Introduction

In the scientific literature on Relationship Marketing, some concepts have already been explored and investigated by many researchers. Quality, customer satisfaction and brand loyalty are constructs that have been permeating the literature for decades. There are a great number of works detailing antecedent and consequent relationships between these constructs, such as the ones carried out by Lipstein (1959), Kuehn (1962), Day, 1969, Jacoby and Kyner (1973), Jacoby and Chestnut (1978), Tse and Wilson (1988), Dick and Basu (1994), Jones and Sasser (1995), Garbarino and Johnson (1999), and Oliver (1997, 1999).

Based on these studies, a direct relationship between quality and satisfaction has been found, which in turn has also been intrinsically associated with loyalty, although these relationships are not linearly symmetric, and not always reciprocal.

Oliver (1999, p. 34) defines customer loyalty as “a deeply held commitment to rebuy or patronize again a preferred product/service consistently in the future, thereby causing repetitive same-brand or same brand-set purchasing, despite situational influences and marketing efforts having the potential to cause switching behavior”. Such loyalty is gradually built on some grade of familiarity and allegiance to the product/service brand that creates a condition called by the author as ‘ultimate loyalty’, which suggests a set of positive experiences (satisfaction) continuously provided to the customers by the supplier/provider.

The concept of ultimate loyalty challenges the previously established loyalty concepts. Despite this challenge, ultimate loyalty has received little attention over the last years. In spite of the detailed and thorough analysis presented by Oliver (1999), the topic has not been enough discussed by the academicians. According to Google Scholar, although the article has deserved more than 10,000 citations, no study has established a suitable scale to measure ultimate loyalty, or even identify its antecedents.

Therefore, it is supposed that is reasonable to carry out a quantitative survey suited to assess a concept that has called so much conceptual attention from researchers, although it is still at an embryonic development stage in quantitative terms. Hence, the present research aimed to: a) identify ultimate loyalty determining indicators; b) identify its antecedent constructs; c) evaluate the relationship between antecedents constructs and ultimate loyalty.

The research was conducted with Harley Owners Group® (HOG), in the city of Belo Horizonte, state of Minas Gerais, Brazil. The members of this group are well-known Harley Davidson’s loyal clients (Oliver, 1999; Malhotra, 2001).

Literature Review

In early studies on loyalty, the focus was simply on the repurchase behavior. From such a viewpoint, loyalty was defined as a repeated purchase from a supplier (Frank, 1967), from an operational measurement perspective. By taking the same approach, Brown (1952) classified loyalty into four patterns, named: undivided loyalty; divided loyalty; unstable loyalty; and no loyalty.

From the same perspective, and focusing on repurchase standard, Lipstein (1959) and Kuehn (1962) measured loyalty based on the odds of gaining new purchases. Some authors such as Day (1969) and Jacoby and Chestnut (1978) stated that a simple repurchase behavior analysis would not be enough to distinguish true loyalty from spurious loyalty.

Other authors have suggested the integration of loyalty into other concepts published in the marketing-related literature (Jacoby & Kyner, 1973). Selnes (1993) stated that customer loyalty will call for future consumption probability, continuity of service agreements, reduction of brand switching likelihood, and build positive Word-of- Mouth communication (WOM).

To measure loyalty, Fornell (1992) proposed a method based on repeated purchase intention and price tolerance, while Srinivasan, Anderson and Ponnavolu (2002) pointed two factors they’ve considered to be the most important, i.e., positive word-of-mouth and price tolerance.

Traditional research based the concept of loyalty in behavioral indicators such as the proportion of purchases made from the same supplier, purchase sequence, and the probability of buying from a given supplier (Dick & Basu, 1994). Following the recommendation of Jacoby and Chestnut (1978), these authors developed a model considering attitudinal and behavioral loyalty aspects and by combining these two dimensions, they suggested the existence of four types of loyalty, namely no loyalty, spurious loyalty, latent loyalty, and loyalty.

