BRAND LOVE AND POSITIVE WORD OF MOUTH: THE MODERATING EFFECTS OF EXPERIENCE AND PRICE

Structured abstract Purpose – This study tests and extends Carroll and Ahuvia’s (2006) model on the antecedents and outcomes of brand love by examining the moderating effects of experience and price on the relationship between brand love and offline word of mouth (WOM), and between brand love and electronic WOM (eWOM). Design/methodology/approach – An online survey was conducted on Facebook, which generated 342 valid responses. The data were analyzed using partial least squares structural equation modeling. Findings – The results mainly support the model by confirming that the self-expressiveness of the brand and trust are positively associated with brand love. No association between hedonic product type and brand love was found. The study’s findings also suggest positive associations between brand love and both types of WOM. Experience and price were found to strengthen the relationship between brand love and offline WOM, but not the relationship between brand love and eWOM. Originality/value – This study contributes to the literature by introducing and testing an extended model of brand love. It tests two moderating effects on the relationship between brand love and WOM.


Introduction
Phrases such as "I absolutely love this bag!" or "I love Coca Cola!" are commonly used by consumers (Ahuvia, 2005).Although these expressions are used more lightly when talking about products than when talking about other human beings, prior research shows that consumers are able to experience a deep emotional attachment towards brands (Albert and Merunka, 2013;Batra et al., 2012;Huber et al., 2015;Langner et al., 2015).In fact, previous research suggests that a customer's deep emotional attachment to a brand is the most reliable measure of a strong brand, and should therefore be a primary objective for brand management or customer relationship marketing (Pawle and Cooper, 2006).For example, lasting relationships may be forged based on financial benefits to customers, such as loyalty discounts (Sheth and Parvatiyar, 1995), or due to the limited number of options available (Albert and Merunka, 2013); however, without emotional connection those relationships are easily dissolved if such "relationship enhancers" disappear.Therefore, repeat purchase behavior alone is an insufficient measure of a strong relationship and an inadequate objective for customer relationship management (Sheth and Parvatiyar, 1995).Thus, while a strong brand with steady and profound customer bonds is acknowledged to be the main objective of brand management (Esch et al., 2006), brand love is the ultimate goal of the customer-brand relationship.
Brand love has its roots in the theories of interpersonal love and relationships (Batra et al., 2012;Langner et al., 2015).Fournier (1998) was among the first scholars to provide evidence that consumers are able to emotionally bond with brands in a similar way to that observed in personal relationships.This has been supported by more recent research, which shows that consumers can develop a relationship with a brand that evokes emotions similar to love (Albert et al., 2008;Carroll and Ahuvia, 2006;Wallace et al., 2014).Emotions towards a brand have been found to play an especially strong role in the propensity to continue a relationship (Drennan et al., 2015;Kudeshia et al., 2016;Nobre, 2011) and when conflicts occur in relationships (Pawle and Cooper, 2006).Although brand love is acknowledged as an important construct of consumer-brand relationships (Batra et al., 2012;Huber et al., 2015), research on its drivers and consequences for such relationships remains limited (Albert and Merunka, 2013;Fetscherin, 2014;Grisaffe and Nguyen, 2011), as does research on the variables that moderate these relationships (Vernuccio et al., 2015).
To develop brand love, a consumer must have a positive brand attitude and positive experiences, as well as a feeling of psychological proximity to the brand (Joji and Ashwin, 2012).A beloved brand is an integrated part of the consumer's self-expressiveness, and by using a certain brand and demonstrating love towards it, the consumer expresses his or her self to others (Albert et al., 2008;Huber et al., 2015;Wallace et al., 2014).In addition to selfexpressiveness, brand trust (Albert and Merunka, 2013;Drennan et al., 2015) and a brand's hedonic value (Batra et al., 2012;Huber et al., 2015) have also been shown to be antecedents of brand love.Trust is suggested to play a crucial role in all long-term relationships (Morgan and Hunt, 1994), and thus is a focal driver of brand love.Hedonic and utilitarian motives in consumer-brand bonding and in the formation