High-fit charitable initiatives increase hedonic consumption through guilt reduction

Purpose – Past research on cause-related marketing (CRM) suggests that these socially beneficial initiatives are more effective when linked with hedonic than utilitarian products. Little is known, however, about the process underpinning this effect. This paper aims to investigate why and under what circumstances CRM enhances the appeal of hedonic products by testing the mediation of guilt and introducing the moderating role of cause-product fit. Design/methodology/approach – The authors test a model of moderated mediation in two studies. Study 1 shows that the effectiveness of combining CRM with hedonic consumption is explained by the mediating role of feelings of guilt. Study 2 demonstrates that this mediation depends on the level of fit or congruency between the cause and the product. Findings – Results suggest that CRM campaigns offer the opportunity to improve the consumption experiences of hedonic products by reducing the feelings of guilt intrinsically connected with these options. Moreover, fit moderates the emotional processes activated by CRM initiatives. When fit is high, CRM reduces guilt and improves consumers’ experiences when purchasing hedonic alternatives. Originality/value – The study extends current understanding of how CRM can promote hedonic consumption and contributes further to research on guilt as an emotion able to promote responsible consumption decisions. Moreover, the study introduces and tests the impact of cause-product fit in predicting consumers’ ethical purchase intention. For managers of hedonic brands, the study offers important implications on how to deploy CRM campaigns to foster better customer experiences.


High-fit charitable initiatives increase hedonic consumption through guilt reduction Introduction
Cause-related marketing (CRM) virtuously activates relationships between non-profit organizations, companies and consumers in support of socially desirable goals (Kotler, et al., 2012;Lafferty et al., 2016). In the past 10 years, CRM has become the fastest-growing category of sponsorships in US and European markets (IEG, 2015). Companies across different industries have implemented campaigns linking them to various social causes, from medical research to care for children in need, or aid initiatives in developing countries (Nielsen, 2014). A transaction-based definition of CRM links the company's donation directly to customers' purchase behavior: "an offer from the firm to contribute a specified amount to The authors wish to thank Professor Silvia Grappi for her feedback on an earlier version of the paper. EJM 51,11/12 2030 a designated cause when customers engage in revenue-providing exchanges that satisfy organizational and individual objectives" (Varadarajan and Menon, 1988, p. 60). Existing research has examined the main effect of CRM programs on intentions to purchase a brand (Chang, 2008;Folse et al., 2010), actual product choice (Arora and Henderson, 2007;Chun and Cheng, 2015), consumer willingness to pay (Koschate-Fischer et al., 2012) and bidding behavior in a charity auction context (Elfenbein and McManus, 2010). Moreover, previous studies (Strahilevitz, 1999;Kim and Johnson, 2013) show that psychological factors, such as personal cause affinity (Arora and Henderson, 2007) or emotional involvement (Baghi et al., 2009), have a central role in responses to CRM, as they influence to what extent consumers engage emotionally with the link between the product and the social cause (see Lafferty et al., 2016 for a review).
An interesting issue however is whether this effect varies across different product categories. Hedonic products  can be linked with CRM successfully. Previous studies (Strahilevitz, 1999;Strahilevitz and Myers, 1998) reveal that a customer is more likely to choose a CRM product in shopping situations which satisfy more frivolous needs (Hirschman, 1980) than when purchasing a utilitarian product with a clear functional goal. Very little research however examines systematically the process underpinning this effect. Strahilevitz and Myers (1998) speculate that CRM initiatives may soothe consumers' feelings of guilt that are often implicit in a hedonic shopping context (Botti and McGill, 2011;Simonson, 2002a, 2002b). Their study however does not test this hypothesis explicitly. Following this approach, Hagtvedt and Patrick (2015) find in a luxury retailing context that an association with a charity at the point of sale can increase purchase intentions and facilitate upselling. The reduction of consumers' guilt is the underlying mechanism that improves willingness to buy luxury brands. Building on these initial findings, this paper aims to answer the following research question: RQ1. Why does CRM enhance the appeal of hedonic products?
To address this question, we examine: guilt reduction as a potential mediator of the effect of CRM associated with hedonic choices on consumer behavior; and product-cause fit as a potential boundary condition that might moderate the relative effectiveness of a CRM campaign.
The study contributes to existing debates in several ways. First, the manuscript extends work by Hagtvedt and Patrick (2015), developing and testing the same mediating process in the domain of hedonic, non-luxury goods. While Hagtvedt and Patrick (2015) examine only the role of guilt in very expensive luxury shopping, this investigation focuses on non-luxury hedonic products that offer a more limited potential for guilt reduction. Examining whether the mediating role of guilt is supported in these more common hedonic consumption situations has significant managerial implications because it widens the applicability of the theory to a vast range of contexts. Second, this research conceptualizes and tests the role that perceived fit, defined as the level of perceived similarity between the product and a charitable cause (Lafferty, 2007;Pracejus and Olsen, 2004), plays in consumers' emotional and behavioral reactions to CRM. Results show that CRM-product fit moderates the feelings of guilt elicited by the purchase of a hedonic (versus a utilitarian) product. Consequently, the present findings extend the theory advanced by Hagtvedt and Patrick (2015), documenting that CRM reduces guilt and enhances the appeal of hedonic products only under high fit conditions.