Oliver (1999) argues that loyalty has been developed in stages, that is to say, cognitive, affective, conative and action loyalty. The author goes deeper into the analysis by using a classification pattern based on two dimensions, named individual fortitude and community/social support. Individual fortitude would be the extent or degree of willingness, a behavior that makes the consumers to ignore, refuse, or fail to see the advertising appeals of competing brands (Oliver, 1999). The combination of these two dimensions entails four types of loyalty:

Product Superiority, reflecting the traditional view of loyalty as a result of the superior quality, which implies a brand preference.

Village envelopment by which the individual is protected from external influences, and is encouraged to repeatedly consume the brand by receiving information and regular updates.

Determined self-isolation that leads customers to appreciate the product superiority and turn away from offers advertised by competitors.

Immersed self-identity, in which consumers want the association with the product or service, they consciously affiliate to the environment knowing it supports the association and, in extreme situation, are rewarded for their patronage. This phase the author calls ultimate loyalty.

Oliver (1999) suggests that ultimate loyalty, the combination of high individual fortitude and high community/social support, is the one that interest managers. It can be truly classified as loyalty as consumers take on a commitment and involvement so intensely that they become brand boosters and promoters.

Some easily noticeable behaviors characterize ultimate loyalty. The consumer’s psychological state favors certain degree of irrational behavior, which explains and justifies the consumer’s preference, even when she/he is exposed to competitors’ advertisings offering higher quality products, readier availability and more competitive prices. Although many advantages offered by competitors, customer remains loyal, and will not switch her/his brand preference. At this stage of loyalty, consumers maintain a state comparable to love in its many manifesting ways, that’s in this case, to goods (Ahuvia 1992 & Fournier 1998). In this consumption context, apart from the sensual side, two other aspects of the phenomenon should be considered, that’s to say, adoration and total commitment.

Oliver (1999) makes clear that, among the factors required to build ultimate loyalty, there is one factor that depends exclusively on the company’s sponsorship, by making local investments in the development, support, and encouragement of the customer’s community.

Other conditions that precede ultimate loyalty are: the product must have a unique configuration that makes it desirable; an identifiable and profitable segment of customers should consider the product desirable; the product must be worshiped, at least by potentially loyal customers; and, the product must be able to be included in a social network, so that the consumer only feels part of the village by having it.

Product Uniqueness (PU)

As Franke and Scheirer (2007) proposed, the more a customer identifies a product tailored to meet her/him expectations, the more she/he will feel motivated to purchase it by comparing it to a benchmarking product regardless its functionality or appearance. According Weidmann’s (2007), regarding value of the framework of Psychographic Traces, some purchases are decided by exclusivity, self-identification, pursuit of prestige, and even by visibility (Cypriano, 2013). These luxury brand consumers match many psychographic profiles such as the one that identifies individuals searching for unique products classified as ‘single identity’. These customers seek exclusivity to express themselves in a unique way. This discussion suggests the hypothesis one.

H1: product uniqueness has a positive effect on ultimate loyalty.

Adorable Product (AP)

Product shape, transform and give life to the universe, and help individuals find and play their role (McCraken, 2003). As previously stated by Ahuvia (1992) and Fournier (1998), love has many manifesting ways that may sometimes take the form of purchasing. Oliver (1999) explains that a general commitment results from a real fondness for the product or service. This kind of commitment is considered to be an adoration or devotion based on maintaining something to avoid a feeling of loss previously experienced by the customer. Such a behavior is common among children. However, adults also express love for goods such as automobiles they may name or treat in ways typically reserved for their loved ones. According to Wiedmann’s (2007) model, these situations explain the consumer’s willingness to pay more for one brand than for another similar, even if it has a lower quality one.

H2: product adorability has a positive effect on ultimate loyalty.