of brand love have also been acknowledged as being of high importance (Grisaffe and Nguyen, 2011).Prior evidence suggests that hedonic value affects shopping experiences (Babin et al., 1994), as well as the way brands are assessed -that is, hedonic brands are assessed more intuitively compared to brands offering utilitarian value (Suh, 2009).Carroll and Ahuvia (2006) also show that consumers might be less loyal to purely hedonic brands than utilitarian brands or brands delivering both types of value.The effect of self-expressiveness, brand trust, and hedonic value on brand love have been shown in prior studies (e.g.Albert and Merunka, 2013;Drennan et al., 2015;Huber et al., 2015); however, none of these studies have examined the compound effect of these constructs on brand love.
Several desirable outcomes of brand love have been proposed, such as brand loyalty, positive word of mouth (WOM) (Carroll and Ahuvia, 2006;Albert and Merunka, 2013;Wallace et al., 2014), stronger purchase intentions (Fetscherin, 2014;Kudeshia et al., 2016;Pawle and Cooper, 2006), and greater willingness to pay price premiums (Albert and Merunka, 2013;Fedorikhin et al., 2008;Thomson et al., 2005) and accept product failures (Grisaffe and Nguyen, 2011).Owing to the growing influence of peer reviews and suggestions on consumers' buying decisions, positive WOM in both offline and online environments (i.e., electronic WOM [eWOM]) are the key objectives of today's marketing activities (Chen and Xie, 2008).Although Batra et al. (2012) and Kudeshia et al. (2016) report WOM as an outcome of brand love, research has yet to provide a thorough explanation of how brand trust, self-congruence, hedonic product brands, and brand love together foster WOM in offline and online channels.
In addition, constraining a study only to direct relationships of WOM and its antecedents is insufficient to fully understand the phenomenon.The effects of factors that might strengthen or weaken these relationships also need be taken into account.Previous literature suggests at least two important moderators -brand experience (Cooil et al., 2007;Yi and La, 2004) and price perception (Monroe, 2005;Xia et al., 2004) -which are found to be important variables affecting consumer behavior.For example, Ranaweera and Menon (2013) show that brand experience moderates the satisfaction and WOM relationship, and De Matos and Rossi (2008) and McKee et al. (2006) link price perception to consumers' brand attachment and recommendation intentions.However, no prior evidence exists regarding how these variables affect the brand love-WOM relationship.
This study aims to address these limitations.First, we propose and empirically test a conceptual model that explains the formation of brand love and positive WOM through the constructs of the self-expressive brand, brand trust and the hedonic product.Second, we explore the direct and indirect effects of these constructs on positive WOM in the offline and online context.Third, we examine the moderating effects of price perception and brand experience on the effects of brand love in relation to positive WOM.This paper reaffirms previous evidence about the influence of brand trust, self-expressiveness and hedonic value on brand love.In addition, novel findings are provided in the form of evidence for the compound effect of the three factors on brand love and WOM behavior, and how brand love affects offline WOM and eWOM through the moderation of experience and price perceptions.
The remainder of the paper proceeds as follows.First, we define our study constructs, describe our framework and develop hypotheses regarding how a self-expressive brand, brand trust, and hedonic value drive brand love, offline WOM and eWOM.In addition, we discuss how experience and price affect the relationship between brand love and positive WOM.This is followed by a description of the methods and measures used to test the framework.
Subsequently, the results and implications of the study are presented.

Research hypotheses
The conceptual model and hypotheses of the present study are shown in Figure 1.We propose that the self-expressive brand (brand-self connection), brand trust and the hedonic value that the brand delivers are the antecedents of brand love.WOM serve as the behavioral outcomes of brand love.Brand experience and price perception moderate the effects of brand love on offline WOM and eWOM.Two control variables -age and gender -are included in the model relating to WOM and eWOM, since the literature suggests that there are significant differences in how younger generations spread WOM compared to older generations (Sago, 2010), and females compared to males (Zhang et al., 2014).
Place Figure 1 about here

