The impact of dynamic pricing on perceived price fairness,
transaction specific satisfaction and customer loyalty.
The moderating role of individual level factors; privacy concerns, price sensitivity and product involvement.
Written by: Joline Nijholt Studentnumber: 5883520 Date of submission: 30th January 2015
Msc Business Administration – Marketing University of Amsterdam Supervisor: drs. F. Slisser
This document is written by Student Joline Nijholt who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
This study aimed to extend the scarce literature about dynamic pricing by focusing on the moderating role of three individual level factors; (1) privacy concerns, (2) price fairness, and (3) product involvement on the relation between dynamic pricing, perceived price fairness, transaction satisfaction and customer loyalty. Respondents between 18 and 65 were invited to fill out an online questionnaire which included four different purchase scenarios. 170
respondents completed the online questionnaire and were randomly assigned to the four scenarios. Results show that dynamic pricing negatively influences perceived price fairness and that perceived price fairness mediated the relation between dynamic pricing and
transaction satisfaction. Also, customers who perceived price as being fair were more satisfied with their transaction and more loyal to the company. Transaction satisfaction appeared to mediate the relation between perceived price fairness and customer loyalty. Privacy concerns and price sensitivity were found to moderate the relation between dynamic pricing and perceived price fairness, whereas no moderating effect was found for product involvement. Finally, some suggestions for future research are given.
Key words: dynamic pricing, perceived price fairness, satisfaction, customer loyalty, price sensitivity, privacy concerns, product involvement
Table of contents
Abstract ... 3
1 Introduction ... 8
2 Literature review ... 12
2.1 Dynamic pricing ... 12
2.2 Perceived price fairness ... 13
2.3 Transaction satisfaction ... 16
2.4 Customer loyalty ... 18
2.5 Moderating individual level factors ... 22
2.5.1 Moderating role of privacy concerns... 22
2.5.2 Moderating role of perceived price sensitivity ... 24
2.5.3 Moderating role of product involvement... 26
3 Method ... 28 3.1 Research design………..28 3.2 Sample………...…..32 3.3 Measurement of variables ... 35 3.4 Statistical procedures………..37 4 Results………...40 4.1 Correlations………...……….40 4.2 MANOVA………...43 4.3 Direct effects………..44 4.4 Mediation………49 4.5 Moderation……….50
5 Discussion……….….54
5.1 Conclusion………..54
5.2 Practical and theoretical implications……….56
5.3 Limitations………..59
5.4 Future research………...60
References ... 62
Appendix A A.1 Pre-test questionnaire (English) ... 69
A.2 Pre-test questionnaire (Dutch)………...76
A.3 Main questionnaire (English) ... 83
A.4 Main questionnaire (Dutch)………...90
A.5 Purchase scenarios ... 97
Appendix B Analysis outcomes mediation (Table 7/8/9/10)………...100
List of tables and figures
TablesTable 1 Demographic characteristics of main survey 33
Table 2 Means, Standard Deviations, Correlations (two-tailed) and Scale Reliabilities 42
Table 3 MANOVA results 44
Table 4 Hierarchical regression model for perceived price fairness 47
Table 5 Hierarchical regression model for perceived transaction satisfaction 48
Table 6 Hierarchical regression model for customer loyalty 49
Table 7 Bootstrap results for mediating effect of perceived price fairness on the 100
relation between dynamic pricing and transaction satisfaction Table 8 Regression model for transaction satisfaction (Bootstrapping 2,000 resample) 100 Table 9 Bootstrap results for mediating effect of transaction satisfaction on the 101
relation between perceived price fairness and customer loyalty Table 10 Regression model for customer loyalty (Bootstrapping 2,000 resample) 101
Table 11 Results for privacy concerns as moderator of the relation between dynamic 52
pricing and perceived price fairness Table 12 Results for price sensitivity as moderator of the relation between dynamic 52
pricing and perceived price fairness Table 13 Results for product involvement as moderator of the relation between 52
dynamic pricing and perceived price fairness Table 14 Bootstrap results for conditional indirect effect of perceived price fairness 54
on the relation between dynamic pricing and transaction satisfaction moderated by price sensitivity, privacy concerns and product involvement.
Figures
Figure 1 Conceptual model 9
Figure 2 Conceptual model: the influence of price sensitivity, privacy concerns 26
and product involvement on the relation between dynamic pricing – perceived price fairness – transaction satisfaction and customer loyalty.
1 Introduction
In 2007 customers of Apple found out that the company had dropped its price of the iPhone by $200 60 days after launch. Loyal customers were furious, as they felt they were punished for being loyal to Apple by buying their new iPhone right after its launch (Macintosh News Network, 2007). As companies have massively entered the Internet as a new medium for their businesses, marketers are confronted with the traditional marketing mix; product, price, place and promotion, wondering how they can effectively use these components in an online setting (Dai, 2010; Yelkur and DaCosta, 2001). Price has become an important factor in surviving the massive online competition. On the one hand companies are continuously striving to find the most effective pricing strategies, while on the other hand researchers are interested in customer reactions on specific pricing strategies (Dai, 2010; Xia, Monroe and Cox, 2004).
Among different pricing strategies, dynamic pricing is a common used version of price discrimination. Sellers try to maximize profit by charging customers different prices for very similar products according to the customers’ willingness to pay (Dai, 2010; Jallat and
Ancarani, 2008; Klein and Loebbecke, 2003). In September 2000 Amazon.com experimented with dynamic pricing, using their customer data to offer products for different prices to different customers. The e-tailer used customer data to imply customers’ willingness to pay and charged them accordingly. Although the new Internet technology and infrastructure makes it possible to use dynamic pricing strategies in a positive way, in the Amazon.com case customers found out that the firm offered exactly the same DVD movies at different prices to different customers, which led to a huge wave of bad publicity and unsatisfied customers (Adamy, in Cox, 2001; Dai, 2010).
Previous research on dynamic pricing shows that dynamic pricing is strongly related to perceived price fairness. Xia et al. (2004) found that there is a negative relation between dynamic pricing and perceived price fairness. They state that when customers find out that they are paying a higher price for a product than other customers, they will feel disadvantaged and perceive price as being unfair. Haws and Bearden (2006) confirm this statement as they found that customers perceive price differences as being unfair when it turns out that the discrepancy is to their disadvantage. According to Martín-Consuegra, Molina and Esteban (2007) price does not only influence perceived price fairness, but also customers’ satisfaction judgments. The extent to which customers are satisfied with their purchase depends on various factors like quality, service, price, environment and personal factors (Zeithaml and Bitner, 1996, in Martín-Consuegra et al., 2007). Satisfied customers are more likely to repeat purchases, build up a relationship of trust with the seller and “consider themselves as loyal customers” (Xia et al., 2004, p. 5). Even though price seems to be an important tool for marketers to become profitable, customer loyalty can also contribute to a firm’s sales as they form a stable income base to the firm (Dai, 2010; Xia et al., 2004).