Hedonic consumption through guilt reduction
The paper is organized as follows. First, we present the theoretical background and develop the hypotheses that underpin our research. Second, we present evidence from two studies that test the proposed mechanisms explaining consumers' reactions to CRM in a hedonic versus utilitarian purchase context. Study 1 demonstrates the mediating role of guilt in explaining consumers' willingness to buy a hedonic product, and Study 2 replicates this finding while also testing the moderation of fit in a CRM context. Finally, we conclude with a discussion of our research findings and the implications for theory and practice.

Research background
Hedonic vs utilitarian consumption and cause-related marketing Products are often classified as hedonic or utilitarian, a categorization that summarizes a host of related concepts . While hedonic experiences are fun , because of their inherent pleasurable nature, they often elicit moral guilt and are linked to vices and frivolousness (Botti and McGill, 2011). Hedonic consumption, comprising multisensory and emotive aspects of consumption experiences Holbrook and Hirschman, 1982), is not always related to expensive or luxury products Simonson, 2002a, 2002b), as accessible goods can also be hedonistic (e.g. candies, alcoholic drinks, chocolate). Utilitarian products, on the other hand, are functional and useful. As a result, utilitarian experiences are easy to justify because they are associated with necessities (Dhar and Wertenbroch, 2000;Simonson, 2002a, 2002b;Okada, 2005). Accordingly, attitudes and behaviors toward hedonic products are more likely to be determined by emotional reactions (Pham, 1998). This intrinsic emotionality (Havlena and Holbrook, 1986;Holbrook and Hirschman, 1982;Mano and Oliver, 1993) can sometimes be characterized by a strong ambivalence (Otnes et al., 1997;Penz and Hogg, 2011). As it is difficult to justify spending on pleasures, hedonic consumption generates guilt Simonson, 2002a, 2002b) and induces emotional ambivalence. Prelec and Loewenstein (1998) point out that when consumers make purchases, they often experience an immediate pain of paying, which can weaken the pleasure derived from consumption. Such pain of paying can lead to guilt in case of weak or flimsy justifications. The pain of paying might also be more acute when spending on enjoyable or hedonic goods than on utilitarian or functional ones because the former are often difficult to justify and by definition not essential (Prelec and Loewenstein, 1998). When buying hedonic products, therefore, consumers experience positive and negative emotions simultaneously (e.g. pleasure and guilt, happiness and sadness). This proposition is supported by the reason-based choice conception, which seeks to explain consumer preferences based on reasons that are constructed to justify purchase decisions (Shafir et al., 1993, Botti andMcGill, 2011). On the basis of this evidence, it is expected that guilt will be stronger in hedonic than in utilitarian purchases: H1. The nature of the product (hedonic vs utilitarian) has a main effect on anticipated guilt in a purchase context.
When purchasing hedonic goods, people look for information in the decision context that will allow them to rationalize the purchase and reduce the negative affect experienced (Okada, 2005). Psychological research suggests that hedonic choices have an internal perceived locus of causality because inherently satisfactory (DeCharms 1968;Deci and Richard, 1985). Accordingly, CRM can offer consumers an avenue that facilitates rationalization and alleviates the emotional ambivalence (Nordgren et al., 2006) linked with hedonic consumption by reducing feelings of guilt. In the so-called EJM 51,11/12 transactional program (Varadarajan and Menon, 1988), for every unit of product sold, the corporation donates a share of the proceeds (or a percentage of the price) to a particular social cause. This kind of program allows consumers to buy a product and donate to charity at the same time, as a share of the revenue is directed in favor of a social cause. In this context, CRM should facilitate the immediate justification of a hedonic purchase and therefore soothe consumer guilt, facilitating the decision to buy a hedonic product. In the case of a utilitarian purchase experience, CRM might not act as a justification because people feel less guilty about buying a utilitarian product. Consequently, when CRM is combined with utilitarian products it will not contribute to make consumers feel better about their necessary purchase (Baghi et al., 2010). Strahilevitz and Myers (1998) demonstrate that charity incentives work better for hedonic products (e.g. chocolate truffles, theme park tickets) rather than for practical products (e.g. pocket dictionaries, correction fluid). They propose that this effect is due to the fact that "the altruistic utility offered by charity incentives may be more complementary with the feelings generated from frivolous products than with the more functional motivations associated with practical products" (p. 444). However, the authors do not explore or test the process that leads to CRM effectiveness in the domain of hedonic consumption. Nonetheless, they suggest that the cause-related hedonic product may offer a valid justification that relieves consumer guilt, and this mediating mechanism could explain the effect identified (Botti and McGill, 2011). Recent work on luxury shopping seems to support this hypothesis, showing that an association with a cause-related campaign can increase decisions to buy expensive shopping items (Hagtvedt and Patrick, 2015). Given this premise, it is expected that combining CRM with a non-luxury hedonic alternative, rather than a utilitarian one, will have a positive impact on consumers' purchase intentions: H2. Combining a hedonic product (versus a utilitarian product) with a CRM initiative increases consumers' willingness to buy.
