Effects of positive and negative switching costs
on emotions and subsequent behaviour.
A mixed methods approach on understanding the customer’s perception of switching costs.
by Johann Lagemann
MSc Marketing Intelligence Master Thesis Defense July 3rd, 2018
University of Groningen Faculty of Business and Economics
Background
•
Economic success depends on ability to maintain long-term relationship (Rust et al. 2004)
•
Competition and costs of attracting new customers increases (Biglaiser et al. 2013)
à
important to understand customer retention and reasons behind repeat purchase
•
Key driver to achieve customer retention à switching costs
Most important consequence
= behavioural loyalty
(Klemperer 1995)
Other consequences ?!
1. Perception of switching costs
(Burnham et al. 2003)Research questions
1. “Which type of switching costs is viewed as either positive or
negative by customers and why?”
2. ”How do switching costs and their perception lead to
emotions?”
3. “How do switching costs and their perception give rise to
different behavioural outcomes?”
Theoretical relevance
à
Switching costs can be perceived
both positively and negatively
à
They give rise to emotions and
subsequent behaviour
Managerial relevance
Conceptual model
Switching
costs
Emotions
Behaviour
Research design
Data
1. Dataset: Secondary qualitative questionnaire data
“Describe the last time you wanted to switch a provider and
why you did (not).”
à
91 narratives transformed (i.e. coded) into quantitative data
Methodology
2. Dataset: Secondary quantitative ACSI data
Indices on customer satisfaction, antecedents and outcomes
à
Calculation of switching cost metric
Coding
Statistical
testing (R)
Human coding
LIWC analyses
T-test
ANOVAs
Results
Study 1
Financial, relational, and procedural SC
à
Groups are not significantly different from each other
à
F(88%), R (35%) and P (32%)
à
Significant more negative perception of procedural SC
Discussion &
Recommendations
à
Rather positive - than negative emotions are affected
à
No sufficient evidence for link of SC through emotions to behaviour
à
Costs decrease positive emotions, benefits have no effect
à
Effects differ per classification i.e. perception
Theoretical implication
à
Switching costs can be perceived
both positively and negatively
à
They in-/decrease positive
emotions, but not negative
emotions
Managerial implication
à
In case of low customer
satisfaction, lower switching costs
à
Use SC as defensive strategy
Limitations & Future research
Limitations
1. Different datasets
2. Sample size
3. Interpretation of LIWC output
4. Respondent are students
Future
research
1. Switching cost source attribution
2. Cultural or individual factors
3. Firm reactions
Effects of positive and negative switching costs on
emotions and subsequent behaviour.
A mixed methods approach on understanding the customer’s
perception of switching costs
Johann Lagemann
MASTER THESIS
MSc Marketing
Intelligence
Effects of positive and negative switching costs on
emotions and subsequent behaviour.
A mixed methods approach on understanding the customer’s
perception of switching costs
Faculty of Economics & Business Department of Marketing Master thesis 18th June 2018 Johann Lagemann Tuinbouwstraat 94b 9717JM Groningen +4915784589763 j.l.lagemann@student.rug.nl S3049604 Supervisor A. Bhattacharya
Assistant Professor of Marketing Second supervisor
Dr. J. van Doorn
ABSTRACT
Switching costs are an important mechanism by which firms increase customer retention and achieve sustained competitive advantage. While the benefits of switching costs to the firm have been shown in numerous analytical studies, research on how switching costs are perceived by the consumer is lacking. We propose that switching costs of each of the three major types identified in the literature (financial, relational, and procedural) can be perceived by customers as either positive or negative. We also suggest that positively and negatively perceived switching costs give rise to positive/negative emotions and subsequent behavioural outcomes, such as complaint behaviour and word-of-mouth. In addition, we argue that switching costs are linked to both benefits/costs of leaving/staying, which have distinct emotional consequences. Divided into three different studies, this paper analyses voice of customer (VOC) data and data from the American Customer Satisfaction Index (ACSI). Our results show that procedural switching costs have both positive and negative characteristics (benefits and costs). We also find that positive switching costs increase a customer’s positive emotions, whereas negative switching costs decrease positive emotions. However, we do not find sufficient evidence for subsequent complaint behaviour or word-of-mouth behaviour. The results furthermore show that costs of leaving/staying decrease positive emotions. These results provide insights to researchers and managers regarding the customer’s perception of switching costs, how to enhance customer loyalty and prevent customer churn.
PREFACE
The preface – I have been looking forward to writing this part of my thesis, since it encompasses the last integral part. Not that I did not enjoy writing my thesis, but still it displays the end of my Master of Science in Marketing Intelligence and I am sure everybody would agree with me when I say that any project needs an end at some point. So, this is it. The day I decided to study in Groningen – it feels like it was yesterday. Time flies and the next chapter is awaiting. I really hope that you will enjoy reading my thesis. I am sure, everybody will be able to identify himself/herself with the topic as it affects most of us. It has been of incredible interest to deep dive in this topic, make my own research and contribute to the world of marketing. There is one thing I have learned in the past – “Research and education is key”. It is the basis for what we become, how we live together and tackle todays challenges in an ever changing world.
Still, I want to thank a couple of people, that are incredibly important to me and helped me in successfully accomplishing this project. First of all, thanks to my parents, without whom all of this would not be possible. Having the possibility to go to a foreign country and study what I strive for – it is not for granted. Second, thanks to all my friends, who made this a very special time in Groningen. Thirdly, thanks to Abhi, a very flexible, valuable and relaxed supervisor, what is astonishing considering the fact, that I was one of 12 students he had to supervise. Also, thanks to Fleur and Harish, my two thesis buddies, for helping out and motivating me.
We made it!