As existing literature shows, there is already much known about the relationship between perceived price fairness, satisfaction and customer loyalty. Therefore this study will extend the existing literature by focusing on the influence of dynamic pricing and various factors moderating the relationship between dynamic pricing and perceived price fairness. Previous research concerning price fairness perceptions focused on the role of customer and transaction similarity (Xia et al., 2004) or the influence of the magnitude and proximity of price differences (Haws and Bearden, 2006). However, little attention has been paid on the processes prior to customers’ fairness perceptions. Different customer segments, in terms of price fairness perceptions, could be distinguished (Anderson, 1996 in Xia et al., 2004). For instance, customers may perceive and respond different to price changes (Monroe, 1973, in
Wakefield and Inman, 2003). Unfair perceived prices could lead to more price sensitivity among customers (Sinha and Batra, 1999, in Xia et al., 2004). Previous research on price sensitivity mainly covers the loyalty – price sensitivity relationship and looks at how loyalty can reduce price sensitivity (Krishnamurthi and Papatla, 2003). Yet, how customers’ level of price sensitivity can influence their perception of price fairness and their loyalty to the firm has not been searched.
Along with the increasing popularity of online shopping, firms are not only able to collect demographic information about their customers, but also gather information about what customers buy, where they look at, how they navigate through websites and what effect promotions have on customers’ online behavior (Dai, 2010; McAfee and Brynjolfsson, 2012). Research shows that a substantial percentage of internet users is to some degree concerned about which personal information is available to other users and how it is used or saved ( Brown and Muchira, 2004; Phelps, Nowak and Ferell, 2000). For instance, customers tend to respond more negatively when it turns out that a price discrepancy is due to a firms volitional actions (Bolton, Warlop and Alba, 2003). While most research has focused on privacy
concerns in a more general marketing sense, this research will look at how different levels of privacy concerns, in case of dynamic pricing, influence customers’ perceptions of price fairness and loyalty to the firm. At last, customers tend to hold on to different search and purchase behaviors according to their product involvement. Although research has been done which looked into the relationship of product involvement on satisfaction and customer loyalty, this has not been studied in the case of dynamic pricing. The concepts price
sensitivity, privacy concerns and product involvement will be discussed in more detail in the next chapter.
The purpose of this research is to investigate the formation of price fairness judgments toward dynamic pricing by looking at three individual level factors (price sensitivity, privacy concerns and product involvement) as moderators of the relationship between dynamic pricing and perceived price fairness. Also the effects on the already familiar relationship between perceived price fairness, transaction satisfaction and customer loyalty will be taken into account. Therefore, this study proposed the following research question:
Is the relationship between dynamic pricing and online perceived price fairness, transaction
satisfaction and customer loyalty moderated by individual level factors; privacy concerns,
price sensitivity and product involvement and if so, how?
The results of this study will be an academic contribution to the existing literature on customers’ price fairness perceptions and customer loyalty in case of dynamic pricing and give a better understanding of the formation of fairness perceptions and behavioral responses. The existing literature about customer loyalty and dynamic pricing will be extended by incorporating price sensitivity, privacy concerns and product involvement as illustrated in the conceptual model (Figure 1). Furthermore this study will give marketers, who consider dynamic pricing strategy, a clearer view of which customer characteristics should be taken into account and what behavioral responses can be expected.
2 Literature review
In this chapter the literature concerning dynamic pricing, price fairness perception, transaction satisfaction and the influence on customer loyalty will be described and discussed in more detail. Also, the individual level factors price sensitivity, privacy concerns and product involvement will be discussed as expected moderators of the relationship between dynamic pricing and perceived price fairness. The first section deals with a more general approach to dynamic pricing and perceived price fairness. The second section describes the role of transaction satisfaction. In the third section customer loyalty is discussed. In the fourth and last section the moderators of the dynamic pricing – perceived price fairness relationship are presented.
2.1 Dynamic pricing
According to Armstrong and Kotler (2000 in Dai, 2010) dynamic pricing is an individual-level price discrimination strategy where prices are charged according to customer, location, product, or time. Dynamic pricing is not a new concept for firms, but rather a method that has been used for years. However, the rapid speed in which technology has developed last decade brings all new opportunities for dynamic pricing (Garbarino and Lee, 2003; Jallat and
Ancarani, 2008; Klein and Loebbecke, 2003; Lee-Kelley and Gilbert, 2003). The Internet has rapidly transformed into an open medium which can be easily used to access information about customers. This customer data makes it possible to gain insight into customers’ wants and needs and give firms the opportunity to use this information in offering products to specific people for different prices (Dai, 2010; Xia et al., 2004). Companies can use previous online search behavior to estimate customers’ willingness to pay for products and use this information to set their prices accordingly. By charging customers according to their
willingness to pay, companies can maximize their profit (Garbarino and Lee, 2003; Jallat, and Ancarani, 2008; Klein and Loebbecke, 2003; Lee-Kelley and Gilbert, 2003). Although
dynamic pricing seems to be a profitable opportunity for firms, the open character of the Internet can give customers insight in this dynamic pricing process and influence their perception of price (Dai, 2010). This research will look at the consequences of dynamic pricing on price fairness perceptions, transaction satisfaction and customer loyalty. Special attention is paid on possible individual level factors influencing these outcomes.
2.2 Perceived price fairness
Customers are aware of firms charging different prices for products in certain situations. For instance, coffee is much more expensive at Starbucks than at AH to go and Ben & Jerry’s ice cream is much more costly than ordinary ice cream at the grocery store (Bijsterbosch, 2005; www.ah.nl/producten/diepvries/ijs). However, these price differences are based on time, geographical factors or other assignable reasons. Customers are willing to pay more for Ben & Jerry’s ice cream, because they know that the price is based on sustainable and fair products (Bijsterbosch, 2005). In case of dynamic pricing, prices are based on private customer information and therefore not identified as a valid reason to charge different prices to different customers (Garbarino and Lee, 2003). Customers’ perception of price fairness is defined by Xia et al. (2004) as the customers assessment of whether a perceived price can be reasonably justified. Thus, perceived price fairness can be seen as a subjective judgment, while it concerns customers perception of price regardless of the true price. Perceived price fairness can be influenced by various factors which will be discussed next (Dai, 2010; Xia et al., 2004).
Social comparison theory
Previous literature shows that customers’ perceptions of price fairness are based on comparisons with various reference points like previous prices, prices paid by others and prices of other companies. As according to Festinger’s (1954) social comparison theory,
people constantly use comparisons to justify their own choices. Most important are similar others in this comparison process, which mostly means people in the customers’ direct environment. With this in mind it is likely that customers also compare themselves with similar others, while forming their price fairness perceptions. When customers buy a product they will compare themselves with other customers who bought the same product. If they find out that they are paying a higher price than these similar others, customers will perceive price as being unfair (Bechwati, Sisodia, and Sheth, 2009, in Dai, 2010; Martin, Ponder and Lueg, 2009).
