al. (2007) • I am satisfied with my overall experience with this company.
• As a whole, I am happy with this company. • Overall, I am pleased with the service
experiences with this company.
24
Post-failure satisfaction Ok et al. (2007) • I am satisfied with my overall experience with this company.
• As a whole, I am happy with this company. • Overall, I am pleased with the service
experiences with this company.
0,954
Complaint satisfaction Ok et al. (2007) • In my opinion, this company provided a satisfactory resolution to the problem on this particular occasion.
• I am satisfied with this company’s handling of this particular problem.
• I am satisfied with this particular service experience.
0,952
Failure severity Maxham and Netemeyer (2002)
• Minor problem/ Major problem;
• Small inconvenience/ Big inconvenience; • Minor aggravation/Major aggravation.
0,835
Man. check redress Homburg and Fürst (2005)
• I received an adequate compensation from the company.
• Overall, the compensation I received from the company was fair.
0,797
Man. check attentiveness
Homburg and Fürst (2005)
• The employee seemed to be very interested in my problem.
• The employee understood exactly my problem.
• Overall, the employee’s behavior during complaint handling was fair.
0,952
* The results concerning the Cronbach’s Alpha in this table are derived from the actual data used in this study.
Table 4.3 gives information of how the transition from variables to adequate responses are made. These attributes can be seen as building blocks for the questionnaire. A complete overview of scenarios is available in appendix II. 3 of 5 variables are fixed in this research and no measures are taken on their behalf. Therefore one can see that these variables, timeliness, apology and credibility, solely have the attribute level high. As for the other 2 variables, redress and attentiveness, the attribute levels are manipulated according to the experimental design.
Table 4.3Attribute levels of scales used in the recovery scenarios
25
Timeliness The customer will be called back within 15 minutes High
Apology The employee offers his apology to the customer High
Credibility The employee tells the customer that the temporary workers were not correctly informed about the right time.
The employee admits that this company is to blame for the service
failure.
The employee states that failures like these never occur
High
High
High Redress The employee will not charge the first day the workers are at work, this
as a compensatory measure for the loss incurred.
The employee informs you that he cannot compensate you furthermore
High
Low Attentiveness The customer is in direct contact with the employee
The employee attends to the service failure directly The employee responds in a correct and polite way The upcoming meeting this failure will be evaluated
(The absence of previously mentioned items with high attribute levels) Employee is not present
High High High High Low Low
Table 4.4 represents a short summary of the service failure scenarios. The complete scenario can be found in appendix II.
Outcome Failure A scenario in which a client temporarily hires flexible workers, but the flexible workers don’t show up.
Process Failure An order is placed for flexible workers, but the order never comes through
26
4.4 Parametric Testing
27
5. Results
Here the results of the actual research will be discussed and evaluated. Answers will be given to the hypotheses stated in Chapter 3 and the data will be transformed into knowledgeable information for the purpose of this research. In the next chapter conclusions will be made, based on the information
provided in this chapter.
5.1 Overview of data
238 respondents filled out the questionnaire. Their reactions were triggered through e-mails which were sent out in two sessions. The total number of emails sent out is 2272. This indicates a response rate of 10,48%.
Actual customers of, were asked to fill out the questionnaire. The customers were targeted where highest response rate was expected. Also the quality of the data was highest in this case, because of the ability of active customers to identify themselves with the situations (scenarios).
Respondents were randomly assigned to the scenarios, such that equally large groups would exist. The groups range from 26 till 36 respondents, where each group minimally represents 11% of the respondent base.
5.2 Manipulation checks
Firstly, an assurance has to be made whether the service failures were evaluated equally in severity. If the service failures would differ in severity, than distorted results would have been found on the complaint satisfaction eventually. The two failure scenarios, the outcome failure and a process failure, were evaluated by the respondents. Fortunately, a Mann-Whitney test resulted in a probability of 0,626 that the groups show equal results on their failure severity.