TABLE OF CONTENTS
1. Introduction 7
2. Literature review 10
2.1 Emotions 10
2.2 Switching costs 11
2.2.1 Financial, relational and procedural switching costs 11
2.2.2 Positive and negative dimension of switching costs 13
2.2.3 Benefits and costs of leaving and - staying 15
2.2.4 Delimitation of positive/negative switching costs from benefits/costs 16
2.2.5 Prisoner, churner, positive stayer, and rational stayer 17
2.3 Behaviour 18 2.3.1 Complaint behaviour 18 2.3.2 Word-of-mouth behaviour 19 2.4 Control variables 21 2.4.1 Time 21 2.4.2 Industry 21
2.5 Overall framework and hypotheses 22
2.5.1 Study 1 23
2.5.2 Study 2 23
2.5.3 Study 3 24
3. Study 1: Switching costs as antecedents of emotions 25
3.1 Research design 25
3.1.1 Data 25
3.1.2 Narrative coding 25
3.1.2.1 Human coding 25 3.1.2.2 Linguistic Inquiry and Word Count (LIWC) analyses 26
3.1.3 Variable operationalization 26
3.1.4 Descriptive statistics 28
3.1.5 Method 29
3.1.5.1 Analysis of variance (ANOVA) 29 3.1.5.2 Linear regression model(s) 29 3.1.5.3 T-test 29
3.2 Results 29
3.2.1 Segmentation on financial, relational and procedural switching costs 29 3.2.2 Perception of financial, relational and procedural switching costs 30 3.2.3 Emotional outcome of positive and negative switching costs 31
3.3 Discussion 32
4. Study 2: Switching costs as antecedents of behaviour 34
4.1 Research design 34
4.1.1 Data 34
4.1.2 Variable operationalization 34
4.1.2.1 Calculation of switching cost metric 35
4.1.3 Descriptive statistics 36
4.1.4 Method 37
4.1.4.1 Analysis of variance (ANOVA) 37 4.1.4.2 Linear regression model(s) 38
4.2 Results 38
4.2.2 WOM behaviour 40
4.3 Discussion 41
5. Study 3: Benefits/costs as antecedents of emotions 44
5.1 Research design 44
5.1.1 Data and method 44
5.1.2 Variable operationalization 44
5.1.3 Descriptive statistics 44
5.2 Results 44
5.2.1 Segmentation on benefits and costs of leaving/staying 44
5.2.2 Emotional outcome of benefits and costs of leaving/staying 46 5.2.3 Emotional outcome of benefits and costs of leaving/staying per group 46
5.3 Discussion 47
6. General discussion 49
6.1 Findings and theoretical implications 49
6.2 Managerial implications 50
6.3 Limitations and future research 51
References 53
1. INTRODUCTION
For most companies, economic success depends on the ability to maintain long-term relationships with customers, who purchase their offerings repeatedly (Rust, Ambler, Carpenter, Kumar, and Srivastava 2004). Customer retention, defined as the continued patronage of a firm’s products and services from existing customers over time, is a critical marketing goal for all firms and has long been viewed as a behavioural manifestation of customer loyalty (Jacoby and Chestnut 1978), also because customer loyalty is a key concern of marketing managers due to its potential impact on firm performance (Reichheld and Teal 2001). As competition and the costs of attracting new customers increase (Biglaiser, Cremer, and Dobos 2013), understanding customer retention and the reasons behind repeat purchase is therefore an issue of considerable theoretical and managerial importance. Extensive efforts have been made by marketing scholars to identify antecedents of customer retention. One key driver identified and primary way in which firms achieve customer retention is through the erection of switching costs, such that the costs to customers of changing to an alternative product or service provider acts as an obstacle to their doing so (Jones, Mothersbaugh, and Beatty 2000; Keaveney 1995; Froot and Klemperer 1989). Dwyer, Schurr, and Oh (1987) posit that all else being equal, a consumer will stay in a relationship with a particular provider in order to economize on switching costs. Bansal and Taylor (1999) provide empirical verification of this, showing that switching costs predict switching intentions and switching behaviour, and are a stronger predictor of customer retention than customer satisfaction. Likewise, Lam, Shankar Erramilli, and Murthy (2004) demonstrate that switching costs directly influence both advocacy and patronage loyalty.
for the firm, such as decreasing growth rates. For example, in the wireless communication industry, grudge-holding incumbent customers have been shown to dissuade potential new customers by spreading “bad” word-of-mouth (WOM) about the firm (Malhotra and Kubowicz-Malhotra 2013). Thus, a more thorough understanding of both the favourable and unfavourable consequences of customer switching costs would aid in judging its utility as a managerial tool. Switching costs are also of great interest and concern to makers. From a policy-maker perspective, this interest is driven by the notion that switching costs harm consumer welfare and may also form a source of provider firm’s market power (Klemperer 1995; Waterson 2002). However, the grounds for such policy-maker interest may not necessarily be well-founded. Specifically, switching costs may be a positive choice made by consumers because of strong emotional attachment to a brand or relationships with provider personnel rather than a result of firm or industry-wide use of "lock-in" mechanisms – i.e., it is likely that there may be both negative and positive dimensions of switching costs. For managers, switching costs are furthermore important because they allow the identification of customers that are more or less likely to defect to a rival provider in the future (Roos and Gustafsson 2007; 2011). Such knowledge allows managers to allocate more resources to protect attractive but high switching potential customers and fewer resources to customers that have a low probability of switching. Switching costs are usually implicitly or explicitly framed as a negative phenomenon for customers (Fudenberg and Tirole 2000) since it restricts their freedom; and a positive phenomenon for the firm (Klemperer 1995) possessing them since it helps firms to retain its customers and charge higher prices to the customers who are “locked-in”. However, a number of researchers have argued that switching costs consist of both positive dimensions, involving switching barriers that encourage customers to willingly remain in a relationship, and negative dimensions, involving switching barriers that make customers feel trapped in a relationship (Jones, Reynolds, Mothersbaugh, and Beatty 2007; Colgate and Lang 2001). Since switching costs are a result of a consumer’s subjective evaluations of experiences in a relationship with the provider, the question of whether customers evaluate and perceive both types of switching costs similarly or differently is important. Of further significance to managers and academics is whether a differential perception of switching cost might have any effects on a customer’s behaviour, such as complaint behaviour and WOM.
Third, how do switching costs and their perception give rise to different behavioural outcomes? In addressing these three questions, this research makes two contributions to the literature. First, we argue that switching costs in all three major domains within which they have been conceptualized (financial, relational, and procedural) can be viewed as either positive or negative by customers. This contrasts with previous conceptualizations by economists, who assume that all switching costs are bad for customers. It also contrasts with previous work in marketing, in which researchers have shown that switching costs can be positive, but only in the domain of relational switching costs (e.g. Jones et al. 2007). Second, building on Lazarus’ appraisal theory of emotions, stress and coping (Lazarus 1991), we argue that perceptions of negative switching costs give rise to negative emotions and customers may cope by increased complaint behaviour and spreading negative word-of-mouth about the firm and are increasingly motivated to exit a relationship with the firm. Conversely, perceptions of positive switching costs evoke positive emotions, why customers are less likely to engage in complaint behaviour, but more likely to spread positive WOM about the firm and are increasingly motivated to stay with the provider.
The results show that switching costs on the financial, relational and procedural dimension can be perceived both positively and negatively (benefits and costs) and contribute to prior literature, where only relational switching costs have been looked at positively (Colgate and Lang 2001). Besides that, we find that the positive dimension of switching costs, more precisely “willingly” staying in a relationship, directly increases positive emotions, whereas negative switching costs, i.e. being “locked-in” decrease positive emotions. However, with respect to behavioural consequences we do not find sufficient evidence that the perception of switching costs leads to complaint behaviour or WOM behaviour. While the effect of switching costs has been studied previously (Haj-Salem and Chebat 2014) past literature has not yet looked at the direct relationship between switching costs and emotions.
