Abstract
Customer satisfaction has been a focus of researchers and marketers as an important antecedent of customer loyalty. Some recent studies propose that customer delight possibly produces greater customer loyalty than satisfaction. Loyalty has also become of greater interest to researchers and marketers as a multiphase concept as well as a useful segmentation of customers with accompanying targeted strategies. As one of few empirical studies on customer satisfaction, delight, and loyalty, the primary objectives of this research is to understand how customer satisfaction and delight influence loyalty and to understand the multiphase framework of loyalty, including cognitive, affective, and conative loyalties. This study contributes to the body of knowledge on customer satisfaction, delight, and loyalty and provides theoretical and applied suggestions for the hospitality industry.
Customer satisfaction has been regarded as an important antecedent of loyalty for many years. The concept of customer delight is gaining attention from practitioners and researchers as a new strategy, as opposed to mere customer satisfaction (Blackwell, Miniard, & Engel, 2006; Chandler, 1989; Keiningham & Vavra, 2001; Schlossberg, 1990). In the hospitality industry, understanding customers’ emotional factors, such as delight, becomes more critical in relation to their consumption experiences and repeat purchase because customers have many opportunities to interact with a product or service provider during the consumption experience (Bigné, Andreu, & Gnoth, 2003, Mattila, & Wirtz, 2000) and customers are more likely to rely on explicit cues to assess a provider’s products and services (Bitner, 1992; McNeilly & Barr, 2006).
The literature has increasingly featured theoretical and empirical scholarship on satisfaction and delight (e.g., Crotts, Pan, & Raschid, 2008; Finn, 2005; Hicks, Page, Behe, Dennis, & Fernandez, 2005; Loureiro, 2010; Oliver, Rust, & Varki., 1997; Schümmer, 2007). The few empirical studies that support the conceptual distinction between satisfaction and delight have reported inconsistent evidence, particularly on their relationship to repurchase behaviors (Finn, 2005; Hicks et al., 2005; Loureiro, 2010; Ngobo, 1999; Oliver et al., 1997). Hicks et al. (2005) showed that delight has a significant impact on repurchase intention but satisfaction did not. Finn (2005), Loureiro (2010), and Ngobo (1999) found that satisfaction had a stronger impact on repurchase intention than delight. Oliver et al. (1997) found that delight has a significant impact on repurchase intention for symphony ticket purchasers, but an insignificant impact for theme park patrons. Additional research on the common and distinct characteristics of satisfaction and delight is ripe for examination in a hospitality and tourism context.
Just as satisfaction, customer loyalty is considered crucial to the success of business organizations. Academic research on customer loyalty has received considerable attention with many studies exploring the linkage between customer satisfaction and loyalty. In the hospitality industry, there is a strong need to assess the level of attitudinal customer loyalty, and some loyalty studies have distinguished between the attitudinal dimensions of loyalty and a multiphase framework of loyalty as a useful way to segment customers with differential strategies (Back, 2001; Back & Parks, 2003; Li & Petrick, 2008; McMullan & Gilmore, 2003; Oliver, 1997). The temporal sequence of a multiphase loyalty framework remains unproven (Li & Petrick, 2008).
This study is aimed at understanding (a) the relationships between customer satisfaction and delight influencing loyalty and (b) the existence of cognitive, affective, and conative loyalty dimensions in a model with satisfaction and delight constructs. Our study addresses a gap in the research by providing an empirical testing of relationships. The findings are discussed in a way that marketers can better serve hospitality customers more effectively by better understanding loyalty creation and development.
Literature Review
The literature is presented under the following topics: customer satisfaction, customer delight, and customer loyalty, followed by hypotheses and the conceptual model.
Customer Satisfaction
As a core concept of marketing, the literature on customer satisfaction supports that satisfaction is an essential factor related to a company’s future profit by increasing customer loyalty (E. W. Anderson, Fornell, & Mazvancheryls, 2004; Homburg, Koschate, & Hoyer, 2005). According to the expectation-disconfirmation theory (Oliver, 1981), customer satisfaction is believed to result from a process of a customer comparing his/her expectations and perceptions of performance; the confirmation or disconfirmation of those expectations then predicts satisfaction. This theory is the most extensively accepted theory and has been applied by many researchers and marketers (Mattila & Wirtz, 2000; Wirtz, Mattila, & Tan, 2000; Yi, 1990). More recent satisfaction definitions concede an emotional response in modeling satisfaction evaluations (Giese & Cote, 2000; Oliver, 1997; Spreng, MacKenzie, & Olshavsky, 1996), and satisfaction is defined as consumers’ evaluative judgments based on both cognitive and affective responses as an emotional response (e.g., Bigné et al., 2003; Oliver, 1997; Phillips & Baumgartner, 2002; Wirtz & Bateson, 1999; Wirtz et al., 2000).
In the hospitality industry, a number of researchers have applied satisfaction-related theories and methods as one of the most frequently examined topics, and many empirical studies show that customer satisfaction is a strong indicator of intentions to revisit and recommend the destination to other people (e.g., Back & Parks, 2003; Jeong, Oh, & Gregoire, 2003; Mattila & Mount, 2003). Other studies have questioned the robustness of the customer satisfaction and loyalty relationship and have suggested other service dimensions may play a role in loyalty formation and sustainability (e.g., Skogland & Siguaw, 2004; Szymanski & Henard, 2001; Yi & La, 2004).
Customer Delight
The literature related to delight is presented under the following topics: definition of delight and the measurement of delight.