Furthermore, an assessment was made on the manipulations that were made in the scenarios. Since the recovery scenarios differentiated in their organizational responses, manipulation checks needed to be done for validation purposes. Respondents were asked to evaluate the manipulations
concerning the redress and the attentiveness. The means, which are displayed in table 5.1, show a variance, which can be explained by the deliberate manipulations of organizational responses appointed to the failure scenarios. The results of a one-way ANOVA test show that the means, which represent the average evaluation of respectively redress and attentiveness per scenario, differ significantly from each other, where p is smaller than 0,001. This indicates that the manipulations, that were implemented, have been remarked by the respondents. Unfortunately, B3 is not consistent with the results described above. In this case redress is considered higher in the case of B3, as
28
expect a service recovery which is in sync with the service failure and could evaluate that as compensatory or redress. Variable Scenario A1 Scenario A2 Scenario A3 Scenario A4 Scenario B1 Scenario B2 Scenario B3 Scenario B4 Redress 4,84 3,60 4,00 4,80 5,53 4,85 5,46 5,12 Attentiveness 5,67 4,54 5,41 5,50 5,83 5,47 6,04 6,09
Table 5.1Means of evaluation of respondents concerning the manipulations in the service recovery scenarios
Finally, some background variables were used in the questionnaire. The literature review in this study mentioned that, referring to the study of Maxham and Netemeyer (2002), the SRP is not likely to occur a second time. Meaning that customers who have already suffered one or more service failures, are less likely to show a SRP. For this purpose the question was asked whether they sustained a service failure in the past. This way, customers that encountered a previous service failure could be excluded from analysis.
A t-test shows whereas customers who encountered a service failure before, show in some satisfaction variables a less satisfied answer. Meaning that, according to Maxham and Netemeyer (2002), the chances for a SRP to occur are less likely in this group. Conclusively, this shows that this variable has to be taken into account, when testing the SRP.
5.3 Hypothesis Testing
Hypothesis 1Testing the hypothesis is done with a 2x2x2 ANOVA test. This analysis will show the influence of each organizational response individually and the interactions effects of these responses.
H1a: Attentiveness will have a greater impact (positively) on complaint satisfaction when a process
failure occurs than when an outcome failure occurs.
H1b: Redress will have a greater impact (positively) on complaint satisfaction when an outcome
failure occurs than when a process failure occurs.
29
satisfaction. Also the different organizational responses tend to react differently to the failure scenario.
Figure 5.1, Satisfaction levels outcome failure Figure 5.2, Satisfaction levels process failure
The hypothesis H1 assumed the superiority of attentiveness and redress in respectively process and
outcome failures, whereas interaction effects were assumed between the variables “failure scenario” and one of each organizational response, redress or attentiveness. The results are shown below in table 5.2.
Table 5.2, Results on ANOVA analyses Hypothesis 1 3,5 4 4,5 5 5,5 6 A1 A2 A3 A4 3,5 4 4,5 5 5,5 6 B1 B2 B3 B4
Tests of Between-Subjects Effects
Dependent Variable: Complaintsatis
90,709a 7 12,958 7,485 ,000 6081,410 1 6081,410 3512,924 ,000 53,519 1 53,519 30,915 ,000 21,710 1 21,710 12,541 ,000 12,220 1 12,220 7,059 ,008 5,752 1 5,752 3,323 ,070 ,590 1 ,590 ,341 ,560 2,854 1 2,854 1,648 ,200 ,605 1 ,605 ,349 ,555 398,165 230 1,731 6633,778 238 488,874 237 Source Corrected Model Intercept Failurescenario Redress Attentiveness Failurescenario * Redress Failurescenario * Attentiveness Redress * Attentiveness Failurescenario * Redress * Attentiveness Error Total Corrected Total
Type III Sum
of Squares df Mean Square F Sig.
30
There are three main effects, three two-way interactions and there is one three-way interaction. All interactions show non-significant results and therefore H1 is rejected. We do however find three
significant main effects. Weak evidence is found for an interaction whereas complaint satisfaction seems to respond almost significantly on the combination of failure scenario and redress.
There was a significant main effect of redress (/F/(1, 230) = 12.54,/p/ < .001), attentiveness (/F/(1, 230) = 7.06,/p/ < .05) and failure scenario (/F/(1, 230) = 30.92,/p/ < .001). Further results can be seen in table 5.2.