The mediating mechanism: relieving feelings of guilt Guilt is an unpleasant emotion experienced when individuals feel that they have not lived up to their self-image (Tangney et al., 2007) or they see themselves as responsible for a negative outcome (Roseman et al., 1996). Research shows that behaving in support of social causes diminishes guilt. For example, donation to charitable causes (Basil et al., 2008) or engagement in responsible consumption choices (Antonetti and Maklan, 2014) are often motivated by feelings of anticipated guilt and implementing such behaviors reduces the level of guilt experienced by consumers. This evidence supports the speculation first suggested by Strahilevitz and Myers (1998) and documented in a luxury retail context by Hagtvedt and Patrick (2015). A CRM initiative counterbalances at least some of the guilt implicit in hedonic consumption. Consequently, guilt acts as a mediator of the effects of CRM on purchase intentions in a hedonic consumption context. In marketing and consumer behavior literature, there is significant evidence that the desire to assuage feelings of guilt can be, in certain circumstances, a strong driver of consumer choices. It is possible that hedonic consumption might increase levels of anticipated guilt because it is perceived as a frivolous decision and a loss of self-control (Kivetz and Keinan, 2006;Steenhaut and Van Kenhove, 2006;Mishra and Mishra, 2011). There is, for example, significant evidence that buying luxury goods generates guilt Simonson, 2002a, 2002b). Guilt is often caused by behaviors that imply an inability to regulate one's own decisions including consumption of comfort food (Mohr et al., 2012) or Hedonic consumption through guilt reduction binge drinking (Agrawal and Duhachek, 2010). In general, when consumers are presented with the possibility of engaging with these behaviors, they might anticipate feelings of guilt, and this would influence their ultimate choice (Steenhaut and Van Kenhove, 2006;Baumeister et al., 2007). In a hedonic context, anticipated guilt therefore acts as a negative influence on consumers' decisions. To cope with the negative feelings of guilt, consumers will be more likely to reduce their consumption of hedonic products. On the basis of these arguments, it is expected that CRM initiatives will increase willingness to buy hedonic products through a reduction of the guilt associated with these products: H3. The inclusion of a CRM initiative reduces consumers' feelings of guilt associated with the purchase of a hedonic product.
H4. Guilt has a negative influence on consumers' decisions to purchase a hedonic product. Guilt reduction therefore mediates the effect of CRM on hedonic purchase intention.
The perceived fit of cause-related marketing campaigns One of the main drivers of CRM effectiveness is the selection of charitable partners that provide a good fit with the firm (Pracejus and Olsen, 2004;Simmons and Becker-Olsen, 2006). Fit has often been considered in the marketing literature as "perceived similarity of the brand and other entity" (Keller, 2003, p. 599). Research on fit developed from the study of brand extensions, which considers customers' perceived fit between the brand and the new product or offering (Aaker and Keller, 1990). In a CRM context, fit evaluates the perceived similarity and consistency between a cause aim and a brand or product. Varadarajan and Menon (1988) described fit in CRM alliances as the perceived link between the company's image, positioning and target market and the cause's image and constituency (Varadarajan and Menon, 1988;Ellen et al., 2006). Trimble and Rifon (2006) differentiate between image-based fit and functional fit. Imagebased fit is the perceived similarity of positioning adopted by the for-profit brand and the social cause (e.g. Avon and a charity promoting research on breast cancer would be expected to have a high fit), whereas functional fit is the contextual congruence of product features and the mission of the social cause (e.g. a medical check-up and support for medical research would be expected to have high fit). In the present study, the focus is on functional fit between the product features and the mission of the social cause supported (Bigné et al., 2012). Fit is a crucial success factor in a CRM campaign. A wide set of contributions have showed a positive effect of cause-product/brand fit in terms of perceived brand image (Gwinner and Eaton, 1999), brand credibility (Rifon et al., 2004;Becker-Olsen et al., 2006b), firm's altruistic attributions (Rifon et al., 2004;Ellen et al., 2006), product purchase intentions (Becker-Olsen et al., 2006a;Gupta and Pirsch, 2006) and consumer choice (Pracejus and Olsen, 2004). Similarly, Becker-Olsen et al. (2006a) showed that low-fit corporate social responsibility initiatives impact consumer attitude and purchase intentions negatively irrespective of the firm's motivations, and that high-fit social responsibility actions lead to an improvement in consumer attitudes toward the firm and the brand while also raising purchase intentions.
There are also results that suggest product-cause fit can have problematic effects. For instance, Yoon et al. (2006) found that the fit between the company and the CSR domain leads to an increase in consumers' skepticism toward the corporation's motives (Yoon et al., 2006). Similarly, Foreh and Grier (2003) suggested that when a company is closely related to the CSR domain it supports, it is likely that the profits the firm might generate from the CSR EJM 51,11/12 activity will lead consumers to questioning the motives of the company (Foreh and Grier, 2003). On the whole, however, the literature indicates that product-cause fit is likely to increase the persuasiveness of a CRM campaign.