Brand love
Love is understood as a subjectively experienced combination of the dimensions of intimacy, passion and commitment (Aron and Westbay, 1996), whereas brand love is "the degree of passionate emotional attachment a satisfied consumer has for a particular trade name" (Caroll and Ahuvia, 2006, p. 5).Although brand love positively correlates with attitude, involvement and satisfaction, it is separate and distinct from those constructs (Albert and Merunka, 2013;Thomson et al., 2005).Positive attitude alone does not require the formation of brand love (Carroll and Ahuvia, 2006), does not link the brand to the self-concept, and does not cause an individual to maintain a relationship with the brand when more attractive options are introduced (Thomson et al., 2005).In addition, although brand involvement shares some characteristics with brand love, such as both being based on needs, values and interests (Brakus et al., 2009) and being driven by perceived importance and personal relevance (Zaichkowsky, 1985), the constructs are distinct from each other.While brand love requires emotional attachment towards a brand in order to emerge, brand involvement is only connected with cognitions (Batra et al., 2012;Thomson et al., 2005).
Brand love also differs from satisfaction.Carroll and Ahuvia (2006) list several ways in which brand love differs from satisfaction, including: (1) satisfaction is conceptualized as a cognitive judgment, whereas brand love is more affective; (2) satisfaction is typically related to transactions, whereas brand love develops over time and is more related to long-term brand relationships; (3) satisfaction is related to expectancy disconfirmation, whereas brand love does not require these conditions and (4) satisfaction does not require integration of the brand into the consumer's identity, or their willingness to declare love, whereas brand love does.
Attachment to a brand, which is a precondition for brand love, develops over time and requires affective memories that connect the object to the self (Holmes, 2000;Mikulincer et al., 2001).Therefore, brand love is found to be constructed on a variety of affective-laden perceptions, such as self-brand integration, passion-driven behaviors, positive emotional connections, a long-term relationship, positive overall attitude (attitude certainty and confidence) and separation distress (Batra et al., 2012;Thomson et al., 2005).Therefore, brand love is a sign of a far stronger consumer-brand relationship than mere positive brand attitude or satisfaction.

Self-expressiveness of a brand
Brands are used as symbols of the self and the ideal self (Schultz et al., 1989), and consumers may connect a brand to certain individuals (Fournier, 1998).According to Kleine et al. (1995), consumers are both motivated to be part of a group and also seek a unique sense of self.The concept of self can be divided into the notions of the social and inner self (Ball and Tasaki, 1992) and the real and ideal self (Joji and Ashwin, 2012).While the social self is public and seeks the approval of reference groups, the inner self is private and pursues individual attainment.Correspondingly, the real self is how the consumer actually considers him-or herself, and the ideal self is the self he or she aspires to be.As a result of their symbolic nature, brands can be used both to support people's self-image (Matzler et al., 2011) and to help consumers construct aspirational self-images (Escalas and Bettman, 2003).For example, Batra et al. (2012) show that consumers describe their beloved objects in terms of possessing similar features and characteristics to themselves.Matzler et al. (2011) and Wallace et al. (2014) highlight how products and brands that consumers are emotionally attached to are not only in harmony with the consumers' selfperceptions, but also prompt consumers to communicate those perceptions to others.Therefore, to develop brand love, a consumer must have a feeling of psychological proximity with the particular brand (cf.Joji and Ashwin, 2012).As a consequence, time, energy and resources are more likely to be invested into beloved objects in the process of self-expansion (Park et al., 2010).Prior research (e.g.Albert and Merunka, 2013;Huber et al., 2015) also suggests that brands that encourage consumers to express themselves are the most likely to be loved.Carroll and Ahuvia (2006), as well as Wallace et al. (2014), confirm this and find a strong association between the self-expressiveness of the brand and brand love.
As brands are used to communicate the self to others (Batra et al., 2012;Matzler et al., 2011), the self-expressiveness of a brand is defined as an antecedent of brand love (Huber et al., 2015;Park et al., 2010).Consumers choose brands they believe members of their reference or aspiration group in a particular field would select.Thus, consumers verify their self-image by associating with a prototypical brand user and thereby distinguishing themselves from users of other brands.On this basis, we wish to confirm the positive effect of self-expressiveness on brand love, and thereby suggest the following: H1: A brand's self-expressiveness has a positive effect on brand love.