From the two organizational responses, redress seems to be the dominant variable in the service recovery process, which is in line with research from Estelami (2000), which argues the dominancy of redress. When looking at the interaction effect of redress with failure scenario, we find some
evidence that redress should be put in a context when used in the service recovery process. As can be seen in figure 5.1 and 5.2 redress seems to be powerful in the outcome scenario, but loses its strength when recovering from a process failure. Adversely, attentiveness always has a positive effect on complaint satisfaction. Attentiveness can therefore be seen as an important organizational response, in a sense where it has effect without accounting for different service failures.
Furthermore the interaction effect of redress with attentiveness is not found, which contradicts Blodgett, Hill and Tax (1997).
Hypothesis 2
In this test, the data gives us the possibility for testing the correlation between two variables. This study uses a Pearson correlation test after verifying for a normal distribution.
H2: Satisfaction with complaint handling is related positively with overall post recovery
satisfaction.
A Pearson correlation test shows a correlation of 0,661, where p < 0,01. H2 is therefore accepted.
This is evidence whereas the complaint satisfaction seems to have a relation on the overall post recovery satisfaction. So, in line with the findings of Homburg and Fürst (2007) and Buttle and Burton (2001), we found that complaint satisfaction has a clear correlation with post failure overall
31
Hypothesis 3
In line with the statistical meaning of the service recovery paradox in this study, this study compares the overall satisfaction levels within the scenarios and consequently the same groups. This study does not compare satisfaction levels between groups that differ in having a service failure and not having a service failure. A Paired T-test between Overall pre failure satisfaction and Overall post failure satisfaction for the different groups (scenarios) is used.
H3: Overall satisfaction (post recovery) is greater than overall satisfaction prior to the service
failure, in the temporary labour market setting.
The literature provided us with solid knowledge about the organizational responses and their subsequent satisfaction levels and how they would react. Therefore it would be logical only to test the satisfaction levels of the scenarios in which we would expect a significant result. For example, scenario A2 and B2 serve as a, more or less, control group and will therefore not be included in this test. In table 5.3 the results are shown for the various t-tests which have been performed.
Recovery Scenario Overall satisfaction Pre-failure (Mean) Overall satisfaction Post-failure (Mean) Service recovery paradox Statistically significant on 95% confidence interval
A1 (high redress and high attentiveness )
5,21 5,29 Yes, although
not significant
No, p=0,613
A4 (high redress and low attentiveness)
5,28 5,2 No No, p=0,416
B1 (high attentiveness and high redress)
5,11 5,63 Yes No, p=0,077
B3 (high attentiveness and low redress)
5,46 5,74 Yes No, p=0,065
(Yes, p=0,045 correcting for previous failure) Table 5.3.Results on the service recovery paradox
H3 is partially accepted. As can be seen in the table above, there is a case in which a service recovery
32
33
6. Discussion
This chapter provides answers to the research questions that are mentioned in chapter 1 of this thesis. Furthermore, the problem statement will be addressed and the managerial implications will be given. We will finalize with the limitations and future research indications.
6.1 Discussion
At the beginning of this thesis the question was, whether a service recovery paradox would exist in this particular industry. Additionally, by knowing how to affect such a service recovery paradox and how to increase the likelihood of causing a SRP, one could do a good job in retaining customers and increase profitability. The literature review has provided us with multiple organizational variables, originally developed by Davidow (2003), which could help companies building a good service recovery process, which in turn helps to better performance. This chapter contains a roadmap to build such a service recovery which helps companies towards an excellent service recovery and ultimately tries to retain its customers better.
First, this chapter will show how the different organizational responses play a different role in the service recovery process. Secondly the outcome of the service recovery process, complaint satisfaction, is placed in a context. The relation between complaint satisfaction and overall satisfaction will be made clear. Finally, the SRP will be discussed and its consequences for organizations.
Redress and Attentiveness
34
the case of an outcome failure. Conclusively, attentiveness should always be included, where redress should be used in the case of an outcome failure. This notion is in line with the findings of Smith et al. (1999).