In cases of high cause-product/brand compatibility, the cause-related information (e.g. the nature of the cause and its mission) is more accessible and easy to remember, and it will therefore have a stronger impact on purchase intentions (Fazio et al., 1989). Product/brandcause compatibility influences consumer choice through associative learning (Shimp et al., 1991;Till and Nowak, 2000), and perceived belongingness: a close link between two stimuli (e.g. product functionality and the cause mission) should make the development of an associative link easier (McSweeney and Bierley, 1984;Till and Nowak, 2000). This is consistent with information integration theory (Anderson, 1981), which suggests that prior attitudes will be integrated with the new congruent information influencing positive evaluation and response. Therefore, greater fit between the product and cause in the CRM alliance should lead to a more favorable consumer's response through the integration process of positive and congruent associations.
On the other hand, when there is a disconnection between CRM and product features, it is more likely that the CRM message will be perceived as manipulative (Drumwright, 1996) and will likely decrease consumers' willingness to buy (Bignè-Alcaniz et al., 2012). Thus, a good fit between a firm and a social initiative can be more easily integrated into a consumer's existing cognitive structure, strengthening the connection between the firm and the social initiative (Rumelhart et al., 1986;Fiske and Taylor, 1991). Alternatively, a low-fit initiative is likely to be perceived as not congruent with prior expectations making it more difficult to integrate new knowledge into existing memory structures (Folkes and Kamins, 1999). Consistent with this stream of previous research, the assumption is that fit moderates the positive impact that companies can achieve from linking a CRM campaign with a hedonic rather than a utilitarian product: H5. The willingness to buy a hedonic (rather than a utilitarian) product associated with a CRM campaign is stronger (weaker) under conditions of high (low) fit between the product and the CRM campaign.
The fit between CRM and the product should also regulate consumers' emotional reactions. The elicitation of feelings of guilt after exposure to a CRM campaign will depend on the persuasiveness of the message. Research shows that consumers resist messages they perceive as manipulative, and in such a situation, they are less likely to experience guilt (Coulter and Pinto, 1995). As past CRM research demonstrates that campaigns with high fit are more persuasive (Becker-Olsen et al., 2006a;Gupta and Pirsch, 2006), it is also reasonable to expect that in this context a high-fit CRM campaign associated with a hedonic product will decrease feelings of guilt more strongly. Associative learning principles (McSweeney and Bierley, 1984;Shimp, et al., 1991) support the idea that a positive affective transfer between associated entities will be stronger when the two objects are closely linked. According to this perspective, pairing a pleasant and congruent stimulus (e.g. a social mission) with a product or a brand can affect overall attitude and emotional response toward the product or the brand (Gupta and Pirsch, 2006). From this point of view, the CRM campaign will be more likely to influence the consumption experience positively when it is closely associated with the product purchased (Till and Nowak, 2000;Lafferty, 2009). The positive affective transfer will thus reduce the ambivalence of a hedonic purchase (Botti and McGill, 2011). In addition, research on the persuasiveness of guilt appeals has demonstrated that increased processing fluency (i.e. ease of processing) leads to stronger emotional reactions Hedonic consumption through guilt reduction (Duhachek et al., 2012). As greater fit is a typical driver of ease of processing, the expectation is that high fit will strengthen the impact of CRM, when linked to hedonic consumption, on guilt. At the same time, CRM campaigns linked with utilitarian products might not be as easy to process and hence will have a weaker effect in terms of guilt reduction. Guilt is not central to the utilitarian purchasing experience and consequently the guilt reducing effect of CRM is less pronounced (Botti and McGill, 2011;Okada, 2005). The ability of a CRM initiative to alleviate guilt should therefore be stronger in hedonic purchase experiences (vs utilitarian ones) and when consumers perceive a high level of fit between CRM and the product features. This effect should further translate in positive intentions to purchase a hedonic product: H6. Consumers will experience stronger (weaker) feelings of guilt under condition of low (high) fit between the hedonic product (rather than utilitarian product) and the CRM campaign.
Figure 1 summarizes the hypotheses tested in this research. The aim is to test the combination of CRM with hedonic products in two different ways. Study 1 compares reactions to hedonic (vs utilitarian) products with CRM and without CRM. This allows testing H2, H3 and H4 (H1 is also tested in this study through the inclusion of a condition where consumers assess utilitarian products without CRM). In Study 2, the focus is on examining CRM combined with hedonic versus utilitarian offerings. This allows testing all research hypotheses and also probing H2, H3 and H4 from a different perspective. The theorizing does not simply suggest that hedonic products can be promoted effectively through CRM initiatives but that these campaigns are relatively ineffective when linked to utilitarian compared to hedonic products. The second study tests specifically this prediction and allows assessing the role of fit, in terms of perceived link between the functional features of the product and the social aim of the cause, as a moderating variable in the mediating mechanism.