In a study by Xia et al. (2004) the influence of transaction similarity, comparative others and the buyer-seller relationship on customers’ price fairness perceptions are
investigated. Xia et al. (2004) state that customers not only compare themselves with similar others, but they also compare their transactions with similar transactions while forming fairness judgments. Transactions that are not similar do not seem to influence fairness perceptions as much as transactions that are highly similar (Xia et al., 2004). In case of dynamic pricing, where companies offer similar products for different prices to customers, this comparing behavior seems an important factor for companies. Customers, who
experience a negative price discrepancy between what they paid and what other customers have paid for the same product, are more likely to evoke negative emotions like anger, disappointment and switching behavior (Haws and Bearden, 2006).
Distributive and procedural justice
As mentioned before, customers’ perceptions of price fairness is an assessment of whether a perceived price can be reasonably justified (Xia et al., 2004). This means that customers make an assessment for themselves of what they perceive as fair and what they perceive as unfair. Homans (1961, in Dai, 2010) confirms this finding and refers to distributive and procedural
justice as mechanisms to form perceptions of fairness. He defines distributive justice as “the judgment of the allocation of rewards on the basis of individual contributions to an exchange relationship” (Homans, 1961, in Dai, 2010, p. 17). In other words, customers make a
consideration between buyers’ and sellers’ investment and outcome in an exchange relation, when they feel that there is a negative discrepancy between both, customers will perceive this as unfair (Homans 1961, in Dai, 2010). Procedural justice, on the other hand, refers to “the process, method and/or the rules used to determine the outcomes influence judgments of fairness perceptions” (Thibaut and Walker, 1975, in Dai, 2010, p. 17). In short, distributive justice is more about the balance between input and output, while procedural justice focuses more on the procedures which lead to certain outcomes.
Fairness perception research has shown that procedural justice plays a more important role in influencing customers’ perceptions of fairness. However, previous research also has shown that customers are little or not aware of the procedures or strategies that companies use to come to certain outcomes (Bolton et al., 2003). For that reason it is interesting to
investigate the influence of customers’ knowledge about companies’ pricing procedures on their perception of fairness some more. In this study we will specifically pay attention to the process of dynamic pricing, wherein companies use customers’ personal information to set their prices. We will look at customers’ perceptions of price fairness when they find out about this pricing procedure and investigate what influence this has on transaction satisfaction and customer loyalty.
Cognitive dissonance
The subjective character of price fairness perceptions, as the distributive justice mechanism points out, corresponds with Adams’ (1965) equity theory, which states that customers are concerned about the fairness of inputs and outputs of all parties involved in a transaction. A
disproportional relationship, that customers experience when they are charged a higher price than other customers for a similar product, will lead to perceived inequity. According to Adams (1965) “the presence of inequity will motivate the perceiver to achieve equity or to reduce inequity; and the strength of motivation to do so will vary directly with the magnitude of inequity experienced” (Adams, 1965, p. 283). In other words, customers will experience cognitive dissonance, as their expectations conflict with reality. In order to align these conflicting beliefs, customers will try to reduce dissonance either by; (a) changing their behavior; (b) justify behavior/cognition by changing conflicting cognition; (c) justify
behavior/cognition by adding new cognitions or (d) ignore any information that conflicts with existing beliefs (Festinger, 1962).
This study will investigate how customers perceive price in case of dynamic pricing. Expected is that customers, who find out that they are paying a higher price than other customers for a similar product, will be more likely to perceive price as unfair. Focus will be on the disadvantage of dynamic pricing as customers are more likely to show their negative responses than their positive ones (Dai, 2010; Xia et al., 2004). Expected is that customers’ individual level factors (price sensitivity, privacy concerns and product involvement) influence the degree to which a higher price leads to unfair price perceptions, which will be explained in more detail later. We state that:
H1: Dynamic pricing will negatively influence perceived price fairness. Specifically,
customers who experience dynamic pricing to their disadvantage will be more likely to
perceive price as unfair.
2.3 Transaction satisfaction
As previous literature shows, satisfaction is a widely searched but difficult to capture concept (Dai, 2010). According to Oliver (1997, in Anderson and Srinivasan, 2003, p. 125),
satisfaction is “the summary psychological state resulting when the emotion surrounding disconfirmed expectations is coupled with a consumers’ prior feelings about the consumer experience”. Which refers to if customers’ expectations about one situation are in line with their actual experience of the situation. Although this definition largely catches what satisfaction is about, researchers have not been able to come up with a widely accepted definition yet. Oliver (1997, in Dai, 2010) distinguishes transaction-specific satisfaction and overall satisfaction, whereas transaction-specific satisfaction refers to the satisfaction of a customer with a specific product purchase. Overall satisfaction, however, not only refers to customers’ satisfaction with the product purchase, but also satisfaction with the company, the brand and other factors of influence. This research will focus on transaction-specific
satisfaction, which will be referred to as ‘transaction satisfaction’ from now on, as we are interested in the effects of dynamic pricing in a specific purchase situation (transaction). The definition we will hold on to is “the contentment of the customer with respect to his or her prior purchasing experience with a given electronic commerce firm (Anderson and
Srinivasan, 2003, p. 125)”.
Previous satisfaction literature focused either on cognition as a predictor of
satisfaction, which referred to customers’ expectations about the performance of a product (Oliver, 1980; Westbrook, 1987, in Martín-Consuegra et al., 2007), or on affect as a predictor of satisfaction, which claims that consumption of the product can influence the customers’ satisfaction judgments (Homburg, Kocheta and Hoyer, 2006). This research will take both concepts in consideration as we think that both customers’ expectations of a product and the consumption (buying behavior) can indirectly influence the satisfaction judgment. In forming satisfaction judgments, price seems to play an important role, as it appears to be the first thing customers think of when they are evaluating products or transactions (Zeithaml,1988, in Martín-Consuegra et al., 2007). Research by Dai (2010) confirms this finding and shows that
the concepts perceived fairness and satisfaction are highly correlated, but yet still are different. Dai (2010) found that perceived price fairness influences customers’ satisfaction, because customers who perceive price as unfair are more likely to be dissatisfied with their purchase and customers who perceive price as fair are more likely to be satisfied with their purchase. Also, dissatisfied customers were more likely to take revenge at the company by for instance spreading negative word of mouth or avoid future purchases at the specific company (Dai, 2010).
Empirical findings suggest that customers use reference points in making purchase decisions. Previous experiences with a product or seller appeared to be a good predictor of future behavior. Therefore, satisfying past experiences can be the key to repeat purchasing behavior and trust in a firm (Dai, 2010). According to Dai (2010) customers who repeat their purchases and consider the firm as being trustworthy are seen as loyal to the firm. In this study we will focus on the relation between dynamic pricing, perceived price fairness and transaction satisfaction. Transaction satisfaction is expected to be indirectly influenced, through perceived price fairness, by dynamic pricing. Customers who experience dynamic pricing are more likely to perceived price unfair and in that way less satisfied with their transaction. We propose that:
H2: Perceived price fairness mediates the relation between dynamic pricing and transaction
satisfaction. Specifically, customers who experience dynamic pricing will perceive price as
more unfair and therefore are less satisfied with their transaction.