2. LITERATURE REVIEW
This chapter contains an extensive discussion of existing literature on the topic of the three studies. First, we review each variable and present relevant research findings in order to provide the reader with background information before we then introduce the overall framework and the conceptual model of each study. This section therefore serves as the basis for the hypotheses, which are developed and tested later on.
2.1 Emotions
Emotions are ubiquitous throughout marketing and have received great attention in the past, especially because emotions play a decisive role in explaining customer purchase behaviour (Kim and Lennon 2013) and human decision-making (Haj-Salem and Chebat 2014). According to Bagozzi, Gopinath, and Nyer (1999, p. 202) emotions “influence information processing, they mediate responses to persuasive appeals and enact goal-directed behaviours.” Though in order to measure emotions, we need to define the latter and distinguish positive from negative emotions.
Bourne and Russo (1998, p. 364) define emotions as a “subjective internal state that has biological, cognitive and social components.” Martin, O’Neill, Hubbard, and Palmer (2008, p. 226) simplify this definition and state that “sudden exposure to stimuli gives rise to a state of positive or negative autonomic arousal, provoking a cognitive analysis of the stimuli, which then provokes a physiological reaction leading to a feeling best described as emotion.” Sheth, Mittal, and Newman (1999, p. 356) follow this idea of external influence and define emotions as “consciousness of the occurrence of some physiological arousal followed by a behavioural response along with appraised meaning of both.”
Positive and negative emotions can be distinguished based on the idea of motivation. Whereas positive emotions can be described best as approach motivation, i.e. attaining positive emotional experiences and/or states, negative emotions can be associated with avoidance motivation, meaning avoiding negative emotional experiences (Martin et al. 2008). These emotions can act as a source of information, which are used to evaluate the stimulus and form an attitude (Martin et al. 2008). Furthermore, positive and negative emotions during pre-, actual and post-consumption of the service encounter are found to be influential on long-term behavioural intentions, such as positive and negative WOM as well as complaint behaviour (Martin et al. 2008; Zeelenberg and Pieters 2004).
emotions and build the systematic causality of subsequent behaviour. This is especially interesting for firms to know as consumer behaviour can indeed be observed, but seldom explained. The following section therefore gives further information on the theory of switching costs and their potential relation to emotions and subsequent behaviour.
2.2 Switching costs
As competition and costs of attracting new customers increase, retaining customers is one of the major concerns of a firm’s marketing efforts (Biglaiser, Cremer, and Dobos 2013). Switching costs receive increasing attention and have been used as a viable tool for keeping customers in relationships (Bansal, Irving, and Taylor 2004).
They are defined as the one-time costs, contrary to ongoing costs of using a product or service, that customers associate with the process of switching from one provider to another, which do not need to be incurred immediately upon switching (Burnham et al. 2003). Farrell and Klemperer (2007) state that consumers experience switching costs between providers when an investment specific to her/his current provider, e.g. investment in equipment or learning how to use a product, must be duplicated for a new provider. This demonstrates that a product will have switching costs if a buyer will purchase it repeatedly and find it costly to move between providers. This approach is centred on the idea of “sunk costs” (already incurred costs that cannot be recovered) made by customers in their relationship with an existing provider that would need to be given up and potentially recreated at a new provider. Economists have therefore defined switching costs in terms of both past and future investments made by customers in a relationship with an existing provider, and the future costs associated with switching to a new provider (Shapiro and Varian 1999).
Jones, Mothersbaugh, and Beatty (2002) define switching costs as perceived economic and psychological costs, associated with changing from one provider to another, thus indicates two dimensions on which switching costs can be experienced. Whereas Klemperer (1995) shows that switching costs originally were thought to be only financial or psychological, more recent research has shown that switching costs are multidimensional, as the following paragraph shall further describe in detail.
2.2.1 Financial, relational and procedural switching costs
Procedural switching costs primarily involve the spending of time and effort by the customer and include evaluation of product/service alternatives, learning related to acquiring and consuming such alternatives, and related setup costs (e.g. the effort required by the customer to reach a similar level of comfort in usage may not be transferable to other brands of the same products). Financial switching costs consist of monetary costs in moving from one product/service provider to another and involve the loss of financial resources (e.g. firm-levied contracts in order to penalize switching by customers). Relational switching costs consist of personal and brand relationship loss associated with switching providers and involve psychological discomfort due to breakage of perceived relational bonds (e.g. the sense of belonging, empathy, courtesy, understanding, familiarity or even friendship which a customer enjoys with the incumbent provider may not be transferable to a new provider) (Burnham et al. 2003).
Past literature has classified relational, procedural and financial switching costs as positive or negative depending on which aspects of the relationship the type of switching cost relates to. Jones et al. (2007) for instance show that procedural switching costs are perceived negatively by customers as they are a negative source of constraint while relational and financial costs are perceived positively since they are a positive source of constraint. Yet did they only consider certain aspects of each switching cost domain in order to build the scales and arrive at conclusions. To give an example, Jones et al. (2007) only look at loyalty rewards when evaluating financial switching costs and do not take other types of financial switching costs into account, such as contracts. Further, while classifying switching costs as positive or negative, research has mostly looked at the disutility from switching, i.e. the costs of leaving, but not at the utility from not switching, i.e. the benefits of staying. Colgate and Lang (2001) for instance only consider the hassle which customers face when trying to exit a service, but do not take into account the convenience of staying.
2.2.2 Positive and negative dimension of switching costs
Usually switching costs are differently framed for customers and firms. Whereas they are implicitly or explicitly framed as a negative phenomenon for customers (Fudenberg and Tirole 2000), as they restrict the customer’s freedom, they are framed as a positive phenomenon for the possessing firm (Klemperer 1995) since it allows firms to retain its customers and charge higher prices for customers, who are “locked-in”. Switching costs have therefore been regarded as a viable defensive strategy used by firms to prevent its customers from exiting the relationship since they act as a constraint, discouraging customers from moving to alternative product/service providers (Dube, Hitsch, and Rossi 2009). However, theoretical work on customer attachment (Thomson, MacInnis, and Park 2005) and willingness to stay in relationships (Hirschman 1970), suggests that from a customer perspective there is likely a difference between being unable to exit a provider relationship and willingly staying. Being unable to exit a relationship refers to the negative dimension of switching costs, which involve switching barriers, that make customers feel trapped in a relationship. Willingly staying pertains to the positive dimension of switching costs, which involves switching barriers that encourage customers to remain.
transaction. For example, Meuter, Olstrom, Bitner, and Roundtree (2003), while explaining the growing use of self-service technologies, show that the speed and ease with which customers can access a service presents a key benefit to customers. Learning costs may also be framed as gained knowledge such that the customer does not need to put additional efforts to learn something again. The customer benefits because s/he gains expert knowledge and information about the service/product and provider and is comfortable using it (Ratchford 2001).