Study 1 Participants and procedures
In total, 82 students at an Italian public university were invited to participate in the research. Only two students in the class refused to participate, leading to a sample of 80 participants (56 per cent females, 96.3 per cent were aged between 18 and 24 years and 3.7 per cent aged between 25 and 34 years). In total, 79 participants completed the entire questionnaire and were retained for the analysis. The choice of a student sample was consistent with previous research on CRM (Baghi et al., 2009;Chun and Cheng, 2015; Figure 1.  Hagtvedt and Patrick, 2015). The first author recruited participants among attendees of a professional training course during class hours. Participants were asked to answer to a survey about their shopping habits. The questionnaire took approximately 10 min to complete. After finishing, respondents were debriefed about the purpose of the study and thanked. The participants were randomly assigned to one of two experimental conditions in a 2 (CRM vs no CRM) Â 2 (product: hedonic vs utilitarian) experiment with the last factor within subjects, leading to a total of 158 evaluations that are used for the analysis. Each participant saw two products (one hedonic and one utilitarian).

Conceptual model
Participants were presented with a hypothetical scenario in which they were asked to imagine being in a purchase situation; they were presented with a written description of the main features (e.g. price, quantity) of two products (half of the participants evaluated the hedonic product first and half evaluated the utilitarian product first). Products used in the scenarios were selected to be accessible (not too expensive) and attractive for students. In the treatment condition, the products were linked to a transaction-based CRM activity in which a percentage of the price was donated to a medical research charity (Strahilevitz and Myers, 1998;Pracejus and Olsen, 2004;Baghi et al., 2009; e.g. coupon of three massages e70.00, 5 per cent of the price will be donated to support medical research). In the control condition, no CRM information was provided. Prices were in line with the Italian marketplace (Table I shows how the information was presented to participants). Each participant saw only two of the four products (one hedonic and one utilitarian).

Materials
A pre-test was conducted to select the hedonic and utilitarian products. A separate sample of 50 students from the same university evaluated, on two 7-point Likert scales, to what extent they perceived seven products as useful and functional or hedonic and pleasure seeking (Botti and McGill, 2011). The most hedonic products were a pair of sunglasses (M hedonic = 6.08, M utilitarian = 3.12; t (1, 48) = 7.97 p < 0.001) and a coupon for three massages (M hedonic = 6.43, M utilitarian = 2.03; t (1, 48) = 12.37 p < 0.001). The most useful products were a laser printer (M hedonic = 2.08, M utilitarian = 6.12; t (1, 48) = 23.22 p < 0.001) and a train transit pass (M hedonic = 3.01, M utilitarian = 6.27; t (1, 48) = 12.22 p < 0.001). Consequently, four products were selected for the main study. Coupon of 3 massage, e70 5% of the price will be donated to support medical research Coupon of 3 massage, e70 and Train transit pass, e70 5% of the price will be donated to support medical research and Train transit pass, e70 Hedonic and utilitarian products Sunglasses, e120 5% of the price will be donated to support medical research Sunglasses, e120 and and Laser printer, e100 5% of the price will be donated to support medical research Laser printer, e100

Measures
In both conditions, participants were asked to rate the level of anticipated guilt experienced (Dahl et al., 2005) if they had decided to buy the products presented (three items on a sevenpoint Likert scale: e.g. "not guilt at all/a lot of guilt"; Dahl et al., 2005). They also evaluated their purchase intentions of the products (three items on a seven-point Likert scale; Robinson et al., 2012). Participants also rated the hedonic versus utilitarian nature of the products on six items (three items for the hedonic nature: e.g. "not enjoyable/enjoyable" and three items for the utilitarian nature: e.g. "not necessary/necessary"; short version of the scale validated by Voss et al. (2003) and used in a previous study by Baghi et al. 2010). Both the wording of the items and the rating scale adopted are consistent with previous studies on CRM (Baghi et al., 2009;Bignè-Alcaniz et al., 2012;Hagtvedt and Patrick, 2015). All scales were randomized to avoid order effects.  (Hair et al., 2011). Discriminant validity was supported as the Fornell-Larcker criterion was respected for all constructs (Fornell and Larcker, 1981) and loadings were always higher than crossloadings (Hair et al., 2011).
Guilt: To test H1 and H3, a repeated measures ANOVA using product typology (hedonic vs utilitarian) as within subject factor and CRM as between subjects' factor was performed. Results showed a main effect of product typology on consumers' guilt (hedonic products without CRM M guilt = 5.76 vs utilitarian products without CRM M guilt = 2.32; F(1, 77) = 31.28; p < 0.001). Hence H1 is supported by the data. The same analysis showed a main effect of the CRM between subject condition on consumers' guilt in hedonic purchases (hedonic product with CRM M guilt = 4.12 vs hedonic products without CRM M guilt = 5.76; F(1, 77) = 8.57; p < 0.001). H3 is supported.
Willingness to buy: A repeated measures ANOVA using product typology (hedonic vs utilitarian) as within subject factor and CRM as between subjects factor showed evidence of a main effect of product typology on consumers' purchase intentions (hedonic products M wtb = 5.11 vs utilitarian products M wtb = 5.53; F(1, 77) = 21.49; p < 0.001) and a main effect of CRM between subjects condition on hedonic consumers' purchase intentions (hedonic products with CRM M wtb = 5.75 vs hedonic product without CRM M wtb = 4.47; F(1,77) = 2.55; p < 0.001). H2 appears also supported by the data. Descriptive statistics are available in Table II. Correlations among the constructs were, as expected, negative and significant. Consumers' willingness to buy had a significant relationship with guilt (M wtb = 6.39; M guilt = 3.32; r = À0.242; p < 0.001).
To test the mediation model proposed, a mediation analysis using PROCESS Model 4 (Hayes, 2012;Muller et al., 2005) with 10,000 bootstrap estimation resamples and 95 per cent confidence intervals was conducted to verify the expectation that guilt reduces willingness to buy (H4), CRM associated with hedonic products reduces consumers' guilt (H3) and improves purchase intention (H2). The independent variable (CRM condition comparison) was coded À1 (hedonic product/no CRM) and þ1 (hedonic product/CRM). This analysis was therefore conducted on a sample of 79 evaluations. The mean of all items for each construct was used in the analysis. The indirect effect of the presence of CRM linked to a hedonic product on purchase intentions was mediated by guilt reduction (indirect effect = 0.23, CI from = 0.01 to = 0.49). All the path coefficients in the model were consistent with the hypotheses. The effect of CRM on hedonic purchase intentions was fully mediated by the consumers' sense of guilt (Table III).