2.4 Customer loyalty
Although price seems to be an important marketing tool for marketers to become profitable, customer loyalty can also contribute to a firms’ sales and profitability. According to
Reichheld and Sasser (1990) attracting new customers costs, on average, more to a company than retaining ‘old’ customers. They claim that keeping ‘old’ customers is more economical,
while they repeat purchases and are less price sensitive. A strong customer base can give companies lots of advantages like; a stable income; positive word of mouth (free acquisition) about the firm or products and at last insight into the customer segments wants and needs (Anderson and Srinivansan, 2003; Jallat and Ancarani, 2008; Lee-Kelley and Gilbert, 2003; Yelkur and DaCosta, 2001). While the online shopping environment does not have the ability to deliver the same buying experience as the offline shopping environment, companies are struggling with how to build a strong relationship with their customers in order to keep them loyal. Dynamic pricing strategies can influence this customer relationship (Dai, 2010; Xia et al., 2004).
Previous literature shows that customer loyalty mostly has been searched and measured as a form of behavior (Day, 1969 in Dai, 2010). Reichheld and Sasser (1990) confirm this finding as they commit customer loyalty to repeat purchase behavior. However, behavior does not seem to cover the whole concept of customer loyalty. Behavior tends to imply that customers had the intention to behave in a certain way, while this does not always apply (Dai, 2010). For instance, when people are limited in their transportation they are limited in their choice of stores which can lead to repeated purchase behavior at a store, while the customer does not feel loyal to this company. Or customers who do feel loyal to a
company, while they never (are able to) buy at this store, for instance Chanel. These situations make it difficult to define what customer loyalty concerns (Kumar and Shaw, 2004, in Dai, 2010). For that reason this research will combine attitudinal and behavioral aspects of customer loyalty to distinguish customer loyalty from repeated purchase behavior.
Dai (2010) refers to customer loyalty as the most important indicator of the
relationship between buyer and seller. The relationship between buyer and seller is based on repeated buying behavior and customers’ beliefs about the trustworthiness of the firm (Xia et al., 2004). Trust is defined by Mayer, Davis and Schoorman (1995, p.712) as “the willingness
of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor, irrespective of the ability to monitor or control that other party”. Trust can be divided into three different dimensions: (1) ability (i.e., skills and competencies of the seller), (2) benevolence (i.e., the extent to which a seller is believed to want to do good to the buyer), and (3) integrity (i.e., the buyers’
perception that the seller is honest and fulfills its promises) (Mayer, Davis & Schoorman, 1995). These three dimensions can differ across the various stages in which the relationship between buyer and seller finds itself and influence customers’ perception of price fairness. For instance, in the early stage of the relationship between buyer and seller, one does not have much information about one another and trust will be based on the reputation of the company and the promised benefits to one another. In the latter stages of the relationship between buyer and seller, both parties will get more information about each other and their wants and needs. In the last stage customers feel like they can trust the company and that they have become loyal customers. In this phase, customers take the relationship between buyer and seller very personal, which means that, in case of dynamic pricing, perceived price discrepancies can be incorporated as volitional actions of the seller towards the buyer (Xia et al., 2004). Depending on the direction of the price discrepancy, this can either negatively or positively influence transaction satisfaction and trustworthiness of the firm (Constantinides, 2004; Jallat and Ancarani, 2008; Lee-Kelley and Gilbert, 2003; Park and Kim, 2003; Yelkur and DaCosta, 2001).
Research has found that when customers feel like they are treated unfavorably, they are more likely to search for explanations than when they are treated favorably. Bolton et al. (2003) found that customers evolve more negative emotions and behaviors towards the company, like switching to competitors, when they find out that the price discrepancy is due to volitional actions of the company. Therefore, companies using customers’ personal data to
charge customers different prices seem to take a huge risk in putting their relationship with their customers at stake. According to Gilliand and Bello (2002, in Dai, 2010) loyal customers are, in some cases, willing to sacrifice their own wants and needs to stay loyal to the firm. Martin et al. (2009) found that when price discrepancies were small, loyal customers were more likely to perceive this price as fair, than non-loyal customers. Which means that customers are, to some extent, willing to pay more for a product when they feel like this benefits their relationship with the seller. Also, customers tend to compare their prices paid to similar others while forming their perception of perceived price fairness. When a price
appears to be higher to a customer than to their reference point, the customer will feel betrayed for being loyal to the company (Sirdeshmukh, Singh and Sabol, 2002 in Xia et al., 2004). This perceived price unfairness could lead to less loyal customers and less new customers (Constantinides, 2004; Jallat and Ancarani, 2008; Lee-Kelley and Gilbert, 2003; Park and Kim, 2003; Yelkur and DaCosta, 2001).
To conclude, satisfaction along with perceived price fairness are key factors in order to acquire loyal customers. Empirical literature shows that satisfied customers are more likely to behave loyal to a firm (Gwinner et al., 1998, in Martín-Consuegra et al., 2007). However, when customers feel like the balance between what they need to sacrifice and what they earn of it is unequal, they will not be loyal and repurchase the product, even when they are
satisfied with the product (Bei and Chiao, 2001). This research will look into the relationship between perceived price fairness, transaction satisfaction and customer loyalty. Customer loyalty is expected to be indirectly influenced, through transaction satisfaction, by customers perceived price fairness. Customers who perceive price as being unfair are more likely to be dissatisfied with their purchase (transaction) and in that way less likely to repeat purchases and be loyal to the firm. We hypothesize that:
H3: Transaction satisfaction mediates the relation between perceived price fairness and
customer loyalty. Specifically, customers who perceive price as more fair are more satisfied
with their transaction and therefore will be more loyal to the firm.
2.5 Moderating individual level factors
In this research focus is on the individual level factors privacy concerns, price sensitivity and
product involvement that are expected to influence the relation between dynamic pricing and
perceived price fairness.
2.5.1 Moderating role of privacy concerns
Internet is an open medium which can be used to access information about customers and their online behavior easily (Brown and Muchira, 2004; Boyd and Crawford, 2012; Phelps, Nowak and Ferell, 2000). New technology allows businesses to use this customer data for several purposes such as being an information retrieval source, a sales tool, a distribution channel, and a customer support tool (Boyd & Crawford, 2012; Peterson, Balasubramanian and Bronnenberg, 1997 in Brown and Muchira, 2004; Phelps et al., 2000;). Internet differs from more traditional channels in three ways 1) increased data creation and collection, 2) globalization of information and communications, and 3) lack of centralized control
mechanisms (Berman and Mulligan 1999). These differences can be used in favor of firms, but can also have negative consequences for both firm and customers. For instance, users of the Internet leave all kinds of information open to other users (Boyd and Crawford, 2012; Phelps et al., 2000). Firms or other users can easily access this customer information and use it to customize their products, target their advertising, set prices or other ways of confronting their customers (Dai, 2010; Phelps et al., 2000). Also, firms can exchange customer
information with other firms and in that way private information is open to the world.
their privacy to some level. Customers are concerned about what information is used, saved or available to other users (Brown and Muchira, 2004; Phelps et al., 2000).