Definition of delight
Recently, the concept of delight has gained growing attention among researchers as well as practitioners. There is limited research about delight; the current literature still has inconsistencies in defining customer delight while the concept of satisfaction is relatively well established (Schümmer, 2007). There are two different approaches to defining delight as shown in Table 1. The first view of delight, founded on psychology literature (Plutchik, 1980), is distinct and separate from satisfaction, as delight is an emotion while satisfaction is a combination of emotion and cognition. As an emotion, delight is a combination of high pleasure (joy, elation) and high arousal (e.g., Berman, 2005; Finn, 2005; Hicks et al., 2005; Plutchik, 1980; Torres & Kline, 2006). Emotions, in general, consist of two dimensions: pleasure and arousal (Mano & Oliver, 1993; Mattila & Wirtz, 2000; Wirtz & Bateson, 1999). Pleasure refers to the degree to which a person feels good, joyful, or happy in a situation, whereas arousal refers to the extent to which a person feels stimulated and active (Bigné et al., 2003). Based on the expectancy-disconfirmation theory (Oliver, 1981), delight occurs when a customer is pleasantly surprised in response to an experienced disconfirmation toward a company or its product /service experiences, while, in satisfaction, the customer receives what is expected (Berman, 2005; Crotts et al., 2008; Finn, 2005; McNeilly & Barr, 2006; Oliver et al., 1997; Torres & Kline, 2006). Some researchers argue that surprise is not required to experience delight and it can occur simply as a result of joy (Arnold et al., 2005; Barnes, Beauchamp, & Webster, 2010; Kumar, Olshavsky, & King, 2001; St. James & Taylor, 2004).
Summary of Delight Literature
The second view of delight is derived from its definition as an extreme level of satisfaction (or complete satisfaction; Berman, 2005; Keiningham, Goddard, Vavra, & Iaci, 1999; Kumar & Iyer, 2001). Most of the research based on this view does not clearly distinguish between delight and satisfaction. Rather, delight is simply assumed to be a higher level of satisfaction (Alexander, 2010; Vanhamme, 2008).
Based on the literature review, this study supports the conceptualization of customer delight as an emotion or affective response. Affective responses have been found as antecedents of customer satisfaction and delight (e.g., Arnold et al., 2005; Finn, 2005; Mano & Oliver, 1993; Oliver et al., 1997; Phillips & Baumgartner, 2002; Schümmer, 2007).
Another antecedent of satisfaction is disconfirmation or the process where consumers compare product performance with their expectations, as defined by expectancy disconfirmation theory (Oliver, 1981). A state of positive disconfirmation is experienced when performance exceeds one’s expectations for the product or service and a negative disconfirmation is when the performance falls below expectation.
When a consumer does not have expectations but is pleased with performance, that affective state is called “delight.” Delight has also been documented when the level of performance is surprisingly positive and exceeds the normal boundaries of expectation disconfirmation cognitive processing (Oliver, 1997). Thus, delight encompasses a surprise or no expectation component or a higher level of performance evaluation than an individual’s normative range of expectation fulfillment; satisfaction encompasses normative expectation fulfillment, which is implicitly not a surprise (Schümmer, 2007; Verma, 2003).
Satisfaction is composed of cognitive and affective elicitation, whereas delight is composed of primarily affective elicitation (Oliver et al., 1997; Schümmer, 2007) with emotions such as pleasure (Russell, 1980; Watson & Tellegen, 1985), joy, and surprise (Oliver et al., 1997; Plutchik, 1980).
Measurement of delight
Researchers define the concept of delight in different ways and employ different scales to measure the concept (Schümmer, 2007). According to the first view, delight has been measured using scales of emotions as a different construct from satisfaction (Finn, 2005; Hicks et al., 2005; Kumar et al., 2001; Loureiro, 2010; Oliver et al., 1997). As shown in Table 1, Oliver et al. (1997) and Kumar et al. (2001) used a single item, “feel delighted,” but that one item is criticized for not being sufficient to measure delight as a construct. Finn (2005), following Oliver et al. (1997), improved the delight measurement by using three items: “delighted,” “gleeful,” and “elated.” Loureiro (2010) used two items, “enchantment” and “delighted.” Crotts et al. (2008) measured delight indirectly and viewed delight as a mixture of surprise and happiness. In a qualitative study by Verma (2003), delight is described as “pleasurable,” “unforgettable,” and “surprisingly memorable.”
Those who see delight as an extreme form of satisfaction have not measured delight directly, but plainly measured delight as the highest rating of satisfaction (Estelami, 2000; Ngobo, 1999). For instance, Estelami (2000) divided the respondents into disappointment/delight as the most extreme score of satisfaction and Ngobo (1999) measured delight as 100% satisfaction. As described, there is no consensus on the measurement of delight, and therefore researchers agree that better measurement is needed (Berman, 2005; Kumar et al., 2001; Kwong & Yau, 2002; Vanhamme, 2008).
Customer Loyalty
Customer loyalty is defined as a deeply held commitment to rebuy or repatronize a preferred product or service consistently in the future, despite situational influences and marketing efforts having the potential to cause switching behavior (Oliver, 1997). Loyalty is desirable since retaining current customers is less expensive and easier than finding and developing new customers (Reichheld & Sasser, 1990). There are two dimensions of loyalty approaches: behavioral loyalty and attitudinal loyalty (Dekimpe, Steenkamp, Mellens, & Vanden, 1997; Dick & Basu, 1994; Jacoby & Chestnut, 1978; Yi & La, 2004). Behavioral loyalty is defined as the consumer’s tendency to repurchase, which is revealed through behavior that is measured and directly affects brand sales (Hammond, East, & Ehrenberg, 1996). Attitudinal loyalty is the customer’s predisposition toward a brand, which is a function of psychological processes (Jacoby & Chestnut, 1978). Much of the previous loyalty research has focused on the attitudinal dimension of loyalty (Li & Petrick, 2008; Morais, Dorsch, & Backman, 2004; Yi & La, 2004) because of the measurement limitations of behavioral loyalty, which simply measured behavioral variables to predict the customer’s purchasing behavior in the future (Back & Parks, 2003). Attitudinal loyalty is measured by the degree of customers’ intentions to revisit the destination and in their recommendations to others (Li & Petrick, 2008; Oppermann, 2000; Yi & La, 2004; Yoon & Uysal, 2005) and attitudinal dimension also has been emphasized in the loyalty research (Li & Petrick, 2008; Yi & La, 2004).