Discussion
The evidence presented supports the expectation that by linking a CRM cause with a hedonic product companies can potentially increase the appeal of their offerings. This effect is explained by CRM's ability to assuage feelings of anticipated guilt often linked with hedonic consumption. When compared to recent findings by Hagtvedt and Patrick (2015), results indicate that the applicability of CRM as a marketing technique could be extended significantly to various areas of non-expensive hedonic consumption and that the mediating role of guilt is also supported for products which are expected to be less guilt inducing than luxury goods Simonson, 2002a, 2002b).
Study 2 replicated these findings on different products and tested the role of fit as a moderating variable of guilt reduction. This moderated mediation mechanism was examined comparing hedonic versus utilitarian CRM products.

Study 2 Participants and procedures
Study 2 examines all the research hypotheses summarized in Figure 1. In total, 190 students at an Italian public university were invited to participate in the research. Only five students refused to participate, leading to 185 participants in total (74 per cent Hedonic consumption through guilt reduction females, 93.3 per cent were aged between 18 and 24 years and 6.7 per cent aged between 25 and 34 years). In total, 182 of them completed the questionnaire and were retained for the analysis. The choice of a student sample was consistent with previous studies in CRM field (Baghi et al., 2009;Chun and Cheng, 2015;Hagtvedt and Patrick, 2015). The first author recruited participants among attendees of two professional training courses during class hours. Participants were asked to answer to a survey about their shopping habits. The questionnaire took approximately 10 min to complete. After finishing, respondents were debriefed about the purpose of the study and thanked. Participants were randomly assigned to one of six experimental conditions in a 2 (products: hedonic vs utilitarian) Â 3 (CRM: fit vs no fit vs no CRM) between-subjects experiment. Each participant saw two products (both either hedonic or utilitarian) in one CRM condition (Table IV shows the information presented to participants). As in Study 1, participants were presented with a hypothetical scenario in which they were asked to imagine being in a plausible purchase situation; they were presented with a paper description of the principal features (e.g. price, quantity) of two products. We chose products that were relatively affordable and attractive for students [1]. Each participant evaluated either two hedonic products or two utilitarian products for a total of 364 evaluations used in this analysis [2]. The design also allowed collecting observations for both expensive and more accessible hedonic products. All other procedures are consistent with Study 1.

Hedonic products
Coupon of 3 professional relaxing massages, e100 5% of the price will be devolved to support medical research Coupon of 3 professional relaxing massages, e100 5% of the price will be devolved to support the victims of environmental disasters Coupon of 3 professional relaxing massages, e100 and and and Ice-cream sundae 250 g, e5 5% of the price will be devolved to support the food bank donation Ice-cream sundae 250 g, e5 5% of the price will be devolved to support the victims of environmental disasters Ice-cream sundae 250 g, e5 Utilitarian products Complete dental check-up, e100 5% of the price will be devolved to support medical research Complete dental Check-up, e100 5% of the price will be devolved to support the victims of environmental disasters Complete dental Checkup, e100 and and and Paper napkins package, 1000 pieces, e3 5% of the price will be devolved to support the food bank donation Paper napkins package, 1000 pieces, e3 5% of the price will be devolved to support the victims of environmental disasters Paper napkins package, 1000 pieces, e3 EJM 51,11/12 Materials A pre-test was used to identify the relevant products and social causes. An independent sample of 25 students from the same university evaluated, on a seven-point Likert scale, several potential cause-related offerings in terms of: perceived product/cause fit (measured on four items and on a seven-point Likert scale: e.g. "not coherent/coherent" Bignè-Alcaniz et al., 2012); perceived functional or hedonic nature of the product (same items used in Study 1) and perceived importance and urgency of the cause (two items; "not important/not urgent"; "important/urgent"; Robinson et al., 2012). The combination presented in Table IV was selected through the pre-test. The pre-test allowed selecting campaigns that did not vary significantly in terms of the perceived urgency and importance of the causes. We control for urgency and importance because discrepancy in participants' perceptions on these variables might influence their reaction to the CRM messages independently of the variables considered in our model Rifon, 2006, Lafferty, 2009).

Measures
The scales of Study 1 for guilt and willingness to buy are retained in this study. As manipulation checks, participants rated the perceived fit between the products and the social causes, the functional and hedonic nature of the products and the importance and urgency of the campaigns (same measures described above). In the control condition, participants rated their intention to buy the products presented without CRM and the functional and hedonic nature of the products. To rule out potential alternative explanations, in Study 2, we included three additional scales. First, we used four items to measure anticipated feelings of disgust, worry, sadness and anxiety (measured on a scale from 1 = not at all to 7 = extremely). Second, four items were used to measure respondents' social desirability (Grappi et al., 2013). Finally, five items assessed individual skepticism toward CRM (Mohr et al., 1998). These variables were used as covariates in the analysis as they might influence participants' reactions to CRM independently of our theory. We tested the conceptual model with and without these covariates to rule out potential alternative hypotheses. The presentation of the scales was randomized.

Findings
Manipulation checks: Participants rated the hedonic and utilitarian products in line with the expectations (e.g. coupon for three massages: M hedonic = 5.73 vs M utilitarian 4.43; t (1, 88) = 11.24; p < 0.001). The level of fit between the cause and the product was also manipulated successfully. For example, the coupon for three massages (M fit = 3.32) and the dental checkup (M fit = 4.28) both were judged as having higher fit with medical research than with the charity in support of the victims of an environmental disaster (M fit = 2.54; t (1, 59) = 6.09; p < 0.001 and M fit = 3.00; t (1, 60) = 19.48 p < 0.001 respectively).