On the other hand, negative switching costs provide the customer with disadvantages or penalties associated with exiting a relationship with an incumbent provider and we classify these as negative sources of constraint. Each of the three major types of switching costs may again be characterized by some negative factors. Negatively perceived relational switching costs involve normative penalties where the customer is obligated to stay because her/his friends and family or society expect him/her to. For example, Fitzmaurice (2005) shows that customers, who consider migrating to a new firm are less eager to do so when they feel that it would bring disapproval from important people in their lives. Such relational switching costs also include customer perceived psychological penalties where the customer is obligated to stay in a relationship with a provider because s/he wants to avoid confrontations or change. Negative perceived financial switching costs concern financial penalties due to contracts and the sunk cost in an existing provider relationship or consumption-related phenomena. An example might be the costs of replacing co-assets in which the customer has invested, such as peripherals (e.g. a new docking station, travel chargers, and adapters for a brand of laptop). Finally, negative perceived procedural switching costs involve transactional penalties, which entail the effort and time a customer needs to incur if s/he wants to exit a relationship with a product/service provider. For example, closing a bank account involves a significant amount of paperwork and investment of time. Such negatively perceived procedural switching costs may also include risk reassessment penalties, which involve the effort the customer would need to make to evaluate the risks involved with the consumption of other alternatives.
As mentioned earlier, switching costs are a result of a consumer’s subjective evaluation of experiences in a relationship. The question whether customers perceive all types of switching costs similarly or differently is important since a differential perception of switching costs might have an effect on loyalty and behaviour to the firm.
a coping behaviour that either functions to alter a troubled person-environment relationship or to sustain a desired one (Lazarus 1991). Further than that, this mechanism assumes that the individual person has a personal stake in an event that either facilitates or thwarts that stake (Lazarus 1991). In this case, the stakes are high as customers face both advantages and disadvantages, that can affect the customer’s relationship with the firm both in a positive and negative way.
We therefore argue that positive and negative switching costs function as antecedents to positive and negative emotions. White and Yu (2005) find that positive emotions are negatively related with switching behaviour whereas for negative emotions the effect appears to be contrary, more precise there is a positive association with switching behaviour. Moreover, Jones et al. (2007) show that commitment towards a firm can lead to different emotional outcomes. Calculative commitment, reflecting a feeling of being “locked-in”, increases negative emotions, whilst affective commitment, displaying a “desire” and “want” to stay, increases positive emotions. As positive switching costs provide the customer with advantages of staying in a relationship and indicate the willingness of the customer to stay, s/he will show positive emotions. Contrary to that, negative switching costs are associated with negative emotions as they interfere with the customer’s desired goal of exiting a relationship. Hence,
H1a: Positive switching costs lead to positive emotions. H1b: Negative switching costs lead to negative emotions. 2.2.3 Benefits and costs of leaving and – staying
As mentioned earlier, we suggest that the switching costs relationship between customer and provider is further characterised and influenced by both benefits and costs of leaving the provider or staying with him.
gains, or in this case benefits (Ho, Lim, and Camerer 2006; Thaler 1980). As a consequence, the value function of each customer is steeper for losses than for gains (Nagengast et al. 2014). Furthermore, in situations of choice this phenomenon often refers to an “Endowment Effect” as individuals tend to value objects that they possess more than objects they do not possess (Thaler 1980). The loss in utility is therefore larger for giving up something that you own, compared to the gained utility of receiving something equivalent, that has not been owned yet.
Literature on loyalty intentions uses prospect theory to determine the potential effect of emotional reactions to customer status changes. Reducing the status of a customer has found to cause negative emotions (Wagner, Henning-Thurau, and Rudolph 2009). As in our case customers are not confronted with changes in terms of status but with regards to advantages and disadvantages of leaving or staying, which can increase or decrease the perceived value of the relationship to a provider. We therefore argue that based on prospect theory customers weigh the individual benefits and costs of leaving and staying in order to make a decision, which subsequently generates emotions.
Frijda (1986) finds that blocking goals, desires, or rights of an individual typically elicits negative emotions, such as anger, frustration, or sadness (Frijda 1986). Similarly, because costs of leaving/staying refer to the disadvantages that constrain the customer in reaching her/his desired goals, they should elicit negative emotions. Conversely, benefits of leaving/staying should evoke positive emotions as they constitute the advantages of staying with the provider or leaving him. Hence,
H2a: Benefits of leaving/staying lead to positive emotions. H2b: Costs of leaving/staying lead to negative emotions.
2.2.4 Delimitation of positive/negative switching costs from benefits/costs
For the sake of clarity, this paragraph shall explain how positive/negative switching costs are related to benefits/costs of leaving/staying and how these two classifications are distinguishable in the later analysis.
Benefits of leaving refer to advantages of leaving the current provider, whereas benefits of staying pertain to the advantages of staying in the current relationship. Conversely, costs of leaving represent disadvantages of exiting the relationship, whilst costs of staying constitute disadvantages of staying with the current provider.
switching costs provide the customer with disadvantages of leaving the provider and pertain to the constraint that makes a customer feel trapped (“locked-in”) in a relationship.
Overall positive/negative switching costs are therefore related to benefits/costs of staying/leaving. Still they can be different and independent from each other as a consumer might experience both benefits/costs, but overall have positive switching costs as s/he might willingly stay. For this reason, we test both positive/negative switching costs as well as benefits/costs of staying/leaving independently from each other.
2.2.5 Prisoner, churner, positive stayer, and rational stayer
Other than the type of switching costs leading to emotions and behaviour, it is important to consider that the perception of switching costs can lead to certain emotions and behaviour. One important aspect of switching costs is therefore the degree to which customers perceive that they are forced to be locked in a relationship, rather than willingly staying and wanting to be in a relationship (Hirschman 1970). A willing customer may be said to be exhibiting affective commitment and have a favourable attitude towards the firm’s offerings (Dick and Basu 1994). We classify these behaviourally loyal customers, which also have a positive attitudinal loyalty, as “positive stayers” because they are exhibiting signs of having positive brand/relationship reasons to stay despite having lower levels of satisfaction with the firm’s products/services, hence have positive switching costs.