Product should be desirable by an important market segment (DN)

Some goods embody social status. They also define the inclusive and exclusive parameters, and limit the access of certain classes of consumers (Douglas and Isherwood, 2004). To build loyalty of these customers, a product must arouse the interest and desire of a large market segment, and be still prohibitive enough to convey status. These products must have the features looked for by a well- defined target market, and must be available for purchase. A demographic profile of well characterized customers must encompass cultural, social and personal aspects.

H3: There is a positive relationship between the product that belongs to a well-defined market niche and ultimate loyalty.

Product must have the capacity to be embedded in a social network (SN)

Ultimate loyalty begins with individual good will and determination, and reaches the apex with consumer self- identification, who buy the product regardless other offers available in the market, provided that it is supported by users from the same community (Oliver, 1999).

The purchase of a new product is often made with the support of a reference group that, in such cases, indicates the existence of a probable interpersonal relationship focused on brand, which may cause an increase in the consumption of such a brand (Schouten, McAlexander and Koenig, 2002). By analyzing the psychographic profile of luxury goods consumers, Cypriano (2013) noticed that there is a group identified as “Always Impressing” who is interested in impressing others, and manage their lives to be seen as someone of prestige, and so show their social importance. The main motivation of this group is to show they belong to a specific social group. In some social networks, some aspects chosen and shared by the group influence behavior, speech, and consumer brand experience. These facts give approval and the necessary reinforcement to members, and provide even more a feeling of security to them.

H4: There is a positive relationship between brand community membership and ultimate loyalty.

Individual Fortitude (IF)

Oliver (1999) highlights that individual fortitude must be present to build ultimate loyalty, that is, the individual, and only him or her, through his own will, determination and strength of character will be involved in a relationship with a brand to the point of becoming an ultimate loyal customer. Therefore, a conclusion arises, the first step depends on customer closeness to product or service.

Oliver (1999) defines fortitude as “the extent or degree of willingness with which a consumer ignores, refusal or failure to see the promotional appeal of competing brands” (pp. 37). Customers gradually go from low to high fortitude that takes place in a continuous step-by-step process rather than big jumps. At low levels of fortitude, consumers have only brand-related information. At higher levels of fortitude, consumers develop action inertia, and as well fierce defenses against competitive influences that come near blind faith (Oliver, 1999). In the same article, Oliver points out that the passage through the barrier between trust that a product is superior to the stage of individual fortitude and determination by buying the same brand remains somewhat nebulous. Thus, it can be proposed:

H5: There is a positive effect of individual fortitude on ultimate loyalty.

Research Design

This case study employs a two steps descriptive approach, qualitative and quantitative. In the first step, it was identified, through semi-structured interviews, determining attributes in shaping group loyalty, to validate the attributes identified in the literature. Based on such data, a survey questionnaire was elaborated with 32 statements, measured on a 7-point interval scale, ranging from (1) strongly disagree to (7) totally agree.

Indicators used to measure the constructs (Appendix) were based on the following authors: Product uniqueness: Srinivani et al. (2002), Cavusgil, Zou and Naidu (1993), Franke and Schreier (2007), Chang and Chen, (2008) and Cypriano (2013); Adorable product: Cypriano (2013); Product with a market segment: Strehlau and Aranha (2004); Product embedded in a social network: Srinivani, et al. (2002); Individual fortitude: Fraering and Minor (2013); and ultimate loyalty: Morgan and Hunt (1994), Zeithalm, Berry and Parasuraman (1996); Blocker (2012); and Human and Naúde (2014). Thirty-two statements were sorted according to the variables, i.e., product uniqueness (5 items), adorable product (4 items), product with a market segment (6 items) product embedded in a social network (5 items), individual fortitude (5 items) and ultimate loyalty (7 items).

A convenience sample was extracted from the members of the HOG-MG and other HOG members in the city of Belo Horizonte, Minas Gerais, Brazil. The Data were collected by using a structured questionnaire form available in Google docs and by in-person interviews with HOG members at events promoted at HD dealers and HOG meetings in the so-called “Harley meeting points” scattered throughout the metropolitan area of the city of Horizonte, MG, Brazil.