Hedonic consumption through guilt reduction
Guilt: To test H1 and H3, a MANOVA was performed. Results showed a main effect of the product nature (hedonic vs utilitarian) on guilt (hedonic products without CRM M guilt = 3.63 vs utilitarian products without CRM M guilt = 2.94; F(1, 118) = 7.89; p < 0.001). The same analysis showed a main effect of CRM condition on consumers' anticipated guilt in hedonic purchases (hedonic products without CRM M guilt = 3.60 vs hedonic products with CRM M guilt = 2.30; F(1, 180) = 5.744; p < 0.001). H3 is supported by the data. Results showed also an interaction effect between CRM and fit conditions on the same dependent variable (F(1, 240) = 14.929; p < 0.001).
Willingness to buy: Results showed a main effect of CRM condition on consumer hedonic purchase intentions (hedonic products without CRM M wtb = 3.30 vs hedonic products with CRM M wtb = 4.47 (F(1, 179) = 22.894; p < 0.001). H2 is supported. A main effect of fit on willingness to buy (high fit hedonic products M wtb = 5.14 vs low fit hedonic products M wtb = 3.84 and high fit utilitarian products M wtb = 4.29 vs low fit utilitarian products M wtb = 4.25; F(1, 240) = 12.342; p < 0.001) was found. Finally, results showed an interaction effect of product typology and fit conditions on the same dependent variable (F(1, 240) = 13.123; p < 0.001). In fact, under conditions of high product fit, the willingness to buy hedonic products (vs utilitarian products) increases significantly (from M wtb = 4.29 to M wtb = 5.14). On the other hand, the willingness to buy of hedonic and utilitarian products is comparable under conditions of low fit (M wtb = 4.25 for utilitarian products and M wtb = 3.84 for hedonic products). This evidence supports H5. Descriptive statistics are available in Table V (60 product evaluations for each group).
Correlations among all dependent variables were significant and consistent with expectations. Table VI shows that anticipated guilt correlates negatively with intentions to buy the product. The covariates (skepticism and social desirability) appeared to be unrelated to the conceptual variables. To replicate the results of Study 1, a mediation model first comparing the influence of CRM condition on guilt and willingness to buy was run. The independent variable was coded consistently with Study 1 (À1: hedonic product/no CRM and þ1: hedonic product/CRM) and all other procedures are retained. The model was calculated on the basis of 180 evaluations, and the findings are presented in Table VII. All results were consistent with the hypotheses. The introduction of CRM reduced guilt significantly (supporting H3) and affected willingness to buy positively (consistent with H4). Evidence of a positive indirect effect (indirect effect = 0.18, CI from = 0.05 to = 0.34) supported the mediation suggested by the research model. To test the model of moderated mediation proposed, a conditional effects analysis using PROCESS (Muller et al., 2005;Hayes, 2012) was conducted. Hypotheses were examined looking at the difference in reactions to CRM when this is linked to hedonic versus utilitarian products. A model of moderated mediation is estimated using the same procedures described above (PROCESS, Model 8). The independent variable (product type comparison) was coded À1 (utilitarian product with CRM) and þ1 (hedonic product with CRM). Similarly, the low-fit condition was coded À1, while the high-fit condition was coded 1. On the whole, this analysis was conducted on 240 evaluations provided by 120 participants who were presented with the treatment conditions. The details of the paths estimated are presented in Table VIII. Consistent with H3, CRM reduced the feelings of guilt associated with hedonic consumption. H4 is supported as higher guilt reduced willingness to buy in line with the findings of Study 1. When the model was tested in its entirety, the direct path from Product type condition and willingness to buy was not statistically significant, suggesting full mediation. The effect of CRM on willingness to buy was moderated by the Notes: **indicates p < 0.01; ***indicates p < 0.000 Notes: *indicates p < 0.05; **indicates p < 0.001 Hedonic consumption through guilt reduction level of fit. When fit was low, CRM did not impact the intention to purchase a hedonic product (effect = À0.19, CI from À0.43 to 0.04). However, when fit was high CRM influenced willingness to buy (effect = 0.31, CI from 0.06 to 0.56). This result was consistent with the descriptive statistics observed above and supports H5. Furthermore, the link between hedonic products and CRM initiatives had also a positive indirect effect on willingness to buy mediated by guilt. The indirect effect showed a positive influence when fit was high (effect = 0.12, CI from 0.05 to 0.22), but no significant impact when fit was low (effect = À0.19, CI from À0.43 to 0.04). This analysis supports H6. Figures 2 and 3 offer a graphical representation of the interactions postulated in H5 and H6, respectively. Figure 2 shows how linking a CRM initiative with a hedonic product reduces the anticipated guilt from the purchase in cases of high fit between products and the campaign. It is interesting to notice that fit does not have the same impact on utilitarian products because these alternatives are less closely connected with feelings of guilt. Figure 3 demonstrates that high fit between the campaign and the product rewards hedonic alternatives disproportionately because when considering these types of purchases consumers search more actively for information that might justify their purchases.
To rule out potential alternative explanations, we estimated our model again including several covariates. First, the model was calculated again including consumer skepticism toward the CRM campaign and social desirability as controls. We also controlled for the effect of two socio-demographic characteristics (i.e. age and gender). The logic here was that,  as CRM presents a socially responsible initiative, consumers' reactions might be biased by a tendency to answer in a socially desirable way (Randall and Fernandes, 1991;Grappi et al., 2013). Equally, perceptions of scepticism might lead participants to discount CRM initiatives, and this could influence the model (Vanhamme and Grobben, 2009). All results were robust to the introduction of these covariates in the analysis. Neither social desirability nor consumer skepticism had any significant effect on any of the variables included in the model. Age and gender were also not associated significantly with any of our conceptual variables. Finally, the analysis was also estimated incorporating four additional negative emotions included as controls (disgust, worry, sadness and anxiety). The idea is to verify that the mediation model is truly based on guilt by ruling out the potential mediating effect of other emotions (Grappi et al., 2013). The feelings examined have the same valence of guilt and therefore are potentially associated with this emotion. The results were supported also when including these covariates in the model. The only significant effects were a positive relationship between anticipated disgust and guilt (effect = 0.13, CI from = 0.04 to = 0.19) and a positive effect of worry on willingness to buy (effect = 0.29, CI from = 0.13 to = 0.45). The overall pattern of results was not affected by these significant effects. This finding supports the unique role of guilt as mediating mechanism in the theory proposed.