Service recovery and satisfaction
When customers are highly satisfied overall, the literature review has shown that this has positive influence on word-of-mouth, retention rates, defections (Tax et al.; 1998) and ultimately improves profitability (Harari, 1992). Therefore it is important to know which factors significantly contribute to these high overall satisfactions. As mentioned before, Homburg and Fürst (2007) and Buttle and Burton (2001) say that complaint satisfaction, which comes forth from well performed service recovery, influences overall satisfaction mostly, when dealing with service failures. However, Smith and Bolton (1998) state that the last transaction dominates and therefore the overall satisfaction is most strongly influenced by the last transaction specific satisfaction. This study found evidence for the relationship between complaint satisfaction and overall post-failure satisfaction. However, no causal relationships could be tested, but we found a strong correlation between the two variables, which is exactly in line with the study of Smith et al. (1998). This shows that in optimizing
performance, a clear and effective service recovery is an essential part of a business today. It is this essential part of service recovery that can increase profitability up to 85% (Harari, 1992). Thus, in establishing a good service recovery process, one can see the usefulness of it.
The service recovery paradox
This study has shown that a service recovery paradox does exist, which means that customers can get more satisfied (overall) after an excellent performed service recovery. Does that mean that SRP’s offers chances that haven’t been seen so far? Well, this study goes as far as the overall satisfaction whereas it assumes the link between satisfaction on one side and retention rates, profitability, repurchase intentions on the other side. Matos et al. (2007) however shows in their study that the SRP only applies to the satisfaction, which is in line with this study, and not on other variables, repurchase intentions and word-of-mouth. This would indicate that customers get more satisfied and might show a more attitudinal loyalty, but would not transform that into behaviour loyalty.
35
The roadmap
To turn back to the issue at hand in the first chapter: “Isn’t doing it right the first time”(Andreassen, 2001; Michel and Meuter (2007); McCoullough, 2000), better than “recovering on a perpetuate basis?” The answer is “off course”, but if service failures do occur, which is normal in the case of a well-performing company, it needs to address to such a failure adequately. Therefore this study indicates that a company should always strive for causing a SRP and use recovery when it is needed. Based on the results of this study we would advise to do it right the first time whereas the main focus lies on the service/product itself.
Furthermore, a SRP does not hold twice. In line with Maxham and Netemeyer (2002), the results in this study indicated that there was a significant difference in responses, concerning the satisfaction levels, between customers that complained before and customers that have not complained before. This argues in favour of not forcing or provoking any failures.
Conclusively, the service recovery should not be given to little attention. Companies should have their resources available, so it can address the failures properly. With the tools given in this study, five organizational responses and its subsequent importance, we show how service recovery can best reach the maximum satisfaction per failure type. However, one has to be reminded about the
generizability of this subject. Many authors have emphasized the non- generizability of the SRP (McCollough et al.; 2000). We support that notion. This means that results apply for the temporary labour market, and those markets that could be analyzed similar to this market.
Finally, Smith et al. (1999) argued that service recovery needs to be in sync with the service failure itself. This study strongly enforces this theory.
6.2 Limitations and future research indications
36
longer time would make it extremely expensive. Also larger numbers will be advisable, since little evidence has been found concerning the interaction effects of different failure scenarios and its organizational responses. Literature is very clear about the different strengths of the organizational responses, but this study has failed in showing significant results. Future directions would be to use a greater sample size and/or a bigger difference between the recovery scenarios whereas the
difference is even more noticeable between high-low redress and high-low attentiveness. Furthermore, some of the data in this research was only appropriate for correlational testing, whereas causal testing would be preferable. This way stronger assumptions can be made.
6.3 Conclusion
37
7. References
Anderson, E.W. and Fornell, C. (1994) Customer satisfaction, market share, and profitability: findings from Sweden. Journal of Marketing 58, 53± 66.
Andreassen, T. Walling (1998), “Antecedents to satisfaction with service recovery,” European Journal
of Marketing, Vol. 34, 156-175.
Andreassen, Tor Wallin (2001), “From Disgust to Delight: Do Customers Hold a Grudge?” Journal of
Service Research, 4 (1), 39–49.