Oliver (1997) suggests four phases of the loyalty development process based on these two dimensions: (a) cognitive loyalty, (b) affective loyalty, (c) conative loyalty, and (d) action loyalty. The first phase, cognitive loyalty, is based on merely functional characteristics, such as costs and benefits and is focused on product performance. Cognitive is the weakest state of loyalty since this type of commitment is to the costs and benefits of a product and not to the brand itself (Oliver, 1997). Consumers are likely to switch when they perceive alternative offerings as being superior with respect to the cost–benefit ratio (Kalyanaram & Little, 1994; Sivakumar & Raj, 1997). The second phase is affective loyalty, which is a favorable attitude toward a specific brand or product. Affective loyalty is an enhanced liking for competitive brands, which is conveyed through imagery and association used in competitive communications (Oliver, 1999). Affective loyalty involves both the liking of the product and experiencing satisfaction with the brand. The third phase of loyalty development is conative loyalty, which is behavioral intention to repurchase and involves a deep brand-specific commitment (Harris & Goode, 2004; Oliver, 1999). This commitment is more like motives (Oliver, 1999). Conative loyalty is stronger than cognitive and affective loyalty, but has vulnerabilities (Blut, Evanschitzky, Vogel, & Ahlert, 2007; Evanschitzky & Wunderlich, 2006). In particular, repeated delivery failures are a main factor in weakening conative loyalty, and consumers are more likely to try alternative offerings if they experience frequent service failures (Blut et al., 2007; Evanschitzky, Vogel, & Ahlert, 2007; Oliver, 1999). Even though the consumer is conatively loyal, he or she has not developed the resolve to avoid considering alternative offerings (Oliver 1999). The last phase of loyalty is action loyalty, which is when attitude is transformed into action (Kuhl & Beckmann 1985; Oliver, 1999). The three previous loyalty states may result in a readiness to act (to buy), and this readiness is accompanied by the consumer’s willingness to search for the favorite offering despite considerable effort necessary to do so (Blut et al., 2007; Oliver, 1999).
Based on Oliver’s (1997) loyalty development process, some studies have been broadening the concept of the loyalty development process and challenged Oliver’s loyalty development process, the cognitive–affective–conative loyalty sequence, by proposing a different loyalty development process (Back, 2001; T. Jones & Taylor, 2007; Li & Petrick, 2008; Rundle-Thiele, 2005). No consensus on a loyalty development process exists, although recent studies have been broadening the concept. Back (2001) and Li and Petrick (2008) suggested that cognitive, affective, and conative loyalty are not sequentially linked but are independent factors of attitudinal loyalty, which lead to action loyalty based on the tripartite model of attitude structure (Breckler, 1984). According to this view, three components of people’s attitudes—cognition, affect, and behavioral intention—are independent and have some positive correlation with one another, and not all three components of attitude are built up through cognitive processes (Breckler, 1984). T. Jones and Taylor (2007) suggested a two-dimensional loyalty construct by combining attitudinal and cognitive loyalty as one dimension and behavioral loyalty as the other dimension. J. Lee (2003) partially supported Oliver’s loyalty development process as arguing that the cognitive stage is an antecedent of loyalty rather than loyalty and measuring attitudinal, conative, and behavioral loyalty.
There has been a long-standing theoretical issue in understanding the cognitive and affective sequence, and the interplay between cognition and affection is still unresolved (Chebat & Michon, 2003; Lazarus, 1991; Shiv & Fedorikhin, 1999; Solvic, 2000; Zajonc, 1980; Zajonc & Markus, 1985). According to Ajzen and Fishbein (1975), affect is derived from attribute beliefs that are evaluated in an expectancy-value manner (Edwards, 1990). Zajonc (1980) suggests that affect is precognitive in nature, occurring without any extensive perceptual and cognitive processes, and affect and cognition are separate and influence each other. Several researchers (Berkowitz, 1993; Epstein, 1993; Hoch & Loewenstein, 1991; LeDoux, 1996) agree with Zajonc’s (1980) findings. For example, Berkowitz (1993) suggests that affective reactions can occur relatively automatically without an active role of cognitive processes. Epstein’s (1993) cognitive-experiential self-theory proposes two conceptual systems: an experiential system and a rational system. An experiential system is affective in nature and is associated with crude and rapid processing, and a rational system is cognitive in nature and is associated with a more refined and deliberative processing. The two conceptual systems tend to operate in parallel (Shiv & Fedorikhin, 1999).
We adopt part of Oliver’s (1997) loyalty conceptualization but we posit that cognitive and affective loyalties are not sequentially related. We suggest that cognitive and affective loyalty influence each other and both cognitive and affective loyalty directly affect conative loyalty based on integrating psychological (Berkowitz, 1993; Epstein, 1993; Leventhal, 1984; Zajonc, 1980) and biological (LeDoux, 1996) theories of affect and cognition.