Brand trust
Trust is a focal factor of the formation of any long-term relationship (Morgan and Hunt, 1994), and is used as a mental shortcut to reduce uncertainty and the need for mental processing (Moorman et al., 1992).As the brand-consumer relationship is considered a substitute for human contact between the company and its customers (Matzler et al., 2008), brand trust is a necessary precondition for all long-term brand-consumer relationships (Matzler et al., 2008).
In this study, we define brand trust as consumers' affect-based experience of a particular brand (Esch et al., 2006), which leads them to rely on the performance of the brand, as exemplified by their having faith in the brand promises communicated (Albert and Merunka, 2013;Li et al., 2008).
The characteristic of brand trust as an affective experience (Esch et al., 2006) related to the expectations of the brand's honesty, altruism and reliability (Albert and Merunka, 2013) means it is gradually built by the consumers' experiences and encounters with the brand.It is considered a driver of consumers' emotional attachment to a brand (see Chauduri and Holbrook, 2001;Langner et al., 2015).Hence, strong trust in a brand leads to positive outcomes such as positive attitudes, stronger commitment and loyalty (Knox et al., 1993), and is a focal factor contributing to brand love (Albert et al., 2008;Drennan et al., 2015).It arises as a result of a consumer's overall satisfaction with a particular brand (Delgado-Ballester and Munuera-Alemán, 2001).To conclude, brand trust comprises the past and future experiences of the brand, and thus, is an antecedent of brand love (Albert and Merunka, 2013;Drennan et al., 2015).Therefore, our second hypothesis is that: H2: Brand trust has a positive effect on brand love.

Hedonic products and brand love
Consumers' purchasing and consumption behaviors are affected by both the utilitarian and hedonic aspects of value (Babin et al., 1994;Dhar and Wertenbroch, 2000).Hedonic value is provided by an affective and sensory experience of aesthetics, sensual pleasure, fantasy and fun (Hirschman and Holbrook, 1982;Joji and Ashwin, 2012).Brands that deliver largely hedonic value are considered affect-rich (Suh, 2009), and are mostly experienced subjectively (Babin et al., 1994).These are consequently evaluated more intuitively compared to brands offering utilitarian value (Suh, 2009).Hedonic value also fulfills the needs of self-expression and prestige, leading to positive emotional arousal, positive feelings and affection for the brand (Bhat and Reddy, 1998;Huber et al., 2015).Thus, consumers are more strongly attached to brands that offer hedonic value than to more functional-value-oriented brands (Lastovicka and Sirianni, 2011;Thomson et al., 2005).
According to Batra et al. (2012), brand love compounds brand-related cognitions and emotions, and the value the brand delivers is an important antecedent of brand love.In prior research, hedonic value in particular has been found to strongly affect brand love (Chaudhuri and Holbrook, 2001).For example, the findings of Carroll and Ahuvia (2006) and Huber et al. (2015) suggest a positive association between hedonic value and brand love.Thus, hedonic value is positively related to consumer propensity to form brand love (Babin et al., 1994;Carroll and Ahuvia, 2006;Huber et al., 2015).Against this backdrop, we further suggest that: H3: The hedonic value of a product has a positive effect on brand love.