The qualitative survey included fifteen interviewees, all of them active members of HOG chapter MG. For the quantitative research, a sample of 230 interviewees were surveyed.

For data analysis, Partial Least Squares Method (PLS) was used, through PLS Smart 3.0 software, because it is a multivariate technique aimed to maximize the explained variance, fit small samples, and also because it is more effective to detect deviations from normality.

At the end of the data collection stage and after excluding questionnaires with some inconsistency, two hundred and eleven interviewees were gathered, out of which twenty-five (25) were filled in via Google docs. An analysis of variance (ANOVA) was estimated, which indicated no significant differences between the two samples.

Findings

Through the semi structured interviews, information and opinions of HOG members were collected regarding the most varied topics related to HD brand, the importance of HOG, complaints, suggestions, other motorcycle brands, among other subjects. These interviews were used in the refinement of the questionnaire used in the quantitative stage.

During the interviews, the answers of HOG-MG members were collected, and some of them were highlighted as shown in Figure 1.

The answers of HOG members listed in Figure 1 show that some profile characteristics of the interviewees are common, and some are even noteworthy:

• The deep sense of affection of HOG members for their motorcycles that are often treated and referred to as a loved person;

• How the HOG members consider HD unique and incomparable to other motorcycles;

• The benefit of bringing youth spirit back to the HOG members is another important point;

• The sense of community is notable when HOG members claim to have the Harley point at home to receive their HOG fellows.

Turning to quantitative step, among the interviewees, 87.20% are male, 68.80% are married and 89.7% are undergraduate or have graduate degrees. Regarding their occupations, 28.3% are business owners, and 25.8% are hired by the private sector. 56% are older than 41 years old. Thirty-one percent earn up to R$ 8,000.00 monthly (+/- US$2,000), however there is a significant group of 21.6% that earns over R$ 20,000.00 monthly (+/- US$5,000).

Hair et al. (2005) and Mesquita (2010) recommend conducting an exploratory data analysis before applying multivariate statistical technics due to the importance of verifying the assumptions. That stage consisted in checking the linearity, normality, and outlier’s assumptions.

To evaluate linearity, the Spearman’s correlation matrix was used, and it is found that most of the correlation coefficients are significant at 1% level, confirming the linearity of the relationship between manifest variables. By analyzing the results from Kolmogorov-Smirnov and Shapiro-Wilks tests, it was observed data non normality distribution, which reinforces the choice of PLS. The evaluation of outliers identified only one variable related to many outliers, but assuming that it is a typical sample behavior, it has not been disregarded.

The analyses of some important descriptive parameters that help characterize the sample being studied are shown in Table 1. The construct with the highest mean values correspond to Adorable Products - 6.1927, while the lowest average is Individual Fortitude - 5.2529. In general, the averages were high, since the maximum value is 7.

By analyzing the indicators individually, it was found that the highest average value was AP 2 - The owner of a HD loves his motorcycle – 6.4 (maximum was 7); the second refers to AP 3- Customers identify with the brand, 6.3; and the lowest average is 4.2 for the variable LOY 6- The HD customer is not very price sensitive when purchasing HD accessories. Even if it is slightly, the result regarding LOY 6 contradicts one of the most expected outcomes in consequences of loyal- ty, i.e., the low sensitivity of customers to price changes. On the other hand, the findings related to the adorable products confirm, and point to that as the most important characteristics to build loyalty.

After a preliminary analysis, the focus turned to model estimation. Initial results indicated the need to exclude variables with low factor loadings (AP1, ND1, ND3, ND4 and SN4). After model estimation, the first step is to evaluate con- structs’ reliability, convergent and discriminant validity. Reliability shows the internal consistency of the measure, conver- gent validity is the extent to which a measure positively correlates with alternative measures of the same construct, and the discriminant validity refers to the degree to which the measure of different constructs differs from each other. The results from reliability tests and convergent validity are shown in Table 2.