Hypotheses and Conceptual Model
Researchers have supported that satisfaction is an important antecedent of customer loyalty (E. W. Anderson et al., 2004; Homburg et al., 2005). If customers are satisfied with the product and service, they are more likely to continue to repurchase and are more willing to recommend to others (Cronin & Taylor, 1992; Yi, 1990). Oliver (1997) stated that customer satisfaction had a positive effect on attitudes through cognition. The positive attitudes were found to change attitudes toward the product or brand with an enhanced level of positive belief, a reinforced level of positive affect, and an increased repurchase intention (Albarracin & Wyer, 2000; Back, 2005; Oliver, 1997; Yi, 1990). Specifically, if customers perceive service as much higher than the expectation level and the offerings of other brands, they tend to show positive attitudes toward the focal brand. If a customer bases a preference level for the focal brand on a recent positive experience with the brand, he or she would upgrade the level of positive attitude toward the brand (Mittal, Ross, & Baldasare, 1998). People assess the attributes of objects using their current attitude. Once the assessment of the attitude object is assessed as positive, and also as consistent with existing beliefs toward the object, evaluation-cognitive consistency confirms the beliefs and enhances the level of belief confidence (Back, 2005; Smith & Swinyard, 1988). Thus, a customer initially becomes cognitively loyal based on beliefs about the brand attribute only.
Oliver’s (1997, 1999) research tested sequential attitudinal loyalty development and found customers’ affective brand loyalty was not affected directly by their satisfaction level. Rather, customers became affective brand loyal after they were cognitive brand loyal (Back, 2005; Back & Parks, 2003; Oliver, 1997). Attitudinal loyalty reflects the customer’’s psychological disposition toward the same brand or brand set. As such, attitudinal loyalty reflects favorable attitudes toward the brand or organization (Dick & Basu, 1994). Attitudinal loyalty toward a brand or firm is enhanced when the relative strength of the attitude toward the brand or firm is stronger as compared with other brands or firms. Building attitudinal loyalty involves more than simple transactional marketing incentives; positive attitudes toward one’’s brands or firm must be cultivated over a longer term relationship horizon (Kumar & Shah, 2004). As described earlier, this study proposes that cognitive and affective loyalties are not sequentially related, but cognitive and affective loyalty distinctly influence each other and both cognitive and affective loyalty are directly related to conative loyalty. Attitudes are formed not only through cognition, but also affect. In some cases, the cognitive component may be dominant, in some the cognitive and affective factors may interact with each other, and in other cases the affective factors may be dominant and primary (Zajonc & Markus, 1982). Therefore, we propose the following hypotheses:
Hypothesis 1a: Satisfaction will be significantly and positively related to cognitive loyalty.
Hypothesis 1b: Satisfaction will be significantly and positively related to affective loyalty.
Hypothesis 4: Cognitive loyalty will be significantly and positively related to conative loyalty.
Hypotheses 5: Affective loyalty will be significantly and positively related to conative loyalty.
Although academics have emphasized the importance of delighting customers because delight is likely to influence loyalty, loyalty-driven profit, and create positive word-of-mouth (e.g., Hicks et al., 2005; Oliver et al., 1997; Torres & Kline, 2006), the specific relationship between delight and cognitive/affective loyalty has not yet been examined. From the literature review of emotional processing mechanisms, emotions can profoundly influence cognitive processes by which individuals may use their apparent affective response to a target as a basis of judgment (Schwarz, 2000; Schwarz & Clore, 1996), and emotions are significant predictor of intentions, repeat purchasing/patronage behaviors, and development of brand loyalty (Y. Lee, Back, & Kim, 2009; Mattila & Enz, 2002; Morris, Woo, Geason, & Kim, 2002). Cohen and Areni (1991) indicate that consumption emotions, such as delight have strong episodic memory and are highly accessible to cognitive operations. Emotional experiences can be easily retrieved and integrated into current evaluative judgments (Arora & Singer, 2006). Moreover, people in a happy mood are more likely to adopt a heuristic processing strategy, such as top-down processing, with high reliance on preexisting knowledge structures and usual routines, and tend to overestimate the likelihood of positive outcomes (Johnson & Tversky, 1983; Nygren, Isen, Taylor, & Dulin, 1996; Schwarz, 2000). Specifically, delight produces emotional bonds between customers and a company, product, or service by “wowing” them, which may provide additional psychological benefits to the customers (Berry, 1995; Hirschman & Holbrook, 1982; Oliver et al., 1997). Based on the review of the relevant literature on the relationship between delight and loyalty, we propose the following hypotheses:
Hypothesis 2a: Delight will be significantly and positively related to cognitive loyalty.
Hypothesis 2b: Delight will be significantly and positively related to affective loyalty.
Many studies proposed that delight is more likely to affect loyalty than satisfaction (e.g., Berman, 2005; Crotts et al., 2008; McNeilly & Barr, 2006; Oliver et al., 1997; Torres & Kline, 2006), yet the empirical findings so far are inconsistent across studies and product types (Hicks et al., 2005; Finn, 2005; Loureiro, 2010; Ngobo, 1999). Furthermore, there is a lack of research on the relationships between customer satisfaction/delight and cognitive/affective loyalty. According to its definition, satisfaction involves both cognitive and emotional components (Dubé & Menon, 2000; M. A. Jones & Suh, 2000; Oliver, 1997; Yu & Dean, 2001), whereas delight is defined as an emotion (Berman, 2005; Crotts et al., 2008; Finn, 2005; McNeilly & Barr, 2006; Oliver et al., 1997; Torres & Kline, 2006).
The cognitive component of customer satisfaction refers to a customer’s evaluation of the perceived performance of products and service in terms of its adequacy in comparison with some kind of expectation standards (Liljander & Strandvik, 1997; Oliver, 1981; Yu & Dean, 2001) and the emotional component is an affective response to one’s perception of the series of attributes that compose a product or service performance (Dubé & Menon, 2000). Previous satisfaction studies were focused mainly on the cognitive component and suggested that there is a positive relationship between satisfaction and cognitive loyalty (Back & Parks, 2003; Oliver, 1997). The cognitive loyalty is focused on product performance in terms of price, benefits, and costs (Oliver, 1997), which also are major factors in determining satisfaction (Heskett, Sasser, & Schlesinger, 1997; Yoon & Uysal, 2005). Delight generates emotional bonds and stronger memory with customers than satisfaction (Berman, 2005; Loureiro, 2010) and involves receiving the unexpected; with satisfaction, the customer receives the expected (Berman, 2005; Crotts et al., 2008; Finn, 2005; McNeilly & Barr, 2006; Oliver et al., 1997; Torres & Kline, 2006). Based on the review of the relevant literature, we propose the following hypotheses:
Hypothesis 3a: Satisfaction will be more significantly and positively related to cognitive loyalty than delight.