Previous literature by Milne and Gordon (2003) states that customers are willing to share personal information with companies in exchange for an economic or social benefit, but without the risk of experiencing negative consequences. Culnan and Armstrong (1999, p. 106) claim that customers are willing to share personal information when “1) information is
collected in an existing relationship; 2) when they feel like they have control over future use of the information; 3) the information collected is relevant for the transaction; 4) they feel like the information will be used to draw relevant inferences about the customer”. Customers think they build up a close relationship with the firm by disclosing personal information to them (Culnan and Armstrong, 1999; Milne and Gordon, 2003). This relationship holds as long as customer benefits exceed the risks of sharing their personal information. As long as buyers are in a close relationship with sellers they believe that the company is trustworthy and their information is safe. Moreover, customers expect to be treated beneficial in return to their loyalty (Culnan and Armstrong, 1999).
However, when companies are practicing dynamic pricing strategies, they use customer data to set different prices to different people according to their previous online behavior. Most customers are aware that companies obtain information about their customers and use this to reach out to them (Brown and Muchira, 2004). But do customers know that companies use this information to imply their willingness to pay and charge variable prices to their customers for exactly the same product (Brown and Muchira, 2004; Dai, 2010; Xia et al., 2004) ? This research will investigate what influence privacy concerns may have on the dynamic pricing – perceived price fairness relationship. Expected is that prices are perceived as less fair when customers discover that companies have used customers’ private information to charge them a specific price. Therefore, the fourth hypothesis to be tested is;
H4: Privacy concerns moderate the relationship between dynamic pricing and perceived
price fairness. Specifically, customers with high levels of privacy concerns will perceive
dynamic pricing as more unfair than customers with low levels of privacy concerns.
2.5.2 Moderating role of price sensitivity
Price sensitivity refers to the “the extent to which individuals perceive and respond to changes or differences in prices for products or services” (Monroe, 1973, in Wakefield and Inman, 2003, p.201). Research has found that customer reactions to prices differ across product categories. For instance, product categories which go along with high risks (drugs) are perceived to be more price sensitive than product categories that are less risky (Tellis, 1988). Wakefield and Inman (2003) claim that, depending on product commitment, customers can be more or less willing to search for a lower price. Customers are supposed to be more willing to look for a good price for products that are valuable to them, such as cameras, than to less valuable products, such as toilet paper. Not only product, but also experienced time pressure on the purchase can influence price sensitivity of customers. Customers who directly need the product will be less price sensitive than customers that have more time to consider other, cheaper variants (Shankar, Rangaswamy and Pusateri, 1999; Wakefield and Inman, 2003).
According to Shankar et al. (1999), price sensitivity is based on the benefits and costs of customers’ information search behavior. Benefits of information search not only include economical benefits but also non-economical benefits. Costs of searching for information includes the costs of searching for pricing information, as well as costs for searching for non-price information. In other words, the higher the benefits of information search to customers, the lower the focus of customers on price, thus the lower the level of price sensitivity
(Ancarani, 2002; Shankar et al, 1999). Furthermore, the costs and benefits of customers’ information search are based on two factors; 1) medium-related factors and; 2) customer factors. Medium-related factors concern perceived content on a website, perceived
interactivity on a website, the perceived depth of information on the website, the ease to look for price information and the range of product and price options on the website (Lee-Kelley et al., 2003; Shankar et al., 1999). First of all, when the website is more price oriented,
customers are found to be more focused on price search and more price sensitive than when the website mostly exposes product features. Second, website interactivity refers to the extent to which the website uses customer input to present their output. More specifically, sellers can offer customized output to specific customers which make them feel more involved and enhance their ease of shopping. An interactive shopping experience makes customers less price sensitive, as the multiple layers of information on the website makes customers curious to search for non-price information (Shankar et al., 1999).
Customer factors influencing customers’ price sensitivity include, according to
Shankar et al. (1999), loyalty, value of time and shopping frequency. Customer loyalty has an important impact on price sensitivity. Martin et al. (2009) claim that loyal customers are less likely to perceive price discrepancies as unfair than non-loyal customers. Thus, loyal
customers tend to focus less on price than non-loyal customers. Furthermore, Shankar et al. (1999) found that customers who have little time to search for a product will be less likely to compare prices and therefore rely on other factors like past experiences in buying their
products. However, frequent online shoppers are more likely to look for price information and be more price sensitive (Kalyanaram and Winer, 1995, in Shankar et al., 1999).
In this research price sensitivity will be studied in case of dynamic pricing, where sellers use customer information to charge customers different prices (Dai, 2010; Xia et al., 2004). The experienced price discrepancies can have various consequences to customers depending on their level of price sensitivity. Expected is that price sensitivity influences the relationship between dynamic pricing and perceived price fairness. Customers who tend to be price sensitive are more likely to perceive prices as being unfair, are more likely to compare
prices and to search for other price options (Shankar et al., 1999; Wakefield and Inman, 2003). Therefore, is hypothesized:
H5: Price sensitivity moderates the relationship between dynamic pricing and perceived price
fairness. Specifically, customers that are more price sensitive will perceive dynamic pricing
as more unfair than customers that are less price sensitive.
2.5.3 Moderating role of product involvement
Customers’ product involvement is a widely researched concept (Dholakia, 2001; Warrington and Shim, 2000; Quester and Smart, 1998). Researchers have pointed out that product
involvement is an important antecedent of the customers’ decision making process in purchase situations. Consistent with previous literature, Dholakia (2001, p. 1341) defines product involvement as “an internal state variable that indicates the amount of arousal,
interest or drive evoked by product class” which refers to the fact that customers differ in their interest and involvement with specific products from various product categories. According to Andrews, Durvasula, and Akhter (1990, in Warrington and Shim, 2000. p 763) involvement refers to an “internal state of arousal comprised of three major properties: intensity, direction, and persistence”. Whereas they claim that ‘intensity’ refers to the degree of involvement, ranging from high involvement to low involvement depending on product, situation and individual. Basically, the level of product involvement varies between customers and situation, however some product categories can be distinguished which in general concern higher product involvement than others. ‘Direction’ is claimed to be the product towards which the customer is involved. At last ‘persistence’ refers to the duration of the involvement, which can vary from short-term involvement to long-term involvement (Warrington and Shim, 2000).