As can be seen, all constructs showed CR above 0.7, attesting to their reliability. Regarding convergent validity, only indi- vidual fortitude showed AVE value slightly below the recommended minimum (0.5), but above 0.4, which is acceptable in dealing with exploratory research (Nunnaly, 1994), besides the fact that all indicators showed loadings greater than 0.5, values that attest to convergent validity.

The cross-loadings matrix assessed discriminant validity. It was noticed that the factor loadings in the original constructs of indicators are all higher (above 0.50) than in others, which testified the discriminant validity criterion (Chin 1998).

The second phase in PLS-SEM implies establishing predictive capacity and analyzing relationships between constructs. The evaluation criteria comprise the level of significance of the coefficients of the structural path and the coefficient of determination R2.The estimated structural model can be seen in Figure 2, while Table 3 shows the PLS output.

By analyzing the results shown in Figure 2 and Table 3, it is concluded:

H1- proposed that there is a positive relationship between product uniqueness and ultimate loyalty: with coefficient of 0.262, t-value of 3.791, significance level 0.000, it can be observed that the regression coefficient is significant to support hypothesis one. There was already an idea of the influence of this construct on shaping ultimate loyalty based on the sample profile. Motorcycle customization is something relevant to HOG members to the point that “each motorcycle is unique”, “there are no two identical motorcycles” and “it is common to find a wide variety of the motorcycles, even at an event that gathers over two thousand HD motorcycles.” This assertion supports the claims of Franke and Scheirer (2007) and Wiedmann (2007) regarding the desire to purchase unique and exclusive products.

H2- the adorable product has a positive effect on ultimate loyalty: coefficient of 0.173, t-value of 2.657, and significance level 0.000, supports hypothesis two. Such a hypothesis is based on the propositions of Ahuvia (1992) and Fournier (1998) regarding love for products, and deals with the relationship between the HOG members and their motorcycles, and the result shows that this is peculiar bond. It is easy to find HOG members who own more than one motorcycle because they cannot afford getting rid of their first motorcycle. To ratify what’s said above, it is enough to observe that the variable with the highest average value is the AP-2 “The HD owner loves her/his motorcycle”, 6.4171.

H3- There is a positive relationship between the product that belongs to a well-defined target market and ultimate loyalty: coefficient of 0.212 and t-value of 4.232 indicate that the regression coefficient is significant at 0.000 level. A support to this hypothesis was also expected because the average profile of HOG member has been well defined and identified. The product type provides identification of the target market as it confers social status to the owner (Douglas and Isherwood, 2004).

H4- There is a positive relationship between belonging to a brand community and ultimate loyalty: with a coefficient of 0.058 and t-value of 1.133, it does not state that the regression coefficient is significant. Thus, the hypothesis was not supported.

H5- There is a positive effect of individual fortitude on ultimate loyalty: with a coefficient of 0.349 and t-value of 5.201, it confirms the regression coefficient significance at 0.000 level. To understand the evidence of such a hypothesis, one must analyze the sample average profile in more details. Most of the surveyed people are older than 40 years and are professional business owners or employees at private companies, with higher educational level, including postgraduate degree, characteristic of more mature individuals, who are usually more self-confident, determined and self-sufficient. Therefore, as highlighted by Oliver (1999), such individuals, and only them, through their own will, determination, and strength of character, will engage in a relationship with a brand to the point of being an ultimate loyal consumer.

The coefficient R2 is 0.742 (adjusted R2 of 0.736), explains 74.2% of the ultimate loyalty variance, which well explains how ultimate loyalty is influenced by the five independent constructs proposed for the model.

Regardless the outcome, rejecting the positive relationship between the participation in a brand community and ultimate loyalty, it is relevant to go deeper in this analysis as it has found the existence of certain antagonism between these two constructs, i.e., social network and individual fortitude.

Although both constructs, in their own way, help build ultimate loyalty, by evaluating the influence of each of them, the conflict between these variables is established. To what extent does social network help build ultimate loyalty? While the paragraph above highlights the group strength, and value the relationships of members, the individual fortitude is based on the will and self-determination of everyone individually. This means that the degree of involvement and willingness to remain loyal to a brand is something very personal.