Hypothesis 3b: Delight will be more significantly and positively related to affective loyalty than satisfaction.
Based on the proposed hypotheses, a model is proposed in Figure 1.
Proposed Research Model
Method
Sampling
To test the five main hypotheses, a study of consumers who purchased and consumed a product was necessary. The researchers selected a hospitality product because satisfaction, delight, and loyalty are likely to occur during experience-based consumption. Specifically, a stay at a resort in a well-known destination was selected. The resort is an independent hotel, not marketed by a national brand, and owned and operated by three generations of a family. The resort has 360 themed guestrooms, 5 indoor pools and 3 whirlpools, a miniature golf course, a large family fun center with more than 100 video/arcade games, and several guest banquet and convention spaces.
The population of this study was overnight resort guests. Travel parties who stayed at the resort were of interest to test the proposed model and hypotheses. Business travelers and groups for events (e.g., weddings, reunions, meetings) were excluded because some of them might not have been directly involved in the decision making or in the making of the reservation. The total sample size was 3,709 (e-mailed surveys 3,459; mailed surveys 250) and comprised all guest stays for the months of July, August, and September 2009. The sample was subdivided into those with and without e-mails for appropriate and cost-effective survey mailing.
Data Collection
A pilot study (n = 32) was conducted using an electronic survey in June 2009 with faculty, administrators and graduate students at a Midwestern United States university to develop and test the questionnaire instrument. Based on the performance in the pilot study, the questionnaire was revised to facilitate easier reading and clearer wording.
The main data collection was achieved via a self-administered online survey using a web-based survey tool, Qualtrics, on a weekly basis. Each week, on completing a stay at this resort, the survey questionnaire was e-mailed to customers with an invitation letter from the general manager of the resort and a university professor. Respondents who completed the survey received discount coupons and were entered into a drawing for one grand prize (i.e., a stay at the resort). The first e-mail reminder was sent to those who had not yet responded 2 weeks after the first survey questionnaire, and a second e-mail reminded those who had not yet responded 1 week after the first reminder. A total of 3,459 surveys were sent via Qualtrics and the survey response rate was 50% (n = 1,743).
A self-administered U.S. Postal Service–mailed survey also was employed to test for biases that an online survey might cause and to ensure an appropriate sample size for statistical power (i.e., the ability to detect and reject a poor model) in the structural equation modeling analysis (Chin, 1998). Participants were randomly chosen based on a limited budget for mail survey costs. After completing a stay at the resort, a survey package was mailed to them, followed by a reminder postcard 2 weeks after the initial mailing. A second mailing of the survey package was sent to those who had yet to respond. The survey package included an invitation letter by the general manager of the resort and a university professor, a survey questionnaire, and a prepaid return envelope. A total of 250 surveys were mailed and the mail survey response rate was 38% (n = 94).
Before combining the two sets of data, demographic characteristics of participants in these two data sets were compared. There were no significant differences between the two sets of data with demographic characteristics, such as gender, ethnicity, income, and number of people in the household. The online survey had more respondents with full-time employment and the paper survey had more Michigan resident respondents.
After the two sets of data were combined, 177 responses (170 from the online survey and 7 from the paper survey) were dropped because they were found to be inappropriate for the analysis (e.g., incomplete data, guests whose primary purpose of the stay was business, or a group/bus tour). A total of 1,660 completed and usable surveys resulted from the original 3,709 surveys (45% composite response rate; number of online surveys = 1,573; number of paper surveys = 87).
Measurement
The survey items for each construct were developed on the basis of previous studies. Customer satisfaction was measured with four items (Finn, 2005; Spreng et al., 1996). Responses were given on a 7-point Likert-type scale from 1 = strongly disagree to 7 = strongly agree with the following questions: “Overall, this hotel was comfortable,” “Overall, this hotel was satisfying,” “Overall, this hotel was pleasing,” and “Overall, this hotel was made me content.”
To measure customer delight, three items (Finn, 2005), such as “I felt delighted at some time during my stay at this hotel,” “I felt gleeful at some time during my stay at this hotel,” and “I felt elated at some time during my stay at this hotel” were employed. Three additional items also were developed based on previous qualitative studies, which described the delightful experiences with the products or services (Arnold et al., 2005; St. James & Taylor, 2004; Verma, 2003) and these items were consistent with the literature about the conceptual definition of delight (e.g., Finn, 2005; Oliver et al., 1997; Plutchik, 1980). Items included “I felt positively surprised at some time during my stay at this hotel,” “I felt overjoyed at some time during my stay at this hotel,” and “The hotel had experiences/services which were unexpected and they delighted me.” Responses to the delight items was given on a 7-point Likert-type scale from 1 = strongly disagree to 7 = strongly agree.
The measurement items for cognitive loyalty were developed by Li and Petrick (2008) and McMullan and Gilmore (2003) and were presented to participants as five items on a 7-point Likert-type scale from 1 = strongly disagree to 7 = strongly agree. Items included “I believe this hotel provides more benefits than other hotels of its type,” “I made the right choice of hotel with this hotel,” “The hotel’s facilities are visually more appealing compared to other hotels of its type,” “This hotel has better value for my money compared to other hotels’ prices of its type,” and “This hotel provides me superior service quality compared to other hotels of its type.”