WOM
WOM can be understood as any positive or negative statement made by potential, actual or former customers about a product or a company, which is made available through offline or online channels (Hennig-Thurau et al., 2004).Although the essence of WOM behavior is consistent whether based offline or online, differences exist.eWOM mostly comprises written communication (whereas offline WOM is mostly spoken), and may occur in one-to-one, oneto-many or many-to-many forms (offline WOM mostly occurs in one-to-one form).eWOM is more exposed to a ripple effect, and may be more easily controlled by companies compared to offline WOM (e.g.Huang et al., 2011).In addition, Lastovicka and Sirianni (2011) suggest that consumers with beloved objects may be selective in their offline WOM behavior and engage in WOM only in a certain context, or only with selected persons.Therefore, it is logical to distinguish between offline WOM and eWOM, although both are conceptualized as consequences of brand love (Carroll and Ahuvia, 2006;Batra et al., 2012).
In connection with WOM, speaking about a brand to others is considered an important part of an individual's identity construction (Holt, 1997).It is motivated by the intention to do the receiver a favor (Steffes and Burgee, 2009), and also meets people's social needs by putting them in contact with others (Sheth and Parvatiyar, 1995).According to Chung and Darke (2006), WOM is more likely to be generated from self-relevant and hedonic products than utilitarian products, because WOM concerning self-relevant products serves as a means of self-presentation and provides social benefits.In addition, Kudeshia et al. (2016) show that a strong positive relationship exists between brand love and WOM.Therefore, brand love is considered to directly predict positive WOM (Carroll and Ahuvia, 2006;Batra et al., 2012;Fetscherin, 2014) and also to act as a mediator between self-expressiveness and WOM (Wallace et al., 2014) and hedonic value and WOM (Carroll and Ahuvia, 2006).Thus, we posit that: H4: Brand love has a positive effect on positive WOM.H5: Brand love has a positive effect on positive eWOM.

Moderating effects of brand experience and price perception
Previous studies show that a long period of time and extensive experience of a brand is needed for brand love to develop (Holmes, 2000;Mikulincer et al., 2001), whereas brand attitude does not necessarily require such a long relationship (Fredorikhin and Thomson, 2008;Park et al., 2010), or even direct contact or personal experience with the brand.For example, Verhoef et al. (2002) show that the length of brand experience positively affects several aspects of relationships, such as satisfaction, trust and commitment.Therefore, all brand relationships change over time because of changes in how individuals process brand-related information, view the brand and show their commitment to it (Ranaweera and Menon, 2013).
The moderating effect of relationship length on the effect of brand love on WOM has not been empirically tested.However, prior evidence suggests a significant moderating effect of relationship length on the relationship between self-expressiveness and brand love, as well as between hedonic value and brand love (Huber et al., 2015).In addition, Ranaweera and Menon (2013) find a significant moderating effect of relationship length on the effect of satisfaction on WOM.Based on the previous evidence, we expect that brand experience strengthens the effect of brand love on WOM behavior.Thus, our next hypotheses are as follows: H6a: Experience strengthens the positive relationship between brand love and offline WOM, such that when a consumer has more experience with the brand, the relationship between brand love and offline WOM is stronger.
H6b: Experience strengthens the positive relationship between brand love and eWOM, such that when a consumer has more experience with the brand, the relationship between brand love and eWOM is stronger.
Brand loyalty results from a company's ability to provide its customers with more value compared to its competitors.Value is understood here as a perception of what is received (i.e.quality) and what is given (i.e.price) (Zeithaml, 1988).In this equation, price has a dual role: first, it is a measure of sacrifice, and second, an indicator of quality.Previous literature shows that prices and price judgments have a focal impact on buying behavior and loyalty formation (e.g.Hartline and Jones, 1996;Thomson et al., 2005).According to Monroe (2005), consumers judge prices relative to their internal reference prices and to the prices of similar products.As a result, they form subjective price judgments, which is interpreted on a mental price scale of inexpensive-expensive.The way customers judge prices and how these price judgments affect behavior depends on the context and buying situation (Monroe, 2005).For example, Xia et al. (2004) suggest that price affects customer behavior differently depending on whether the customer is loyal or disloyal to the brand.
Prior research shows that subjective price perception plays a central role in customer behavior, affecting value perceptions, buying decisions and post-purchase behavior (e.g.Matzler et al., 2006;Monroe, 2005;Munnukka and Järvi, 2012).This evidence also suggests that those with strong brand love hold more positive perceptions of the brand's prices compared to those of other brands.Consumers' perception of a brand's relative prices is also found to directly affect switching (Keanevey, 1995) and recommendation (Matzler et al., 2006;Varki and Colgate, 2001) 2006), we expect that customers who perceive a brand's prices more positively than those of others are more likely to be more attached to the brand and more eager to recommend it to others.Therefore, the following hypotheses are proposed: H7a: Consumer price perceptions strengthen the positive relationship between brand love and offline WOM, such that when the price of the brand is perceived as more affordable in comparison to that of other brands, the relationship between brand love and offline WOM is stronger.
H7b: Consumer price perceptions strengthen the positive relationship between brand love and eWOM, such that when the price of the brand is perceived as more affordable in comparison to that of other brands, the relationship between brand love and eWOM is stronger.