In modern society, one cannot disregard the influence of social networks on the relationships of all the protagonists. According to Almeida et al. (2013), “members of these communities look for this brand as a lifestyle group, adventure, movement freedom, and possibility of meeting people and ...,” despite the importance of social relationships, in this case, individual fortitude prevails.

The second point that deserves to be highlighted is the fact that during the qualitative survey research done mainly with HOG members, it was clear how members value the group, which cannot be said about nonmembers and former members. Furthermore, there was a kind of rivalry or jealousy of some members of other motorcycle clubs like HOG. It is not possible to generalize such feelings as there are many motorcycle club members who keep friendly relationships with HOG members with no problem. It is also observed that some former HOG members have some grievance and complaints about the group, while some show total ignorance regarding HOG functioning, but still make negative comments regarding the group. In Pinto’s work (2011), it seems that in São Paulo the rivalry is even stronger, and HOG members are nicknamed “coxinhas” (a Brazilian Portuguese pejorative term that stands for a kind of “dandy person”) by some members of other HOGs.

Currently, the influence of social networks on many fields of human activities has been much researched, and such influence has been very often supported. In this study, it has not been noticed, probably due to a mistake made to change a part for the whole group, i.e., HOG members by all loyal HOG members. The fact is that no one can disregard the influence of social networks on the relationships among all players, and for this reason it is pertinent to highlight the unsupported hypothesis, and hence the importance of social networks to build loyalty.

In the referential literature, i.e., the theoretical basis for this work, many definitions have been studied and presented throughout this paper.

The model proposed by Oliver (1999), in which the conditions are prerequisites to build ultimate loyalty, was almost all virtually supported (product uniqueness, adorable product, product desired by an important market segment, and individual fortitude).

Regarding loyal customer behavior described by Oliver (1999) and by Srinivasan, Anderson and Ponnavolu (2002), there is a pointed feature that gives the customer higher tolerance to prices. However, this condition has not been supported. In qualitative research, the statements related to price tolerance by HOG members were rejected by a large number of participants.

The suggested repurchase was observed and supported by the large volume of motorcycle accessories purchased such as clothes for motorcyclists and companions, and even decoration items. Nobody purchases a motorcycle so often, but repeated purchases are done at HD® stores.

Several authors such as Srinivasan, Anderson and Ponnavolu (2002) and Oliver (1999) point out the positive word-of- mouth as a typical publicity among loyal customers that was supported by the survey that reported a variable with very high average values.

During the survey period one HOG member, ultimate loyal customer up to that time, left the HOG for dissatisfaction with the HD dealer and said he/she would change the brand of his/her motorcycle. This fact shows how important the role of the dealership is.

Final remarks

In this research, it was identified and assessed the constituent indicators and antecedent constructs of ultimate loyalty. This research approaches the subject from a different perspective when compared to earlier studies as it focuses on a group of known by being ultimate loyal customers of a brand.

The constructs that drive ultimate loyalty, identified in Oliver (1999) work are: product uniqueness; adorable product; product desirable by a defined market segment; product embedded in a social network and individual fortitude. From the results, it is observed that, except for belonging to a social network, they all are important predictors of ultimate loyalty.

As an academic contribution, the study addressed a topic that deserves further investigation, despite the progress achieved over the last years. The concept of ultimate loyalty has been proposed by Oliver (1999) and subsequently received little attention. Few articles addressing the concept were found, although the great number of citations. Thus, identifying antecedent constructs and developing a scale to measure them, even though in the embryonic stage, constitutes the main research academic contribution.