The measurement items for affective loyalty were developed from Li and Petrick (2008) and McMullan and Gilmore (2003) and tested on a 7-point Likert-type scale from 1 = strongly disagree to 7 = strongly agree. The four items included “I feel happy when I stay at this hotel,” “I like this hotel more than other hotels of its type,” “I love staying at this hotel,” and “Staying in this hotel says a lot about who I am.”
Conative loyalty is defined as a customer’s behavioral intention to repurchase and involves a deep brand-specific commitment (Harris & Goode, 2004; Oliver, 1999). Four indicators were used to measure conative loyalty from the previous research (Li & Petrick, 2008; McMullan & Gilmore, 2003). Two items were on a 7-point Likert-type scale from 1 = strongly disagree to 7 = strongly agree with the following questions: “I consider myself to be highly loyal to this hotel” and “I intend to continue staying at this hotel.” Two other items included “I will return to this hotel in the next year” and “I will recommend this hotel to others.” These items were measured on a 7-point Likert-type scale from 1 = extremely unlikely to 7 = quite likely.
Data Analysis
Survey data were analyzed in two steps. First, preliminary statistics were obtained using the Statistical Package for the Social Sciences (SPSS) 17.0. Second, using M-Plus 5.2, structural equation modeling was used to test the measurement model by validating scales, assessing convergent and discriminant validity, and assessing the structural model by evaluating the hypothesized relationships.
Results
Profile of Survey Respondents
The majority of participants were female (69%) and the average age was 48 years. Majority of participants (82%) were European American/Middle Eastern/White. About 60% of respondents had an annual household income between $50,000 and $199,999. The average number of people in a household was three, and most of the respondents (58%) were employed full-time. The majority of respondents (69%) resided in Michigan, followed by those living in Ohio (12%), Canada (7%), and other states (12%).
Traveling Characteristics of Survey Respondents
Respondents reported the information sources or marketing communications that influenced them to book the current stay at the resort, with multiple responses allowed. The top source was a previous stay (57%). Other sources were hotel loyalty club membership holder (20%), friend or family member recommendation (17%), special package rate (17%), Internet website or search engine (16%), and e-mail promotion or newsletter offer (16%). Most of the respondents traveled with their family members (86%) or friends (12%).
Respondents were mostly satisfied with the hotel’s location (Mean 6.66 on 1: very unsatisfied to 7: very satisfied), friendliness of staff (6.43), hotel amenities (pool, game room, free nightly entertainment; 6.42), room cleanliness (6.42), quality of food (6.30), room comfort (6.12), relationship established between staff and customers (6.04), and value for money (5.68).
Testing of the Measurement Model
Before conducting a confirmatory factor analysis, a factor analysis using SPSS program was performed to check whether all six items of the delight construct, including three newly created items, were representing the construct well. All factor loadings for the items were relatively high, ranging from .79 to .92. Next, confirmatory factor analysis was conducted to confirm the relationships between observed variables and their underlying constructs to assess the degree to which the data fit the proposed measurement model. The proposed measurement model was tested first and did not produce a good fit with the data: χ2(220, N = 1,660) = 2780.0, p < .001, comparative fit index (CFI) = .931, Tucker–Lewis index (TLI) = .921, root mean square error of approximation (RMSEA) = .84, standardized root mean square residual (SRMR) = .046. Some items were dropped when observed variables had low factor loadings (less than .5) and large residuals based on Kline’s (1998; 2010) recommendations. During this process, the following four items were dropped, “The hotel had experiences/services which were unexpected and they delighted me” (one of the new delight items), “I made the right choice of hotel with this hotel” (cognitive loyalty item), “I love staying at this hotel” (affective loyalty item), and “I consider myself to be highly loyal to this hotel” (conative loyalty item). After eliminating the observed variables, the goodness-of-fit indices of CFI, TLI, RMSEA, and SRMR were within the recommended range for a model with good fit to the data, χ2(142, N = 1,660) = 1129.81, p < .001, CFI = .966, TLI = .960, RMSEA = .065, SRMR = .038 (Hu & Bentler, 1999), whereas χ2 indicated a poor fit because of the large sample size. With increasing sample size, the χ2 value increases and leads to plausible models being rejected (Bollen, 1989; Cheung & Rensvold, 2002; Jöreskog & Sörbom, 1993; Schermelleh-Engel, Moosbrugger, & Müller, 2003). Chi-square/degrees of freedom (χ2/df) ratio values that are lower than the value of 2 represent a minimally plausible model (Byrne, 1991), but the measure of χ2/df of freedom is also as dependent on sample size as the χ2 test. Wheaton (1987) argued that this direct dependence on sample size makes it impossible to suggest general thresholds as a guideline. Working with a large sample size, we concluded that the model was acceptable based on the recommended parameters in the literature.
As shown in Table 2, the measurement model examined reliability and construct validity, including convergent and discriminant validity. The reliability test was conducted using the Cronbach’s alpha and a composite reliability, which indicates the internal consistency of the observed variables measuring each factor. Cronbach’s alpha of all four factors exceeded the recommended .70 (Nunnally, 1978). Composite reliability was estimated as a second measure of reliability because Cronbach’s alpha may over- or underestimate scale reliability (Raykov, 1998). The acceptable range for composite reliability should be greater than .70 (Chin, 1998), and all five factors were found to be acceptable. To assess construct validity, convergent and discriminant validities were investigated. Convergent validity was supported with factor loadings for the observed variables that were statistically significant (p < .001); and they ranged from .67 to .97 for standardized factors. All the average variance extracted exceeded .50. Discriminant validity was also confirmed by noting that the average variance extracted for each construct was greater than their shared variance (Fornell & Larcker, 1981).
Results of Reliability and Validity
Note: SAT = Customer satisfaction; DEL = Customer delight; COG = Cognitive loyalty; AFF = Affective loyalty; CON = Conative loyalty.
Composite reliability = (Sum of standardized loadings)2/((Sum of standardized loadings) 2 + Sum of indicator measurement error).