However, a customer may be dissatisfied and may wish to exit the relationship but cannot do so due to switching costs. Dick and Basu (1994) classify this set of customers as “spuriously loyal”, who show calculative commitment and would switch if they get a chance and if they cannot, they might resort to retaliatory behaviour against the firm (Haj-Salem and Chebat 2014). We classify these customers as “prisoners” as they are exhibiting lower levels of attitudinal loyalty than assured by their satisfaction with the firm’s products/services, though still remain customers, thus they are exhibiting signs of being trapped and are unhappy about that, hence have negative switching costs.
Contrary to “positive stayers” and “prisoners”, “churners” show little to no loyalty at all and hence leave the provider. Consequently, “churners” can be described as not exhibiting any costs in switching from one provider to another.
rational decisions when weighing benefits and costs of leaving or staying, but stay in the end, as it appears to be the ultimately better choice from a rational point of view.
Lastly “prisoners” and “positive stayers” can mainly be distinguished from “churners” based on the idea of excess- and deficit loyalty. Fader and Schmittlein (1993) find that high-market-share brands have significantly greater loyalty than the levels that would be expected on the basis of a popular consumer purchase model. Similar to this idea, we assume that “prisoners” and “positive stayers” both exhibit significantly greater loyalty than the levels that would be expected on the basis of their satisfaction, i.e. excess loyalty. On the other hand, “churners” exhibit significantly lower loyalty than the levels that would be expected based on their satisfaction, hence deficit loyalty.
Figure 1 depicts the satisfaction-loyalty relationship and the classifications.
Figure 1: Diagram of satisfaction-loyalty relationship and classifications
2.3 Behaviour
2.3.1 Complaint behaviour
company or towards a third entity” and thereby differentiate the recipient of complaint behaviour. According to Crié (2003) complaint behaviour towards a third entity can be understood as negative WOM, which will be dealt with in a subsequent section.
Inward negative emotions are elicited when a customer holds himself/herself responsible for the negative outcome (Lazarus 1991; Tangney and Dearing 2002) and hence reflect the internal attribution of responsibility (Barclay, Skarlicki, and Pugh 2005; Oliver 1993). Outward negative emotions on the other hand are defined as blaming the other party for the situation/outcome (Smith, Haynes, Lazarus, and Pope 1993). Haj-Salem and Chebat (2014) study the effect of switching costs as antecedent to inward and outward emotions and find evidence for two routes through which positive and negative switching costs affect loyalty/exit behaviour. Negative switching costs generate the coping behaviour exit and desire for revenge (Haj-Salem and Chebat 2014), though as exit due to negative switching costs is simply not available, customers might therefore seek revenge through complaining (Grégoire and Fischer 2008). Furthermore, Chebat, Davidow, and Borges (2011) show that these customers exhibiting negative switching costs may be trapped but will take any opportunity to complain. Hence,
H3a: Higher negative switching costs lead to a higher amount of complaints.
Conversely as positive switching costs provide the customer with advantages of staying in the relationship, they trigger inward negative emotions, which motivate the customer to stay with the actual service provider. Thus, compared to negative switching costs, they should engage in less coping behaviour, such as exit behaviour, and therefore show a lower amount of complaint. Hence,
H3b: Higher positive switching costs lead to a lower amount of complaints. 2.3.2 Word-of-mouth behaviour
minimize the stress arising from a stimulus (Lazarus and Folkman 1984). Customer coping behaviours may range from escape, to positive reinterpretation, and overtly negative acts like confrontation (Yi and Baumgartner 2004).
In this case customers might want to exit the relationship with the provider and perceive staying with the provider “unwillingly” as a negative situation. This also refers to the situation of being “locked in”, where being “locked-in” a dissatisfying relationship may lead to engagement in harmful behaviour towards the provider (Jones et al. 2000). Negative switching costs generate the coping behaviour exit and desire for revenge (Haj-Salem and Chebat 2014), though as exit due to negative switching costs is simply not available, customers might therefore seek revenge through making negative remarks to their friends, thus engage in negative WOM behaviour (Grégoire and Fischer 2008). Furthermore, as negative switching costs generally reflect positive attachment, i.e. loyalty to the respective firm or provider, paired with low satisfaction due to potentially perceived costs of leaving, negative switching costs should decrease positive WOM and increase negative WOM. Hence,
H4a: Negative switching costs lead to a lower amount of positive WOM than negative
WOM.
Conversely, if a customer is “willingly” staying in a relationship, i.e. exhibiting positive switching costs, s/he may engage in positive WOM. However, this may not always be true as customers might be staying in the relationship for reasons, that are not linked to the product, service or firm itself. For instance, customers may choose a bank nearby simply due to convenience of the location offered and may not necessarily be attracted by the actual service. In that case firms do not necessarily benefit from higher positive WOM of these customers. Likewise, customers may choose an airline due to its attractive loyalty program even if there are few differences in the attractiveness of the overall service with respect to its competitors. Moreover, positive switching costs generally reflect satisfaction and positive attachment, i.e. generate loyalty (Haj-Salem and Chebat 2014) to a firm due to perceived advantages of the relationship, which may increase positive WOM and decrease negative WOM.
H4b: Positive switching costs lead to a higher amount of positive WOM than negative
2.4 Control variables
Our hypothesized relationships might be affected by variables, such as time and industry. The following paragraph shall explain why it is important to control for these two factors. Since these variables are only added in order to control for their potential effects, we do not formulate hypotheses on the control variables.
2.4.1 Time
As switching costs are defined as the one-time costs, that customers associate with the process of switching from one provider to another (Burnham et al. 2013) we assume that this process, i.e. the ease of switching a provider has changed over time. This might be due to technical changes, such as number portability service in the mobile communications market, which has been found to significantly decrease switching costs since number portability has been in force (Lee, Kim, Lee, and Park 2006). On the other hand, contractual agreements, such as payment plans in the telecommunication industry, are proactively used by providers in order to lock-in customers and increase switching costs (Klemperer 1987). Moreover, as market-related variables, such as product alternatives and competition, change over time, they can reduce switching costs by affecting the customer’s quality perception of relationship and offerings. Furthermore, increasing market knowledge of customers indicates the growing awareness of switching costs (Pick and Eisend 2014).
Considering that with time the extent to which switching costs can be exhibited as well as the process of switching have changed, we use time as a control variable in the second study. The first study does not allow to control for time, as this information is not given for study one.
2.4.2 Industry
telecommunication standards make switching incredibly difficult, hence create switching costs (Cullen and Shcherbakov 2010). More importantly it also needs to be considered, that switching costs do vary and tend to be higher for services than for goods (De Ruyter, Wetzels, and Bloemer 1998).
Although the data of study 1 provides industry information, we will not control for industry effects, as the majority of narratives refers to the telecommunication industry and hence does not show enough variance. Study 2 will control for industry effects as the variation and diversity of industries could have significant effects on the tested relationships.