As a managerial contribution, observing the literature on customer’s loyalty, usually there are two examples of companies (Harley Davidson and Winnebago) which have achieved this degree of preference of their customers and it is possible to conclude that there are more than mere coincidences between the strategies of these companies. There is a deliberate strategy in these organizations, established by the administration that shows even for the most basic levels of frontline employee the importance of the management model. There is a philosophy that supports and maintains the process, all with great effort and professionalism. From the results of the research, the constructs most influential in the development of ultimate loyalty were identified and with this information, it is possible to implement and manage loyalty programs, focusing efforts on the most determining variables. Therefore, it should be established strategies to enhance brand image, highlighting the product uniqueness and adorability. As the target market is composed with individuals of high fortitude, special attention should be given aiming to direct communication efforts to this public.

This study has limitations of various orders. The research focused in the region of Belo Horizonte, Brazil, limiting the capacity of generalization because the subject is of far greater importance and goes beyond the boundaries of this state. The number of respondents is two hundred and thirty and only two hundred and eleven were valid, which is a small sample.

Moreover, the findings of this study should be analyzed from the perspective of a group loyal to a particular mo- torcycle brand, so for any application to other products or markets corrections and adjustments should be properly carried out.

Ultimate loyalty arouses interest in all those who teach and work in relationship marketing. As suggestion for future research, other industries such as jewelry, electronic, cloth- ing, among others, should study the topic and get deeper knowledge in order to enhance academic knowledge and the improve management practices.

The product embedded in a social network showed little influence on building ultimate loyalty, however it is pos- sible that, in another market segment, or even in the mo- torcycle market segment, with another approach, different results may be achieved.

Other issues that should be studied include characteristics and strategies of the companies that were successful in building and keeping the loyalty of their customers and ultimate loyalty consumer behavior. The latter are also sug- gestions for further research.

References

Ahuvia, A. (1992). “For the Love of Money: Materialism and Product Love,” in Meaning Measuring, and Morality of Material- ism, Floyd Rudmin and Marsha Richins (Eds.). Provo, UT; Association for Consumer Research, 188-98.

Almeida, L. O. V., Tavares, M. C., Urdan, A. T., Teixeira, D. J. (2013, set/ dez). Comunidades de Marcas e os proprietários de Harley-Davidson de Belo Horizonte. Revista Gestão & Tecnologia, Pedro Leopoldo, v. 13, n. 3, pp. 235-257.

Blocker, C. P. (2012). The Dynamics of satisfaction and loyalty after relational transgression. Journal of Service Marketing, v. 26, ed.2, pp. 94-101.

Brown, G. H. (1952). Brand Loyalty- fact or fiction? Advertising Age, 23 (9), pp. 53-55.

Cavusgil, S. T., Zou, S., Naidu, G. M. (1993, third quarter). Product and Promotion Adaptation in Export Ventures: An Empirical Investigation. Journal of International business Studies.

Chang, H. H., Chen, S. W. (2008). The Impact of Customer interface quality, satisfaction and switching costs on e-loyalty: Internet experience as a moderator. Computer in Human Behavior, Elsevier Ltd.

Chin, W. W. (1998). The partial least square approach for structural equation modeling. Marcoulides, G. A. (ed). London Erlbaum Associates, pp. 295- 336.

Cypriano, I. M. B. M. P. (2013). O consumidor brasileiro de marcas de luxo: Uma descrição e segmentação psicográfica por valores. Dissertação apresentada à Escola de Administração de Empresas da Fundação Getúlio Vargas. São Paulo, SP, Brasil.

Day, G. S. (1969). A Two-dimensional concept of brand loyalty. Journal of Marketing Research, 9, pp. 29-36.

Dick, A. S., Basu, K. (1994). Customer loyalty: toward an integrated conceptual framework. Journal of the Academy of Marketing Science, 22(2), pp. 99-114.

Douglas, M., Isherwood, B. (2004). O Mundo dos Bens- para uma antropologia do consumo. Rio de Janeiro: UFRJ.

Fornell, C. (1992). A national customer satisfaction barometer: The Swedish experience. Journal of marketing, 56 (1), pp. 6-21.

Fournier, S. (1998). ”Consumer and Their Brands: Developing Relationship Theory in Consumer Research.” Journal of Consumer Research, 24, pp. 343-73.