According to Laurent & Kapferer (1985) and Zaichkowsky (1986) the degree to which customers are involved with a specific product or product category can be explained by three factors, namely (1) individual characteristics such as a customers’ needs, interests, values, and goals; (2) situational factors like purchase situation; and (3) characteristics of the purchase object, such as variations in product categories. Previous research by Suh and Youjae (2006) claims that product involvement is mainly based on personal relevance of the product to the customer varying between either situational or enduring product involvement. Situational product involvement mostly concerns external stimuli, which only temporarily influence the relevance of the product towards the customer. Enduring product involvement is more concerned about intrinsic stimuli, like past experiences which are mostly stored in the long-term memory and in that way more stable (Suh and Youjae, 2006). Products concerning high involvement are more likely to concern more extensive search behavior, more time spent on decision making and more detailed product attributes which makes it hard to choose between products (Zaichkowsky, 1986). In other words, customers who conduct more ‘information seeking behavior’ and ‘product/price comparing behavior’ in order to look for the best option will be more involved with the product than customers who are shopping for groceries. Namely, high involved customers tend to collect more knowledge about a product and its substitutes before purchasing than customers looking for frequently bought products (Engel et al., 1993 in Schmidt and Spreng, 1996; Warrington and Shim, 2000).
This research will focus on the moderating effect of product involvement on the dynamic pricing - perceived price fairness relation. Expected is that customers’ degree of involvement with a product influences their perception of price fairness. Customers who are highly involved with a product are more likely to perceive price as unfair as they have much more knowledge about a product and its substitutes than low involved customers. Customers
who are low involved will be more likely to perceive price as fair as they are less aware of product information and price. Finally, we hypothesize that:
H6: Product involvement moderates the relationship between dynamic pricing and perceived
price fairness. Specifically, customers who are more involved with a product will perceive
dynamic pricing as more unfair than customers who are less involved with a product.
Figure 2
Conceptual model: the influence of price sensitivity, privacy concerns and product
involvement on the relation between dynamic pricing – perceived price fairness – transaction satisfaction and customer loyalty.
3 Method
In the following chapter the collection of the sample and the research design is discussed. First the data collection method and characteristics of the collected sample are described. Then, an operationalization of the variables and their measurements will be given. At last, a brief description of the statistical procedures to test the expected relations is presented. See the appendices for the pre–test questionnaire and the main questionnaire (Appendices A1 to A4).
3.1 Research design
This study used a 2 (no dynamic pricing versus dynamic pricing) x 2 (high involvement (electronics) versus low involvement (household goods)) experimental design, in which
respondents were randomly assigned to one of four different purchase scenarios. The independent variables were dynamic pricing, product involvement, price sensitivity and privacy concerns. Perceived price fairness, satisfaction and customer loyalty were dependent variables. First of all respondents were either placed in a situation where they bought a product and found out that a friend paid a lower price for the same product, or in a situation where they bought a product and did not find out what other customers paid in order to test dynamic pricing. Second, respondents in each pricing situation were randomly assigned to either buy an iPhone (electronics) or toothpaste (household goods) in order to test product involvement.
The questionnaire started with measuring the individual level factors; price sensitivity, privacy concerns and product involvement. Then the respondent was randomly assigned to one of the four purchase scenarios in which it was depicted to purchase a product at Bol.com. Bol.com was chosen as it is one of the largest Dutch e-tailers, which makes it more likely that respondents are familiar with the company and could contribute to a more realistic purchase scenario (Meijsen, 2014). Bol.com offers a wide variety of products like books, DVD, electronics and personal health care. For this research two products were used according to their level of general product involvement; electronics (iPhone), which is expected to be a high involvement product, and household goods (toothpaste), which is expected to be a low involvement product (Dholakia, 2001). The clarity of the expected product involvement levels were tested during the pre-test.
The four purchase scenarios all included information about the product and a visual image of the product. The image and information were similar for each product in both the ‘dynamic pricing group’ and ‘no dynamic pricing group’. The price difference between the respondent and the other customer in the ‘dynamic pricing group’ was set at 15% for each of the two products, which is similar to the research by Haws and Bearden (2006). Also, the
clarity of this perceived price difference will be measured during the pre-test. After receiving the information and image of the product, respondents were either told about a friend who bought the same product for a lower price or no additional information was given, depending on the group they are in (dynamic pricing versus no dynamic pricing). At last respondents in the ‘dynamic pricing group’ were made aware about the fact that the firm used personal customer data to set the specific price to the specific customer, in order to test the influence of privacy concerns (Appendix A5). After the purchase scenario, respondents were asked about their perceived price fairness, transaction satisfaction and customer loyalty. The questionnaire ended with some demographical questions and the option to fill in a personal email adress in order to win a giftcard of choice (value €30).
Pre-test
In order to make sure that all items of the questionnaire were clear and valid a pre-test was distributed to 16 respondents. All respondents were able to fill in the questionnaire
sufficiently. After a short introduction about online shopping, respondents were asked about their price sensitivity, privacy concerns and product involvement (Dai, 2010; Xia et al., 2004). Then respondents were randomly assigned to one of the four simulated purchase scenarios at Bol.com in which they either experienced dynamic pricing or did not experience dynamic pricing. Important was that the ‘dynamic pricing group’ found out about the fact that they were paying a higher price for a similar product than others and that this was due to the firm using private customer data to set these prices. The ‘no dynamic pricing group’ served as a control group and did not found out about the dynamic pricing (Dai, 2010; Xia et al., 2004). Also, the respondents were randomly assigned to either buy an iPhone (high involvement product) or toothpaste (low involvement product) (Dholakia, 2001). Respondents that were assigned to the dynamic pricing scenario were afterwards asked about the magnitude of the
price difference and if they perceived this as being a major of minor difference. These questions tested whether the price differences in the dynamic pricing scenarios were clear. Furthermore, all respondents were, after the confrontation, asked about their perceived price fairness, satisfaction with the transaction and customer loyalty (Dai, 2010; Xia et al., 2004). Also, some demographical information about the respondents was asked. At last the pre-test included three (extra) open-ended questions in order to test if the manipulation of the
purchase conditions was clear and valid and to test if the questionnaire was free of
ambiguities. The first question asked what respondents thought that would be a reasonable price for the product in their purchase scenario, in order to test if the prices in the purchase scenarios were realistic. The second question tested if the purchase scenario could be more realistic, in order to check if respondents perceived the purchase scenario as being comparable with a real life situation. The third question considered whether the questionnaire items and purchase scenario were clear to the respondent (Dai, 2010) (Appendix A1 and A2).
The results of the pre-test showed that all respondents perceived the price difference in the dynamic pricing condition as a major difference, therefore it was assumed that the price difference in the dynamic pricing condition is clear. The prices for both the iPhone and the toothpaste were overall perceived as being realistic. All respondents reported a realistic price as being somewhat around the stated prices of €599 (iPhone) and €10,99 (toothpaste 5-pack). Most of the respondents in the pre-test reported that the purchase scenario was realistic and recognizable. Although, one respondent suggested to include a ‘Buy Now’ button in the scenario to make it more realistic. In the main questionnaire this button is included. The results of the third open ended question included some suggestions of respondents whether the questionnaire was clear. The term ‘e-tailer’ appeared to be confusing for some respondents and therefore was changed in the main questionnaire into ‘online retailer’.