2.5 Overall framework and hypotheses
This chapter shall provide the reader with an overview of the overall framework, the conceptual model of each study as well as the associated hypotheses.
As indicated in the introduction, this paper is divided into three studies that jointly build the overall framework (see figure 2). Study 1 aims to determine how switching costs lead to positive and negative emotions. Study 2 subsequently looks at how switching costs serve as antecedent for subsequent complaint behaviour as well as positive and negative WOM behaviour. Past literature, as extensively reviewed in the preceding chapter, outlines the relationship between emotions and behaviour. Study 3 sheds light on how benefits and costs, which are related to switching costs, lead to emotions. Overall, we therefore aim to explain how switching costs lead to certain behaviour (study 2), based on the causality that this effect goes through emotions (study 1).
2.5.1 Study 1
In light of the research questions study 1 focusses on the following three parts. First, we aim to find out how the identified groups differ with regards to financial (F), relational (R) and procedural (P) switching costs. Second, we test for a positive as well as negative perception of each of the three switching cost dimension (F, R, and P). Third, we determine which emotional outcome positive and negative switching costs have.
Study 1 therefore consists of the following hypotheses, which are visualized in the conceptual model (see figure 3):
H1a: Positive switching costs lead to positive emotions.
H1b: Negative switching costs lead to negative emotions.
Figure 3: Conceptual model study 1
2.5.2 Study 2
As previously discussed, switching costs reflect (dis-)satisfaction and positive or negative attachment to a provider and hence generate coping behaviours, such as exit and desire for revenge (Haj-Salem and Chebat 2014; Jones et al. 2000). Study 2 therefore aims to find out whether positive/negative switching costs lead to complaint behaviour as well as positive/negative WOM behaviour. Furthermore, we aim to distinguish whether the identified groups differ with respect to their behaviour.
H3a: Higher negative switching costs lead to a higher amount of complaints.
H3b: Higher positive switching costs lead to a lower amount of complaints.
H4a: Negative switching costs lead to a lower amount of positive WOM than negative WOM.
H4b: Positive switching costs lead to a higher amount of positive WOM than negative WOM.
Figure 4: Conceptual model study 2 Figure 5: Conceptual model study 2
2.5.3 Study 31
Study 3 finally looks at the hypothesized effect of benefits/costs on emotions. First, we determine whether any of the customer groups perceives benefits/costs differently. Second, we show which emotions arise through benefits and costs and how these differ per group.
Study 3 therefore tests the following hypotheses, which are visualized in the conceptual model (see figure 6):
H2a: Benefits of leaving/staying lead to positive emotions.
H2b: Costs of leaving/staying lead to negative emotions.
Figure 6: Conceptual model study 3
1
STUDY 1: Switching costs as antecedents of emotions
3.1 Research design2 3.1.1 Data
To test the hypotheses, we analyse secondary qualitative data from a questionnaire, which has been carried out in a US-American college mall. The sample consists of 40 female and 60 male students from the University of Indiana, aged between 19 and 21, who were asked to describe the last time they wanted to switch a provider and why they did or did not. The data is unstructured in a sense that narratives are not clearly numbered and assigned to an individual respondent. After thorough reading and deleting two invalid narratives, i.e. those that do not give an answer to the above mentioned question, we end up with a total sample of 91 narratives from individual respondents.
3.1.2 Narrative coding
In light of the aforementioned research questions we coded the narratives on the one hand manually and on the other hand by means of the advanced text analytics computer program (LIWC). In order to increase face validity of the narrative coding the narratives have been coded individually and independently by three different persons. Afterwards all three codings showed little deviance from each other and were merged under close alignment ending up with a final narrative coding that serves as a basis for study 1, 2 and 3.
3.1.2.1 Human coding
According to past theory we first classify each narrative on the different types of switching costs, that a consumer can exhibit, namely financial, relational and procedural switching costs. In that case, it is important to mention, that those types are not mutually exclusive from each other, but can be experienced all at the same time. Further than that, we also check whether the respondent is clearly expressing benefits and/or costs of leaving and –staying. Here again, benefits/costs are not mutually exclusive from each other. Having classified each narrative on these two dimensions, we assign either positive or negative switching costs to each narrative. Lastly, based on the described classifications (chapter 2.2.5) each narrative is assigned to one of the four different type (prisoner, positive stayer, rational stayer or churner). Throughout the classification we determined a fifth type, which is different from the four given ones, as this respondent first exhibits negative switching costs, i.e. feels “locked-in”, but in the end still exits
2
the relationship, i.e. churns. We classify this type as “First prisoner, then churner” (FPTC). This follows from the fact, that the time frame in which the narratives are formulated may have not been clearly defined, as the time frame of being a FPTC is significantly longer, then the one of a prisoner.
In order to give the reader a clear understanding and picture of how the narrative coding has been performed, the appendix 3.1.2.1 provides an exemplary narrative including its coding.
3.1.2.2 Linguistic Inquiry and Word Count (LIWC) analyses
We ran each narrative through the Linguistic Inquiry and Word Count (LIWC) software to extract quantitative data about the underlying emotions of each narrative. The text analysis is based on the dictionary “Internal Dictionary 2015”, which is composed of almost 6,400 words, word stems, and selected emoticons. LIWC (pronounced “Luke”) reads files word by word, matching each word against a dictionary of words that defines different word types, such as positive and negative emotional words. The output consists of 80 distinct measures varying from general descriptors (e.g. total word count) to linguistic elements (e.g. auxiliary verbs) and psychological constructs (e.g. cognitive words). We focus on the emotional output from LIWC and receive a percentage value for positive and negative emotions in relation to the total number of words per narrative. We further divide this value by the total number of words per narrative in order to quantify emotionality and receive a relative score of positive and negative emotional words, that allows for comparison across all 91 narratives.
Past literature has shown that emotionality can be drawn from textual expression, though the context is still important to consider, as LIWC results of an article on “anger management” likely have little to do with how angry the author was during writing (Pennebaker Conglomerates, Inc. 2018). Hirsh and Peterson (2009) show that word usage during the production of self-narratives were significantly associated with the Big Five personality traits across a variety of psychological categories and thereby indicate a strong connection between language use and personality. Kahn, Tobin, Massey, and Anderson (2007) prove that word use is a meaningful indicator of emotion, whereas Alpers, Winzelberg, Classen, Roberts, Dev, Koopman, and Taylor (2005) show that LIWC ratings of positive and negative emotional words correspond with human ratings of the writing excerpts.