Bamford, D. and Xystouri, T. (2005), “A case study of service failure and recovery within an international airline”, Managing Service Quality, Vol. 15 No. 3, pp. 306-20, ISSN 0960-4529.
Berger, Eechambadi, George Lehmann Rizley Venkatesan (2006) From CUSTOMER LIFETIME VALUE TO SHAREHOLDER VALUE, journal of service research, vol 9 no 2, 156-167.
Blodgett, Jeffrey G., Donna J. Hill, and Stephen S. Tax (1997), “The Effects of Distributive, Procedural, and Interactional Justice on Postcomplaint Behavior,” Journal of Retailing, 73 (2), 185-210.
Boshoff, Christo (1997), “An Experimental Study of Service Recovery Options,” International Journal
of Service Industry Management, 8 (2), 110-30.
Boshoff, Christo and Jason Leong (1998), “Empowerment, Attribution and Apologising as Dimensions of Service Recovery: An Experimental Study,” International Journal of Service Industry Management, 9 (1), 24-47.
Buttle, Francis and Burton, Jamie (2001), “Does service failure influence customer loyalty?,” Journal
of Consumer Behaviour, vol. 1, 3, 217-227.
Conlon, Donald E. and Noel M. Murray (1996), “Customer Perceptions of Corporate Responses to Product Complaints: The Role of Explanations,” Academy of Management Journal, 39 (4), 1040-56.
38
Davidow, Moshe (2003), “Organizational responses to customer complaints: What works and what doesn’t,” Journal of Service Research, vol. 5 no 3, 225-250.
DeWitt, T., Doan T Nguyen and Roger Marshall (2008), “Exploring customer loyalty following service recovery,” Journal of service research, 10, 269-281.
Estelami, H. (2000), “Competitive and procedural determinants of delight and disappointment in consumer complaint outcomes”, Journal of Service Research, vol 2 no. 3, 285-300.
Gilly, M.C. and Gelb, B.D. (1982), ``Post-purchase consumer processes and the complaining customer'', Journal of Consumer Research, Vol. 9, December, pp. 323-8.
Goodwin, C. and Ross, I. (1990), ``Consumer evaluations of responses to complaints: what's fair and why'', Journal of Services Marketing, Vol. 4, Summer, pp. 53-61.
Harari, Oren (1992), “Thank Heaven for Complainers,” Management Review, 81 (January), 59–60
Hart, Christopher W., James L. Heskett, and W. Earl Sasser, Jr. (1990), “The Profitable Art of Service Recovery,” Harvard Business Review, 68 (August), 148-56.
Heskett, J.L., T.O. Jones, G.W. Loveman, W. Earl Sasser & L.A. Schlesinger (1994), Putting the Service- Profit Chain to Work, Harvard Business Review, 72, (March-April), 164-174
Hoffman, K. Douglas, and Beth G. Chung (1999), “Hospitality Recovery Strategies: Customer Preference versus Firm Use,” Journal of Hospitality and Tourism Management, 23 (1), 71-84.
Holowell, R. (1996) The relationships of customer satisfaction, customer loyalty, and profitability: an empirical study. International Journal of Service Industry Management 7(4), 22± 42
Homburg, Christian and Fürst, Andreas (2005), “How organizational complaint handling drives
customer loyalty: An analysis of the mechanistic and the organic approach,” Journal of Marketing, 69, 95-114.
Kelley SW, Davis MA. (1994), “Antecedents to customer expectations for service recovery”, Journal
39
Luo, Xueming and Homburg, Christian (2008), “Satisfaction, complaint, and the stock value gap,”
Journal of Marketing, vol. 72, 29-43.
Luo, Xueming and Homburg, Christian (2007), “Neglected outcomes of customer satisfaction,”
Journal of Marketing, vol. 71, 133-149.
Magnini, V.P., Ford, J.B. and Markowski, E.P. et al. (2007), “The service recovery paradox: justifiable theory or smoldering myth?”, Journal of Service Marketing, Vol. 21 No. 3, pp. 213-25.