3.2 Sample
During a period of 21 days, respondents were approached to fill in the main questionnaire by sending electronic invitations to people using snowball sampling. Respondents who filled out the questionnaire could choose to fill out their email address and have a chance of winning a gift card of their choice (value €30). The gift card was randomly distributed to one of the respondents who left their email address after closing the survey and was aimed to enthuse people to fill out the questionnaire. People who received the link to the questionnaire could in turn send the link by email to their own network (Saunders and Lewis, 2009). This sampling method was chosen so, in contrast to Dai (2010), not only university students were included, but respondents aged between 18 and 65 were invited to participate. In that way participants covered a wider part of society and outcomes can be more generalized. By approaching a more diverse group of people, chances of including people with different attitudes towards the searched variables, price sensitivity, product involvement and privacy concerns, are more likely (Dai, 2010; Xia et al., 2004).
From the 254 respondents that started filling out the questionnaire, 170 respondents completely finished the questionnaire (response rate 66.5%). Respondents who did not completed the questionnaire were excluded from analysis. From the remaining 170
respondents 59% were female and 41% were male. The mean age was 33 years (SD age = 12.3), ranging from 18 to 64 years. A small majority was educated at research university level (40.5%), while the rest of the respondents were educated at an university of applied sciences (HBO = 39.3%. MBO = 11.9%) or at a secondary school (Vwo/Gymnasium = 3%, Havo = 2.4%, Mavo = 2.4%). 1 respondent had only completed primary school (0.6%). 26.3% of the respondents reported to have a modal income, while 38.3% reported to have an income beneath modal and 35.3% reported to have an income above modal. A total of 95.9% of the respondents reported that they had purchased products online before (Table 1).
Table 1
Demographic characteristics of main survey
Demographic characteristics Frequency Percentage
18 to 25 61 39.1 26 to 35 41 26.3 Age 36 to 45 23 14.7 46 to 55 19 12.2 56 to 65 12 7.7 Gender Man 68 41 Vrouw 98 59
Online shopping experience Yes 163 95.9
No 7 4.1 Education Basisschool 1 0.6 VMBO 4 2.4 HAVO 4 2.4 VWO/Gymnasium 5 3.0 MBO 20 11.9 HBO 66 39.3 WO 68 40.5
Income level Beneath modal 64 38.3
Modal 44 26.3
Above modal 59 35.4
Total N=170 100%
By filling out the online questionnaire respondents were at random evenly distributed over the four purchase conditions. Due to exclusion of unfinished questionnaires there are small deviations in sample sizes of the four conditions; dynamic pricing and mobile phone (N = 46 (27.1%)), dynamic pricing and toothpaste (N = 45 (26.5%)), no dynamic pricing and mobile phone (N = 40 (23.5%)), and no dynamic pricing and toothpaste (N = 39 (22.9%)). Condition 1 in which respondents were experiencing dynamic pricing while buying a mobile phone 38.5% was male and the mean age was 35 year (SD = 13.3). A majority (55%) was educated on an university of applied sciences (55%). Condition 2, where respondents were not
experiencing dynamic pricing while buying a mobile phone, 31.1% was male and the mean age was 32 year (SD = 12.4). Most of the respondents (45.7%) were educated on university level. In condition 3, where respondents were experiencing dynamic pricing while buying toothpaste, 37.5% was male and the mean age was 33 year (SD = 11.7). A majority (44.7%) of respondents was educated at university level. In condition 4, where respondents were not experiencing dynamic pricing while buying toothpaste, 52.2% was male and the mean age was 32 year (SD = 11.9). A majority of the respondents were educated at an university of applied sciences. In all four conditions the mean income of respondents was modal, only in condition 2 the respondents had a mean income slightly above modal.
An independent t-test showed that there was no significant difference in respondents perception of price sensitivity and privacy concerns between the four conditions. In the dynamic pricing conditions respondents did not report higher scores on price sensitivity for mobile phones (M = 5.35, SE = .20) than for toothpaste (M = 5.46, SE = .15) as the difference between means was not significant t(77) = -.44, p > .05. The difference between privacy concerns between mobile phones (M = 4.41, SE = .16) and toothpaste (M = 4.61, SE = .15) was also not significant t(77) = -.86, p > .05. In the no dynamic pricing conditions there also was no significant difference between price sensitivity for mobile phones (M = 4.96, SE = .20) and toothpaste (M = 5.39, SE = .18) t(89) = -1.57, p > .05. The difference between privacy concerns for mobile phones (M = 4.49, SE = .17) and toothpaste (M = 4.69, SE = .13) was not significant t(89) = -.90, p > .05. For product involvement there was a significant difference found between the conditions. For the dynamic pricing conditions respondents reported higher scores of involvement for mobile phones (M = 2,56, SE = .06) than for toothpaste (M = 1.48, SE = .06) and the difference was significant t(77) = 12.25, p < .05. Also, for the no dynamic pricing conditions this difference in product involvement is found. Respondents reported higher scores on product involvement for mobile phones (M = 2,58, SE
= .06) than for toothpaste (M = 1.47, SE = .05) and this difference is significant t(89) = 14.00, p < .05.
3.3 Measurement of variables
Questionnaire translation
As all participating respondents were native Dutch speakers and all items were derived from English studies, the questionnaire had to be translated. In order to make sure that the meaning of all items remained the same, a third person translated the Dutch questionnaire back to English again. Deviations from the original questionnaire were corrected in the final version of the questionnaire.
Perceived price fairness
To measure respondents perceived price fairness a six-item scale designed by Darke and Dahl (2003, in Dai, 2010) was used. This scale showed a Cronbach’s alpha of .90 in a similar kind of research which appears to be very reliable. An example of one of the items is “The price I paid was fair “. The six-item scale for perceive price fairness used a seven-item Likert scale ranging from (1) ‘Strongly agree’ to (7) “Strongly disagree” (Dai, 2010; Haws and Bearden, 2006) (see Appendice A1 to A4).
Transaction satisfaction
To measure respondents satisfaction with their transaction a five-item Likert scale was used by Oliver (1980). This scale by Oliver (1980) originally consisted of eight-items, but three items were eliminated as they concerned overall satisfaction. An example of one of the items is “I am satisfied with my purchase decision”, which should be answered using a seven-item Likert scale ranging from (1) ‘Strongly agree’ to (7) “Strongly disagree” (Dai, 2010)
Customer loyalty
Customer loyalty was measured using a 17-item Likert scale. 16- items were used from the originally 28 customer loyalty items by McMullan and Gilmore (2003, in Dai, 2010). Additionally one item was used from the Dai (2010) study. As mentioned in the previous chapter, this thesis explores both the attitudinal and behavioral aspects of customer loyalty. McMullan and Gilmore’s (2003, in Dai, 2010) both include these dimensions in their scale which explains why this scale was used. Dai (2010) developed a four-items scale additional to the 16-items, in order to include repeat purchase behavior and word of mouth. As this study only takes into account repeat purchase behavior, only this item from Dai’s (2010) four-item scale is included in the questionnaire. An example of one of the used items is “Bol.com is more than a mere retailer to me”. Also this scale is chosen as it showed Cronbach’s alpha values ranging from .70 and .81 (McMullan and Gilmore, 2003 in Dai, 2010) for each dimension of the construct. At last, Dai (2010) rescaled the scale to an 7-item Likert scale, ranging from (1) “Strongly agree” to (7) “Strongly disagree” (Appendix 1).