3.1.3 Variable operationalization
Financial SC Monetary costs in moving from one product/service provider to another 1 = applicable, 0 = not applicable Relational SC Personal and brand relationship loss associated with switching providers 1 = applicable, 0 = not applicable Procedural SC Time and effort necessary by the customer in order to evaluate, learn and consume alternatives as well as
related setup costs 1 = applicable, 0 = not applicable Positive SC Advantages of staying with the current provider and indicator of willingness to stay. 1 = applicable, 0 = not applicable Negative SC Disadvantages associated with leaving the current provider and negative source of constraint. 1 = applicable, 0 = not applicable Word count (WC) Total number of words per narrative LIWC coding (in %)
Positive emotions Percentage value of positive emotions in relation to total number of words per narrative LIWC coding (in %) Negative emotions Percentage value of negative emotions in relation to total number of words per narrative LIWC coding (in %)
Relative positive emotions Amount of positive emotions per narrative, that allows for comparison of all narratives Positive emotions / WC (in %) Relative negative emotions Amount of negative emotions per narrative, that allows for comparison of all narratives Negative emotions / WC (in %)
Classification Definition Measurement items
Prisoner An unhappy customer that is constrained in leaving the provider by negative switching costs, exhibits excess loyalty and a relatively low satisfaction level. 1 = applicable, 0 = not applicable Positive stayer A happy customer that is willingly staying with the provider, exhibits positive switching costs, excess loyalty
and a relatively high satisfaction level. 1 = applicable, 0 = not applicable Rational stayer A customer exhibiting neither excess nor deficit loyalty, that is neither completely satisfied nor dissatisfied and
stays for rational reasons. 1 = applicable, 0 = not applicable Churner A customer that leaves the provider and exhibits deficit loyalty. 1 = applicable, 0 = not applicable First prisoner, then churner (FPTC) A customer that first shows signs of a prisoner, i.e. being locked-in and not able to leave, but in the end still churns 1 = applicable, 0 = not applicable
After coding the narratives, we can determine that 35.16 % of the narratives exhibit positive switching costs whereas the majority of the narratives (64.84 %) exhibits negative switching costs. Furthermore, almost 88 % of the respondents describe having financial switching costs, whereas 35.16 % experience relational switching costs and 31.87 % exhibit procedural switching costs, what gives an indication on the importance of monetary costs in moving from one provider to another (see table 2).
Variable Frequency Relative frequency (in %)
Positive switching costs 32 35.16 % Negative switching costs 59 64.84 % Financial switching costs 80 87.91 % Relational switching costs 32 35.16 % Procedural switching costs 29 31.87 %
Table 2: Descriptive statistics of independent variables
The majority of the narratives (36.26 %) is classified as prisoner, 26.37 % of the narratives are positive stayers, 12.09 % are rational stayers, 12.09 % are churners and 13.19 % are coded as FPTC (see table 3).
Classification Frequency Relative frequency (in %)
Prisoner 33 36.26 % Positive stayer 24 26.37 % Rational stayer 11 12.09 % Churner 11 12.09 % First prisoner, then churner (FPTC) 12 13.19 %
Table 3: Descriptive statistics of classifications
Each narrative has an average word count of 200, whereas the shortest narrative consists of 59 words and the longest one comprises 348 words, indicating the varying extent to which the respondents formulate their experience. On average, the narratives show a higher amount of positive emotions (µ = 2.4) than negative emotions (µ = 1.1) and additionally positive emotions also have a higher range (see table 4).
Variable Mean Min. Max. Std. dev.
Word count (WC) 200 59 348 52.596 Positive emotions 2.4 0.00 5.94 1.308 Negative emotions 1.1 0.00 2.94 0.779 Relative positive emotions 0.013 0.00 0.063 0.011 Relative negative emotions 0.006 0.00 0.030 0.006
3.1.5 Method
In order to test our hypotheses several statistical methods are applied. The following paragraphs shall motivate and describe the choice of each research method, that will be applied throughout the analysis.
3.1.5.1 Analysis of variance (ANOVA)
One-way analysis of variance (ANOVA) post-hoc tests are used to determine whether and how our five identified groups (churner, positive stayers, rational stayers, prisoners and FPTC) show statistically significant differences with regards to financial, relational and procedural switching costs as well as benefits/costs of leaving/staying. As we are performing the test on five different groups simultaneously, we apply multiple-comparison post-hoc correction, more precisely Bonferroni correction, in order to compensate for type 1 errors. Furthermore, one-way ANOVAs are used in order to determine significant associations of positive and negative switching costs with positive/negative emotions and benefits/costs of leaving/staying.
3.1.5.2 Linear regression model(s)
As one-way analysis of variance is limited to mean comparison we apply multiple linear regression models in order to find evidence for significant relationships between switching costs (IV), benefits/costs of leaving/staying (IV) and emotions (DV).
3.1.5.3 T-test
A t-test is applied in order to find evidence for the hypothesized positive and negative perception of financial, relational and procedural switching costs. The mean of each switching cost dimension is therefore mapped onto positive/negative switching costs, with the aim to find significant differences. As the sample size and variance are unequal, where Student’s t-test can be unreliable, a Welch’s two sample t-test is performed.
3.2 Results
This chapter presents the results from the above mentioned methods that were applied in order to determine whether the groups show different behaviour as well as which emotional outcome positive and negative switching costs have.
3.2.1 Segmentation on financial, relational and procedural switching costs
(see table 5), it appears that the five groups are not significantly different from each other on either the financial (p = .143), relational (p = .367) or procedural (p = .908) switching cost dimension. Despite the fact, that none of the results is not even moderately significant, we can determine that the groups are most likely to be different on the basis of financial switching costs. This result might be due to the little variance in the dataset, considering that it only consists of 91 observations. An extension might increase the power and make this analysis richer.
Study 3.2.1 ANOVA
Sum of Squares df Mean Square F Sig. Financial SC Between Groups .735 4 .184 1.768 .143 Within Groups 8.936 86 .104 Total 9.670 90 Relational SC Between Groups 1.001 4 .250 1.090 .367 Within Groups 19.746 86 .230 Total 20.747 90 Procedural SC Between Groups .228 4 .057 .251 .908 Within Groups 19.530 86 .227 Total 19.758 90
Table 5: ANOVA output and means for study 3.2.1
3.2.2 Perception of financial, relational and procedural switching costs
As we argue that financial, relational and procedural switching costs (SC) can be perceived both positively and negatively we perform a t-test on the means of positive and negative switching costs on each switching cost dimension. Looking at the results (see table 6) it seems that only procedural switching costs have a meaningful and significant (p = .036) difference in perception. As they are perceived more negatively (µ = 0.39) than positively (µ = 0.19) we support Jones et al. (2007) finding that procedural switching costs are perceived negatively and conclude that they can also be perceived positively. Financial switching costs turn out to be slightly more negative (µ = 0.90) perceived than positive (µ = 0.84). However, since the t-test is not significant (p = .4779) we cannot confirm the finding of Jones et al. (2007), that financial switching costs are perceived positively. With regards to relational switching costs we again observe that the test statistic turns out to be insignificant (p = .2236), the results though say that relational switching costs are perceived more positive (µ = 0.43) than negative (µ = 0.31).