Matos de, Celso Augusto, Jorge Luiz Henrique and Carlos Alberto Vargas Rossi (2007), “Service Recovery Paradox: A meta analysis,” Journal of Service Research, Vol. 10 no. 1, 60-77.
McCollough, M.A. and Bharadwaj, S.G. (1992), “The recovery paradox: an examination of customer satisfaction in relation to disconfirmation, service quality, and attribution based theories”, in Allen, C.T. (Ed.), Marketing Theory and Applications, American Marketing Association, Chicago, IL, p. 119.
McCollough, Michael A. (2000), The Effect of Perceived Justice and Attributions Regarding Service Failure and Recovery on Post-Recovery Customer Satisfaction and Service Quality Attitudes,” Journal
of Hospitality and Tourism Research, 24 (4), 423-47.
McCollough, Michael A., Berry, Leonard L. and Yadav, Manjit S. (2000), “An Empirical Investigation of Customer Satisfaction After Service Failure and Recovery”, Journal of Service Research, Volume 3, No. 2, 121-137.
Maxham, James G. (2001), “Service recovery’s influence on consumer satisfaction, positive word-of-mouth, and purchase intentions, “ Journal of business research, 54, 11-24.
Maxham, James G. and Netemeyer, Richard G. (2002), “A Longitudinal Study of Complaining
Customers’ Evaluations of Multiple Service Failures and Recovery Efforts”, Journal of Marketing, Vol. 66, 57–71.
Michel, Stefan and Meuter, Matthew L. (2007), “The service recovery paradox: True but overrated?,”
40
Ok, C., Back K. and Shanklin C.W. (2007) “Mixed Findings on the Service Recovery Paradox”, The
Service Industries Journal, Vol.27, No.6, pp.671–686.
Oliver, R. (1993), “Cognitive, affective, and attribute bases of the satisfaction response”, Journal of
Consumer Research, Vol. 20 No. 1, 418-30.
Smith, Amy K., Ruth N. Bolton and Janet Wagner (1999), “A model of customer satisfaction with service encounters involving failure and recovery,” Journal of marketing research, vol. 36, 356-372.
Smith, Amy K. and Ruth N. Bolton (1998), “An experimental investigation of customer reactions to service failure and recovery encounters: Paradox or Peril”? Journal of Service Research, vol. 1 no. 1, 65-81.
Tax, Stephen S., Stephen W. Brown & Murali Chandrashekaran (1998), “Customer evaluations of service complaint experiences: Implications for relationship marketing,” Journal of Marketing, vol. 62, 60-76
Voorhees, Brady and Horowitz, 2006, A Voice From the Silent Masses: An Exploratory and Comparative Analysis of Noncomplainers, JOURNAL OF THE ACADEMY OF MARKETING SCIENCE
Yeung, Matthew C.H. and Ennew, Christinet. (2000), “From customer satisfaction to profitability”,
41
Appendix I, Constructs and measurement items
Complaint satisfaction (Ok et al., 2007)• In my opinion, this company provided a satisfactory resolution to the problem on this particular occasion.
• I am satisfied with this company’s handling of this particular problem. • I am satisfied with this particular service experience.
Overall satisfaction (pre-failure and post-recovery) (Ok et al., 2007) • I am satisfied with my overall experience with this company • As a whole, I am happy with this company
• Overall, I am pleased with the service experiences with this company. Failure Severity (Maxham III & Netemeyer, 2002)
In my opinion, the banking problem that I experienced was a • Minor problem (1)/major problem (7).
• Big inconvenience (1)/small inconvenience (7). • Major aggravation (1)/minor aggravation (7).
Distributive justice, basis for manipulation check on compensation (Homburg and Fürst, 2005) • I received an adequate compensation from the company.
• In solving my problem, the company gave me exactly what I needed. • Overall, the compensation I received from the company was fair.
Interactional justice, basis for manipulation check on attentiveness (Homburg and Fürst, 2005) • The employees seemed to be very interested in my problem.
• The employees understood exactly my problem. • I felt treated rudely by the employees. (R)
• The employees were very keen to solve my problem.
42 Appendix II, scenarios.