Privacy concerns
Respondents’ privacy concerns were measured using a scale created by Sheehan and Hoy (1999), which consists of a combination of different privacy concern measures from previous studies. The scale consists of fifteen items concerning ‘privacy concern situations’, which are answered using a seven point scale ranging from (1) “not at all concerned” to (7) “extremely concerned”. An example of one of the used items is “A notice on a web page states that information collected is used by other divisions of that company”. The fifteen items included five items indicating low concerns, five items indicating moderate concerns and five items indicating high concerns (Sheehan and Hoy, 1999) (Appendices A1 to A4).
Price sensitivity
Price sensitivity was measured using a three-item scale which was used by Wakefield and Inman (2003) and approved satisfactory by previous pricing researchers. All items scored a Cronbach’s Alpha ranging between .86 and .89. An example item is “I’m willing to make an extra effort to find a low price for (iPhone/toothpaste)“. Participants were asked to indicate their answers using a 7-point Likert scale ranging from (1) “Strongly agree” to (7) “Strongly disagree” (Wakefield and Inman, 2003) (see Appendices A1 to A4).
Product involvement
Respondents’ product involvement was measured using a 10-item scale by McQuarrie and Munson (2003), who added two more items to the originally eight-item scale designed by Zaichkowsky (1985, in McQuarrie & Munson, 2003) in order to improve reliability and validity. An example of one of the used items is “I have a most preferred brand of this product”, which should be aswered using a 3-point scale ranging from (1) “Agree” to (3) “Dissagree” (see Appendices A1 to A4).
3.4 Statistical procedures
The data were collected using an online survey, which was created using Qualtrics online survey software. The survey started at 15th of December 2014 and ended three weeks later on 6th of January 2015. For analyzing the data the Statistical software Package for Social
Sciences (SPSS) was used. Prior to analyzing the data, the reverse coded items were recoded and the data was checked for abnormalities using scale reliabilities, descriptive statistics, skewness, kurtosis and normality tests. The normality check showed that only privacy concerns and customer loyalty were normally distributed. For price sensitivity a majority of respondents reported high levels of price sensitivity. This can be explained by the fact that
opportunities to compare prices and find the best price deals. For product involvement a majority of people reported high levels of involvement for mobile phones, while for toothpaste respondents were more likely to report a low level of involvement. This can be explained by the fact that respondents might feel more involved to products that are a bigger investment like mobile phones. Also, respondents might feel more involved to products for intensive personal usage rather than products for ordinary usage. For perceived price fairness and transaction satisfaction respondents both slightly reported lower scores, which refers to lower perceived price fairness and transaction satisfaction.
After checking the normality of the variables, scale reliabilities were tested using Cronbach’s alpha. The results can be found in Table 2. Each multi-item construct was tested on internal consistency and to check whether it contributed to the overall scale reliabilities. A Cronbach’s alpha score of >.70 assumes a reliable scale (Field, 2005). Price sensitivity(α = .83), privacy concerns (α = .88), product involvement (α = .72), perceived price fairness (α = .92), transaction satisfaction (α = .90) and customer loyalty(α = .89), all scored above the threshold point of α = .70, which implies they are good reliable scales.
As all data was cleaned and checked, correlation analysis was conducted to get a first impression of the existing relations between the variables. Next, two MANOVA’s were conducted for dynamic pricing (no dynamic pricing versus dynamic pricing) and product (mobile phone versus toothpaste) to test if the four conditions differed on their levels of perceived price fairness, transaction satisfaction and customer loyalty.
Afterwards, regression analysis was used to test the proposed hypotheses. First of all direct effects were measured using hierarchical regression analysis for the dependent variables dynamic pricing, perceived price fairness, transaction satisfaction and customer loyalty. Hierarchical regression was chosen, opposed to putting all variables in the regression model at
once, as it can control for the influence on the dependent variables for several groups of variables. It enables exploration of the effect of adding a variable to the regression model (Field, 2005). In the first step the control variables gender, age, education and income level were entered into the regression model, to check their influence on the outcome variables. In step 2 dynamic pricing, perceived price fairness and transaction satisfaction were entered as independent variables, to test the expected effects controlled for the control variables. In step 3 the individual level factors price sensitivity, privacy concerns and product involvement were entered, to explore if the expected effects are strengthened or weakened by adding these variables. As the variables perceived price fairness and transaction satisfaction were not normally distributed and as the sample sizes of the four purchase conditions is limited, bootstrapping was used as recommended by Preacher and Hayes (2008). Bootstrapping implies a non-parametric resampling procedure to get a sample that is a better representation of the population and to reduce inaccurate outcomes. As this research includes four
experimental conditions, stratified bootstrapping was chosen to make sure that the samples were equally drawn from the observed data. In this case 2,000 random samples of size 170 with replacement were conducted from the observed data (Preacher and Hayes, 2008).
The mediation effects of perceived price fairness on the relation between dynamic pricing and transaction satisfaction and of transaction satisfaction on the relation between perceived price fairness and customer loyalty were measured using a SPSS macro for mediation by Preacher and Hayes (2008). Again bootstrapping of 2,000 samples of size 170 was used to reduce inaccurate outcomes, while the variables perceived price fairness and transaction satisfaction were not normally distributed.
The moderation effects of price sensitivity, privacy concerns and product involvement on the relation between dynamic pricing and perceived price fairness were measured using a
outcomes due to the not normally distributed variables price sensitivity, product involvement and perceived price fairness, bootstrapping was used. As recommended by Preaches and Hayes (2008) a resampling procedure of 2,000 times was used of size 170. Also, to avoid multicollinearity of the independent variables, variables were mean-centered (Preacher and Hayes, 2008).
At last the complete conceptual model was measured using another SPSS macro for moderated mediation by Preaches and Hayes (2008). Bootstrapping of 2,000 resamples of size 170 was used to reduce inaccurate outcomes due to not normally distributed variables.
4 Results
In this chapter, first the existing relations between variables are discussed using correlation analysis. Next, MANOVA was used to test for differences in variance across the four
purchase conditions. Then, direct relations between variables are discussed using hierarchical regression analysis. At last the regression analysis for mediation, moderation and moderated mediation are outlined.
4.1 Correlation analysis
Before testing the proposed hypotheses, a correlation matrix was computed to explore the existing relations between variables. A two-tailed correlation matrix was presented (Table 2) to observe the relations between variables. Missing values are excluded pairwise, in order to use as much information as possible in the analyses. Pairwise deletion only analyzes cases without missing data in each pair of variables being analyzed (Peugh and Enders, 2004). The measured descriptives and scale reliabilities, as described in the method section, are also presented in Table 2.