Data collection
The hypotheses were tested with data from an online survey collected in spring 2014.The convenience sampling method was used to attract respondents to the survey.This was carried out through Facebook, where a link to the survey was posted and advertised.The survey was open for one week, and 342 valid responses were obtained.
The sample is presented in Table 1.A slight majority of the respondents were female (59%) and between 18 and 25 years of age (51%).At the very beginning of the questionnaire, the respondents were asked to name any brand they were satisfied with and which they had considerable experience of.The remaining questions in the survey concerned that particular brand.The most popular brand was Apple (38 mentions; 11% of all mentions).The second most popular was Nike (N=25; 7% of all mentions) and the third most popular was Adidas (N=16; 5% of all mentions).Almost all respondents (96%) mentioned having at least two years of experience using the brand they named.

Measurement
The remaining items were measured using five-point Likert-type scales, except for the hedonic brand, which was measured on a semantic differential scale (see Table 2 for the full list of items).The four-item brand trust scale was derived from Chaudhuri and Holbrook (2001).
The scales for self-expressiveness of the brand (eight items), the hedonic value (four items), brand love (nine items) and willingness to engage in positive offline WOM (four items) and eWOM (three items) were adapted from the scales used by Carroll and Ahuvia (2006).All the model constructs are reflective in nature.The two moderators -experience and price perceptions (see Table 1) -were measured using single-item scales.Experience with the brand was measured with a scale anchored at "1 = Less than a year" to "5 = 16+ years".Price perceptions were measured using a scale adapted from Matzler et al. (2006): "Compared to other similar brands I regard this brand as affordable" (scale ranging from "1 = Strongly disagree" to "5 = Strongly agree").The survey was administered in Finnish.Therefore, all the scales were professionally back-translated to ensure conceptual equivalence (Mullen, 1995).

Measurement model
To test the hypotheses, we performed both exploratory and confirmatory factor analyses.The exploratory factor analysis identified all of the proposed six factors.At this stage, two items from the hedonic value construct and one item from brand love were removed due to their low factor loadings and cross-loadings with other factors.The hypotheses were then tested using SmartPLS 3 (Ringle et al., 2014).PLS-SEM was chosen because of our interest in predicting the key target constructs and identifying their key drivers (Hair et al., 2014, p. 19).
Table 2 presents the results of the confirmatory factor analysis.It shows that all of the factor loadings were equal to or greater than 0.649, and the Cronbach's alphas were close to or exceeded the cut-off value of 0.70 (Bagozzi and Yi, 2012).The model also passed the tests of convergent and discriminant validity (Table 3), as average variance extracted (AVE) values exceeded the cut-off of 0.50 (Table 2) and the squared AVE values were larger than the correlations shared by the respective paired constructs (Fornell and Larcker, 1981).

Place Table 2 about here
Place Table 3

about here
In the data collection phase, common method bias from the self-reported data was minimized by keeping the respondents' identities confidential and by randomizing the items in the questionnaire.In the data analysis phase, we followed Liang et al. (2007) and included a common method factor in the Partial Least Squares (PLS) model that linked to all of the converted from observed indicators (see Podsakoff et al., 2003).The results showed that the indicators' substantive variances (0.543) were significantly greater than their method variances (0.015).Thus, common method bias is unlikely to be a serious concern.