Fraering, M., Minor, M. S. (2013). Beyond loyalty: customer satisfaction, loyalty and fortitude. Journal of Service Marketing, 27(4),pp. 334-344.

Frank, R. E. (1967). Correlates of Buying behaviors for grocery products. Journal of Marketing, Octuber, pp. 48-53.

Franke, N., Schreier, M. (2007). Product Uniqueness as a driver of customer utility in mass customization. Marketing Letters, A Journal of Research in Marketing, Springer Science+ Business Media, LLC.

Hair, J. F., Jr. et al. (2005). Fundamentos de Métodos de Pesquisa em Administração. Porto Alegre: Bookman.

Human, G., Naudé, P. (2014). Heterogeneity in the quality-satisfaction- loyalty framework. Industrial Marketing Management, 43, pp. 920-928.

Jacoby, J., Chestnut, R. W. (1978). Brandy loyalty. New York: John Wiley & Sons Inc.

Jacoby, J., Kyner, D. (1973). Brand loyalty vs. repeat purchasing behavior. Journal of Marketing Research, X, pp. 1-9.

Kuehn, A. (1962). Consumer brand choice as a learning process. Journal of Advertising Research, 2, pp.10-17.

Lipstein, B. (1959). The dynamics of brand loyalty and brand switching. Fifth Annual Conference of the Advertising Research Foundation. New York: Advertising Research Foundation, 101-108.

Malhotra, N. K. (2001). Pesquisa de Marketing: uma orientaçãoaplicada. (3a ed.). Porto Alegre: Bookman.

McCraken, G. (2003). Cultura & Consumo: novas abordagens ao caráter simbólico dos bens e das atividades de consumo. Rio de Janeiro: Mauad.

Mesquita, J. M. C. (2010). Estatística Multivariada Aplicada à Administração (1a ed). Curitiba: Editora CRV.

Morgan, R. M., Hunt, S. D. (1994). The commitment-trust theory of relationship marketing.Journalof Marketing, 58 (3), pp. 20-38.

Nunnaly, J. C.; Bernstein, I. H. (1994). Psychometric Theory. New York. Oliver, R. L. (1997). Satisfaction: A behavioral perspective on the consumer. Mc Graw- Hill.

Oliver, R. L. (1999). Whence consumer loyalty. Journal of Marketing, v. 63, special issue, pp. 33-44.

Oliver, R. L. (1997). Satisfaction: A behavioral perspective on the consumer. Mc Graw- Hill.

Pinto, F. R. T. (2011). Você tem uma moto ou uma Harley? Vínculos com a marca Harley-Davidson em São Paulo. Tese de Doutorado em Antropologia Social- USP, São Paulo, Brasil

Schouten, J. W., McAlexander, J. W., Koening, H. F. (2002). Building Brand Community. Journal of Marketing- v. 66, pp. 38-54, jan.

Selnes, F. (1993). An examination of the effects of product performance on brand reputation, satisfaction and loyalty. European Journal of Marketing, 27(9), pp. 19-35.

Srinivasan, S. S., Anderson, R., Ponnavolu, K. (2002). Customer loyalty in e-commerce: an exploration of its antecedents and consequences. Journal of Retailing, 78, pp. 41-50.

Strehlau, S., Aranha, F. (2004). Adaptação e validação da Escala de consume de status (SCS) para uso no contexto brasileiro. Revista Administração. Belo Horizonte, v. 3, n. 1, pp. 9-17.

Wiedmann, K. P., Hennigs, N., Siebels, A. (2007). Understanding and measuring luxury value: a multidimensional model of consumer’s luxury perception. American Marketing Association Winter Educator’s conference proceedings, San Diego, CA, 18, pp. 393-395.

Zeithalm, V. A., Berry, L. L., Parasuraman, A. (1996). The behavior consequences of service quality. Journal of marketing, v. 60, n. 2, pp. 31-46.