Study 3.2.2 Positive SC Negative SC t Df p-value Means Financial SC 0.71463 54.238 0.4779 0.8437 0.8983 Relational SC -1.2298 59.391 0.2236 0.4375 0.3051 Procedural SC 2.1309 76.029 0.0363 * 0.1875 0.3898 * Bold printed t-values are significant at 5 % sig. level
Table 6: T-test output for study 3.2.2
3.2.3 Emotional outcome of positive and negative switching costs
Our findings (see table 7) show that switching costs (SC) in general are significantly more associated with positive emotions (p = .0459) than with negative emotions (p = .859). The results of our regression (see table 8) show that switching costs have a significant (p = .0459) and positive (β-estimate = .0046) effect on positive emotions, thus we cannot find a significant (p = .859) effect on negative emotions.
The found effect can be explained as follows. As the variable switching costs varies between 1 (= negative switching costs) and 2 (= positive switching costs), whereas the variable emotions is continuous, an increase in switching costs by 1 (1 to 2) significantly increases positive emotions, but not negative emotions. We can therefore conclude that positive switching costs increase positive emotions and accept H1a. With regards to negative switching costs we look at the reverse effect (2 to 1). A decrease in switching costs by 1 significantly decreases positive emotions, but not negative emotions. We can therefore conclude that negative switching costs decrease positive emotions, why H1b is only partly supported. The summary of means (see table 9) confirms this finding, where the average of positive emotions is greater than the average of negative emotions for both positive and negative switching costs.
Considering that positive switching costs provide the customer with advantages of staying and hence give a reason to be rather satisfied than unsatisfied the found effect seems reasonable. Surprisingly negative switching costs do not influence negative emotions, but rather decrease positive emotions. This indicates that switching costs rather affect positive than negative emotions.
Study 3.2.3
Positive emotions Negative emotions
Df Sum Sq Mean Sq F value p-value Df Sum Sq Mean Sq F value p-value SC 1 0.0004 0.0004 4.098 0.0459 * 1 0.0001 9.850e-07 0.032 0.859 Res. 89 0.0095 0.0001 89 0.0028 0.0003
Study 3.2.3
Positive emotions Negative emotions
Estimate S.E. t value p-value Estimate S.E. t value p-value (Intercept) 0.0071 0.0033 2.179 0.0320 * 0.0059 0.0018 3.347 0.0012 ** SC 0.0046 0.0023 2.024 0.0459 * 0.0002 0.0012 0.178 0.8590 F(1, 89) = 4.10 R2 = 4.40 % Adj. R2 = 3.33 % F(1, 89) = 0.03 R2 = 0.04 % Adj. R2 = -0.11 % * Bold printed coefficients are significant at 5% sig. level
Table 8: Regression output of study 3.2.3
Study 3.2.3
Positive switching costs Negative switching costs Obs Mean Std. Dev. Min Max Obs Mean Std. Dev. Min Max Positive
emotions 32 0.1626 0.0108 0 0.0595 59 0.1167 0.1009 0 0.0625 Negative
emotions 32 0.0063 0.0059 0 0.0298 59 0.0061 0.0054 0 0.0212
Table 9: Summary of means for study 3.2.3
3.3 Discussion
In this chapter we discuss the results of study 1 and consequently answer the research questions. We cannot find any significant difference between the five
identified groups with regards to financial, relational and procedural switching costs. The degree to which the three
switching costs dimensions are being perceived therefore does not significantly differ. However, we detect that almost 88% of all narratives mention financial switching costs, whereas the other two dimensions are only mentioned in around 30% of the cases. This indicates a higher importance of monetary costs in the switching process compared to procedural and relational switching costs.
Moreover, we find evidence for a significant relationship between switching costs and emotions. We discover that positive switching costs increase positive emotions and thereby confirm Jones et al.
(2007) finding that affective commitment, i.e. the willingness to stay, increases positive emotions. Contrary to prior literature (Jones et al. 2007) we find that negative switching costs decrease positive emotions. The feeling of “being locked-in” therefore seems to rather affect positive than negative emotions.
Table 10 and 11 provide an overview of the hypotheses, their expected and found relationship as well as on overview of the observed effects in study 1. As this study shows which emotional outcome is consequential for positive and negative switching costs, study 2 investigates behavioural consequences of switching costs.
Hypothesis Supported Expected relationship Found relationship
H1a Supported + + H1b partly supported + -
Table 10: Overview of hypotheses
Study Effect observed
Financial, relational and procedural switching costs
Segmentation - No significant between group difference - Mean of F (88%), R (35%) and P SC (32%)
Perception - Significant more negative perception of procedural switching costs
- No significant difference in perception of financial and relational switching costs
Positive switching costs
Positive emotions Significant and positive effect Negative emotions No significant effect
Negative switching costs
Positive emotions Significant and negative effect Negative emotions No significant effect
STUDY 2: Switching costs as antecedents of behaviour
4.1 Research design 4.1.1 Data
To test the hypotheses, we make use of secondary data from the American customer satisfaction index (ACSI), which is a uniform, national, cross-industry measure of satisfaction with the quality of goods and services available to household consumers in the United States (ACSI Methodology Report). ACSI provides customer satisfaction indices (on 0 to 100 scales), indices of antecedents (drivers or causes) and outcomes of satisfaction with the products and services of over 200 companies, government agencies, and 41 industries. Each company or government service, industry, and sector is measured annually, whereas the national index is updated quarterly. ACSI then measures satisfaction with large companies in representative industries by means of customers of these companies, which are identified by interview screening and are then interviewed about their satisfaction with the specific company.
The dataset contains 2,982 observations over a time period of 30 years, i.e. 1983 - 2013. Several categorical variables give information on the company, industry, stock exchange code etc. Further than that, indices on customer expectations, perceived product/service quality, perceived value, customer satisfaction, customer complaints and customer loyalty are split into individual variables, that each apply to a survey question, which the consumer is asked to answer. More information on each of these and for this study important variables follow in the subsequent section.
4.1.2 Variable operationalization
The ACSI dataset contains data on complaint behaviour, which is classified as a manifest variable (survey question) and measures whether or not a customer has complained to the company within the same time period that qualifies the potential respondent as a customer. After preparing and cleaning the dataset we end up with 1,039 observations, which is partly due to a large amount of missing values. Furthermore, the dataset contains data on positive and negative WOM behaviour. Of the total 2,982 observations there are 2,557 missing values for positive WOM and 2,790 missing values for negative WOM, why we end up with 357 and 162 observations respectively.