Average variance extracted (AVE) = Sum of squared standardized loadings/ (Sum of squared standardized loadings + Sum of indicator measurement error).
Along the diagonal is the average variance extracted (AVE).
Off the diagonal is the shared variance (square of the correlations between any two constructs).
Testing of the Structural Model
The five main hypotheses testing the relationships among the factors are reported as structural equation model results in Table 3. The model fit indices indicated an adequate fit to the data whereas χ2 indicated a poor fit because of the large sample size, χ2 (144, N = 1,660) = 1179.63, p < .001, CFI = .965, TLI = .958, RMSEA = .066, SRMR = .038.
Results of Hypotheses Testing
Four of the five hypotheses were supported. Customer satisfaction had significant influences on cognitive and affective loyalty (support for Hypotheses 1a and 1b). Specifically, these findings show that satisfaction positively influences cognitive (β = .59) and affective loyalty (β = .53). Customer delight had a significant influence on cognitive (β = .30) and affective loyalty (β = .41; support for Hypotheses 2a and 2b). Hypothesis 3a was supported. Customer satisfaction had a greater influence on cognitive loyalty (β = .59) than the influence of delight on cognitive loyalty (β = .30). Hypothesis 3b was not supported because customer delight did not have a greater influence on affective loyalty (β = .41) than the influence of satisfaction on affective loyalty (β = .53). To examine differences in the path coefficients between the effects of customer satisfaction/delight on cognitive/affective loyalty, the chi-square differences test for the comparison of path coefficients was conducted (J. C. Anderson & Gerbing, 1988). The result indicated that there was a significant difference in the path coefficients between the effects of customer satisfaction on cognitive loyalty (Satisfaction → Cognitive loyalty) and the effect of customer delight on cognitive loyalty (Delight → Cognitive loyalty; Δχ2 = 11.748, p < .001) and there was a significant difference in the path coefficients between the effects of customer satisfaction on affective loyalty (Satisfaction → Affective loyalty) and the effect of customer delight on affective loyalty (Delight → Affective loyalty; Δχ2 = 33.749, p < .001).
Hypotheses 4 and 5 were supported as cognitive and affective loyalty had significant influences on conative loyalty. The influence of affective loyalty on conative loyalty (β = .78) was larger than that of cognitive loyalty on conative loyalty (β = .12). From the chi-square differences test for the comparison of path coefficients, there was a significant difference in the path coefficients between the effects of cognitive loyalty on conative loyalty (Cognitive loyalty → Conative loyalty) and the effect of affective loyalty on conative loyalty (Affective loyalty → Conative loyalty; Δχ2 = 142.669, p < .001). Seventy-eight percent of variance in conative loyalty was explained by the influences of cognitive and affective loyalties. Customer satisfaction and delight explained 61% of variance in cognitive loyalty and 72% in affective loyalty. By adding the delight construct, 5% of additional variance in cognitive and 10% of additional variance in affective loyalty was explained.
Testing of the Alternative Models
As shown in Figure 2, the proposed model was compared with two competing models as conceptual alternatives to determine the best fitting model (Bagozzi & Yi, 1988). The two alternative models also were constructed from the same latent variables and indicators but differ in the order of cognitive, affective, and conative loyalties. The alternative Model 1 indicates that there is temporal sequence among cognitive, affective, and conative loyalties according to Oliver’s (1997, 1999) sequential attitudinal loyalty development. The alternative Model 2 indicates that there is no temporal sequence among cognitive, affective, and conative loyalties based on the tripartite model of attitude structure (Back, 2001; Breckler, 1984; Li & Petrick, 2008). The two alternative models are nonnested models, which means that the one is not a constrained version of the other. Thus, the model fit indices and Akaike information criterion was compared among three models (Akaike, 1987; Rust, Chol, & Ernest, 1995). As shown in Table 4, the model fit indices for the two alternative models were similar to the proposed model. The model fit indices for the alternative Model 2 is superior to other models, and Akaike’s information criterion for the alternative Model 2 is smaller than other models, which indicates that the alternative Model 2 is more robust than the proposed model and the alternative Model 1. Therefore, we conclude the alternative Model 2 may be the best representation of these data.
Proposed Model and Alternative Models
Structural Model Results for the Proposed and Alternative Models
Note: N/A = Not available.
Significant at .001 level.
Significant at .05 level.
Discussion and Conclusions
This study examines the impact of customer satisfaction and delight on loyalty by empirically testing a model. The findings of this study have both theoretical and practical implications. The discussion presents the theoretical contributions of this study to existing hospitality literature and its practical implications for hospitality managers.
Theoretical Implications
The present study has several theoretical implications for consumer behavior research. First, as one of a few empirical studies in the context of the hospitality industry, the findings of this study broaden the knowledge on customer satisfaction, delight, and loyalty, and provide a foundation for researchers in understanding relationships among customer satisfaction, delight, and loyalty incorporating the loyalty development. As many researchers have previously highlighted the importance of customer delight (e.g., Hicks et al., 2005; Kumar et al., 2001; Oliver et al., 1997), this study extends support for the conceptualization of customer delight as one of the antecedents of loyalty; specifically, the present study shows that customer delight has significant relationships with cognitive, affective, and conative loyalties. This study challenges the proposition that customer delight is related more to loyalty than satisfaction, as many studies argued (e.g., Berman, 2005; Oliver et al., 1997; Torres & Kline, 2006).
Instead, this study showed that customer satisfaction is more strongly related to loyalty than delight. A possible explanation of the finding is that customer satisfaction influences loyalty based on the cumulative overall evaluation, which is a set of discrete service encounters or transactions with the service provider over a period of time, whereas delight affects loyalty as episodic and ephemeral emotional experiences (Bitner & Hubbert 1994; Oliver 1997; Rust & Oliver 1994). Thus, customer loyalty can be affected more by the cumulative overall satisfaction, which is summing satisfaction as both cognitive and emotional responses from the total experiences, rather than delight, which is a transient and encounter-specific emotional experience.