Structural model
Next, the hypotheses were tested by running the structural model (Figure 2).The model's explanatory power was high for brand love (R 2 = 0.596) and offline WOM (R 2 = 0.380), and medium for eWOM (R 2 = 0.225).H1, which suggests that the self-expressiveness of the brand has a positive effect on brand love, received support from the data (β = 0.542, p < 0.01).Trust (β = 0.303, p < 0.01) and hedonic product type (β = 0.216, p < 0.01) were also found to be positively associated with brand love, thus supporting H2 and H3. Brand love had positive associations with both offline WOM (β = 0.583, p < 0.01) and eWOM (β = 0.416, p < 0.01), supporting H4 and H5.Thus, consumers who are emotionally attached to a brand are more likely to spread positive WOM about it.Of the control variables, age had a significant positive effect on offline WOM and eWOM, in that older respondents were more active in providing positive offline WOM and eWOM, confirming that age impacts positive WOM behavior (Zhang et al., 2014).Gender had no association with WOM, which might be explained by the convenience sampling technique.

Moderating effects
To test H6 and H7, assessed the moderating effects of experience and price perceptions using the product indicator approach (Chin et al., 2003).The moderating tests were run so that only one moderator was considered concurrently.The results of the moderating effects test (Table 4) showed that two of the four hypothesized moderating effects are significant.Specifically, experience and price perceptions were found to moderate the relationship between brand love and offline WOM by strengthening the relationship, but not that between brand love and eWOM, thereby supporting H6a and H7a and leading to the rejection of H6b and H7b.These findings indicate that when a consumer has more experience of a brand, the relationship between brand love and offline WOM becomes stronger, and with less experience, the relationship weakens.With respect to price perceptions, we find that the more affordable a brand is considered to be, the stronger the relationship between brand love and offline WOM will be; and the more expensive the brand is considered to be, the weaker the relationship will be.

Discussion and conclusion
This study contributes to the literature by introducing and testing a conceptual model that shows how the self-expressiveness, trust and hedonic value of a brand drive brand love, and how brand love in turn is related to WOM in both offline and online channels.In addition, we test the effects of experience and price perceptions on the relationships between brand love and WOM and brand love and eWOM.In sum, the findings of the study show that: (1) selfexpressiveness, trust and hedonic product value are all positively associated with brand love; (2) brand love has a strong effect on both types of WOM and (3) experience and price moderate the positive relationship between brand love and WOM, such that when experience is high or a brand is considered affordable, the relationship between brand love and WOM will be stronger.The findings support our expectations that a brand's self-expressiveness, and the trust and hedonic value of the brand, all positively affect brand love.This contributes to the prior research, which has provided only limited evidence so far about the compound effect of these factors on brand love (Albert and Merunka, 2013;Drennan et al., 2015).In this respect, our findings are in line with the extant literature (Fetscherin, 2014;Huber et al., 2015;Kudeshia et al., 2016;Lastovicka and Sirianni, 2011;Vernuccio et al., 2015;Wallace et al., 2014).The results of the current research also mirror those of Caroll and Ahuvia (2006) in confirming that self-expressiveness has a stronger effect on brand love than hedonic product value does.
Peer reviews and suggestions are increasingly influencing consumers' buying decisions; consequently, marketers have made obtaining positive WOM the focal objective of marketing activity (Chen and Xie, 2008).However, prior research is very limited in showing how brand love or its antecedents affect consumers' willingness to engage in positive WOM in offline or online contexts (Batra et al., 2012;Kudeshia et al., 2016;Wallace et al., 2014).This study extends the literature by showing that brand love has a strong positive effect on both types of WOM.Although the effect (path coefficient) is stronger for offline WOM, the difference between the strength of the effects is not statistically significant (at p < 0.05).
The moderating role of experience and price perception have not been widely studied in the context of brand love and WOM, although the concept of brand love itself is conceptualized from the long-relationship perspective (Carroll and Ahuvia, 2006).Our study is among the first to confirm that the more experience a person has with a brand, the stronger the relationship between brand love and WOM will be.Similar findings have been reported concerning the effect of relationship length on the connections between self-expressiveness, hedonic value and brand love (Huber et al., 2015) and satisfaction and WOM (Ranaweera and Menon, 2013).However, the finding that experience does not have the same moderating effect on the relationship between brand love and eWOM needs further research.Similar observations were made regarding the effect of price on the relationship between brand love and offline WOM, where we found that when the subjective price perception is affordable, the link between brand love and offline WOM will be stronger.Again, we did not find this effect with respect to the relationship between brand love and eWOM.These non-significant effects might have arisen from the idea that brand love in our case was more strongly linked with offline WOM, and thus the moderating effects were only significant in these stronger relationships.

Managerial implications
The current research offers three important managerial contributions.First, managers should be aware that having a self-expressive brand is the main predictor of brand love.We recommend that brand managers create an affinity between their brands and consumers by communicating the brand's values and personality traits to nurture consumers' identification with that brand.As the self-expressiveness of a brand also includes the capability to express one's personality to others, a brand should have a clear and strong brand image and symbolic meanings in order to enable consumers to effectively express themselves to others through use of the brand.Second, consumers are increasingly utilizing Web-based communication channels -which is also reflected in their WOM behavior.This study shows that consumers communicate their brand perceptions and experience through both online and offline channels.
Despite the growth of eWOM, consumers' brand love was found to affect their traditional (offline) WOM behavior even more strongly.This could indicate that the online channel is considered a less secure place to share feelings and experiences about brands that produce strong emotions.As the eWOM context dictates that consumers cannot always know whether the people they are communicating with are trustworthy (Steffes and Burgee, 2009), people may be more inclined to engage in traditional WOM about a beloved brand when the other parties' credibility is known (Huang et al., 2011).Consequently, companies could further enhance the growth and sustainability of brand love and the promotion of positive WOM by providing secure online brand communities for devotees of their brands.Third, as our results showed, the more experience consumers have with a brand and the more affordable they perceive the brand to be, the stronger the relationship between brand love and WOM will be.This suggests that managers should take into account the length of consumers' experience with the brand and how affordable they perceive the brand to be in comparison to other brands.Managers should then use these two important variables in developing their customer relationship strategies with the aim of driving more positive WOM about the brand.For example, weighting experienced consumers in the company's marketing programs, such as by providing them with additional benefits for staying with the company, would effectively increase these consumers' propensity to engage in positive WOM behavior.In addition, marketing communications that stress the brand's beneficial price-quality ratio would induce consumers to spread positive WOM.

Limitations and future research
This research has some limitations, which also offer opportunities for future research.The first limitation concerns generalization, in that the results cannot be extended beyond the current sample frame.This is due to the relatively small sample size, which consisted solely of Finnish consumers.In addition, due to our use of the convenience sampling method, the sample may be biased towards certain types of respondents.The second limitation relates to other factors that could improve the predictive power of the proposed model.Although the model's ability to explain brand love is relatively good, there are probably additional indicators of brand love.Third, the study's non-experimental research design limits the reliability of the results concerning the causal relationships.Finally, the results are limited in that the respondents suggested and evaluated a large variety of brands and brand types that they self-selected in the study setting.
Future research should strive to gather larger samples from different market areas that are representative of consumer segments across industries.This would enable an investigation of cultural variations and individual differences.A second avenue for future research relates to improving the proposed conceptual model by including additional factors that would enhance its explanatory power.Moreover, as the characteristic of self-expressiveness seems to strongly affect brand love, it would be interesting to observe how the ability to identify not only with a specific trademark, but also with a typical user of a particular brand, affects the formation of brand love.Another interesting future study avenue would be to examine the effect of online brand communities on the conceptual model.Previous research suggests that consumers' commitment and behavior in online brand communities affects brand loyalty formation (Munnukka et al., 2015), but its effects on brand love and its antecedents lack empirically verified evidence.Future studies should also apply experimental (possibly longitudinal) designs to validate the causality of the proposed model.For example, a fertile research case for a longitudinal research design would be to examine brand love of Volkswagen owners during and after the emissions scandal of autumn 2015.
behavior.Based on the propositions of De Matos and Rossi (2008) and McKee et al. (