Although this study showed that customer satisfaction is related more to loyalty than customer delight, understanding customer loyalty related to customer’s emotions such as delight becomes more critical and helps better explain loyalty along with satisfaction. Customers appear to form attitudinal and emotional commitments to a product and service provider (Back, 2005; Schall, 2003) as there are many interactions with customers and staff during the consumption experiences in a service environment where emotions exist or are created. Recently, many researchers have emphasized the importance of emotions in marketing strategy (Arora & Singer, 2006; Dube et al., 2003; Y. Lee et al., 2009; Mattila & Enz, 2002). Specifically, delight produces emotional bonds between customers and a company, product, or service by “wowing” them, which may provide additional psychological benefits to the customers (Berry, 1995; Hirschman & Holbrook, 1982; Oliver et al., 1997). For example, one may choose to stay at a hotel for emotional experiences, such as excitement and enjoyment.
This study also provides empirical support to the existing literature on the link of customer satisfaction and loyalty. Many studies emphasized the importance of customer satisfaction on loyalty (Ajzen & Driver, 1991; Chen & Gursoy, 2001; Shoemaker & Lewis, 1999; Yoon & Uysal, 2005), although some studies found customer satisfaction does not always have a significant influence on loyalty (T. O. Jones & Sasser, 1995; Stewart, 1997). However, the finding of this study reinforces the traditional view that there is a statistically strong and critical relationship between customer satisfaction and loyalty and customer satisfaction is one of the main antecedents of loyalty (Mittal & Kamakura, 2001).
Another theoretical implication of this study is its extension to the existing literature on the customer loyalty formation process by demonstrating an alternative structure. Many earlier studies on the loyalty formation process originated from Dick and Basu (1994) and Oliver (1997, 1999), which suggested that loyalty starts with cognitive loyalty, followed by affective loyalty, conative loyalty, and ultimately action loyalty (Oliver, 1997, 1999). This present study proposes a new approach of loyalty formation process based on psychology literature (Berkowitz, 1993; Epstein, 1993; Leventhal, 1984; Zajonc, 1980); the two loyalty dimensions, cognitive and affective loyalty, are not sequentially related, but suggest that cognitive and affective loyalty influence each other and both are directly related to conative loyalty. However, this study empirically supports the model, which suggests that cognitive, affective, and conative loyalty are not sequentially linked but that all three loyalty dimensions are independent factors of attitudinal loyalty as suggested by Back (2001) and Li and Petrick (2008) based on a tripartite model of attitude structure (Breckler, 1984). According to this view, cognition, affect, and behavioral intention are separate attitude components and not all attitude components are elicited through cognitive processes (Breckler, 1984).
Practical Implications
Hospitality managers are facing a dynamic and competitive market environment as customers are becoming more sophisticated. This study provides several practical implications for hospitality managers on how to increase customer loyalty. First, managers have focused mostly on improving customer satisfaction to increase customer loyalty. However, this study shows that managers need to understand the importance of not only customer satisfaction but also delight in directly driving customer loyalty. They need to take steps to ensure the generation of both customer satisfaction and delight, which can provide stronger influences in the creation of customer loyalty because the changes in the hospitality industry demand more than customer satisfaction–style management.
Second, the finding that affective loyalty has a greater impact on conative loyalty, in comparison with cognitive loyalty, can provide important knowledge and a tool for managers to develop marketing strategies. Therefore, managers need to understand which products and services can drive more customers to affective loyalty. As an example, a loyalty program is becoming more important as one long-term marketing strategy to build customer loyalty. Most present-day loyalty programs offer points or savings as the key benefit. Based on the findings of this study, managers need to understand customers’ emotions and develop a loyalty program to effectively enhance affective loyalty, such as benefits around lifestyle events.
Limitations and Future Studies
Although this study provides several theoretical and practical implications for the hospitality industry, there are some limitations and recommendations for further research. First, this study did not evaluate prepurchase or consumption expectations. Before product consumption, preconsumption factors may anchor postconsumption evaluations and emotions. We were unable to measure these factors before consumption because a third party managed the guest information system, which required us to create a weekly schedule to pass postconsumption data. Also, some guests may not have had reservations and would not have been enumerated in the population at the preconsumption stage.
The study also did not examine any group differences between first-time and repeat guests. Such an examination may provide better understanding of loyalty creation and development related to customer satisfaction and delight because repeat visitors have expectations from their previous experiences, while first-time visitors have expectations based on external information (McKercher & Wong, 2004; Reid & Reid, 1993). Previous research suggests repeat visitors are less likely to be satisfied (McKercher & Wong, 2004), but have a stronger intention to revisit in the future than first-time visitors (Juaneda, 1996; Petrick & Backman, 2002; Sönmez & Graefe, 1998).
Future research could benefit from pursuing additional factors that may influence satisfaction, delight, and loyalty relationships. For example, action loyalty could be studied by measuring actual future stays with the hotel studied and their actual recommendation of this hotel. Extending this current study by examining unexpected negative experiences and emotions can provide a deeper understanding of customer behavior and loyalty (Schneider & Bowen, 1999; Torres & Kline, 2006). Sociodemographics, such as gender, might be different for customer satisfaction, delight, and loyalty. Particularly, for high-involvement products and services, empirical results have shown a significant difference in customer satisfaction and loyalty by gender (Mittal & Kamakura, 2001). Finally, to be generalized to other populations, the theoretical structure should be tested with different samples, such as types of accommodations (e.g., business hotel, bed and breakfast), places (e.g., other states, other countries), and across a variety of service industries (e.g., restaurant, airline, cruise).