Discussion
In summary, the model of moderated mediation is supported by the data. The indirect effect of CRM through guilt is significant and moderated by the level of fit. Specifically, when fit is high, CRM coupled with hedonic (rather than utilitarian) products reduces guilt. This is because hedonism is intrinsically linked with guilt (Okada, 2005) and when buying hedonic products consumers search for ways to assuage their negative feelings (Prelec and Loewenstein, 1998). Furthermore, high product-cause fit influences the combination between CRM and hedonic products and makes it more appealing while it does not have a positive influence on a utilitarian offering coupled with a CRM campaign. This is because, when buying hedonic alternatives, consumers search for a justification to their decisions actively (Botti and McGill, 2011). On the other hand, in the case of utilitarian choices, external justifications are not necessary and therefore a high fit CRM campaign is less relevant to consumers.

Implications for research
This study offers the first examination of the process that explains why hedonic products can be successfully combined with CRM initiatives. While Hagtvedt and Patrick (2015) have demonstrated that linking CRM with the purchase of very expensive luxury goods assuage consumers' feelings of guilt, this study shows that the same mechanism applies also to more accessible products, supporting the applicability of the theory to a much wider range of purchase situations. Furthermore, the study extends work by Hagtvedt and Patrick (2015) showing that the mediation model advocated is dependent on the level of fit. When fit is low, anticipated guilt does not mediate the effect of CRM on willingness to buy. The findings of this research suggest that CRM campaigns should predominantly be used in conjunction with products that contain guilt-inducing hedonic features. CRM campaigns can offer the opportunity to improve such consumption experiences by reducing the feelings of guilt intrinsically connected with pleasurable offerings (Botti and McGill, 2011;Simonson, 2002a, 2002b;Okada, 2005). This effect, however, is dependent on high fit between the cause and the product features. This study extends significantly Hedonic consumption through guilt reduction existing research on how and why CRM increases the appeal of hedonic products (Hagtvedt and Patrick, 2015;Strahilevitz, 1999;Strahilevitz and Myers, 1998) by clarifying the underlying mechanism explaining this type of consumption. Moreover, our results complement research on hedonic-utilitarian consumption showing that people look for ways to justify their hedonic choices because these often entail feelings of guilt Simonson 2002a, 2002b;Okada 2005;Strahilevitz and Myers 1998). Importantly, our study extends research on hedonic consumption showing how CRM campaigns represent a way to reduce the sense of ambivalence often associated with these types of products. The findings are therefore also relevant for research on the connection between CSR and luxury or other forms of hedonic purchase. While there is some evidence that consumers are less interested in ethical considerations when purchasing luxury products (Davies et al., 2012;Torelli et al., 2012), CRM campaigns appear to be useful tools to keep CSR relevant in consumers' decisions even in the context of luxury choices. This is because, in addition to increasing the salience of socially responsible initiatives, CRM can also enhance the purchase experience and therefore offers an immediate, additional benefit to consumers (Hagtvedt and Patrick, 2015).
Findings of the present study extend previous research on perceived fit in CRM by demonstrating that this variable is also important in regulating emotional processes activated by CRM initiatives. Past work focused on fit as a moderating variable on the overall effectiveness of CRM initiatives (Pracejus and Olsen, 2004;Bignè-Alcaniz et al., 2012), while the current research demonstrates the psychological impact this variable plays in shaping emotional reactions.
The study contributes further to research on guilt as an emotion able to promote responsible consumption decisions. First, the study provides additional evidence, in line with past research, of guilt's strong connection with hedonic consumption experiences (Okada, 2005). Second, the study extends past research describing how guilt can promote the purchase of products displaying responsible/ethical labels (Antonetti and Maklan, 2014) demonstrating that this emotion also underpins support for CRM initiatives. Third, the findings demonstrate that feelings of guilt are reduced by the level of fit between the CRM initiative and the features of the product considered (Duhachek et al., 2012). This is an important finding because it clarifies in what contexts of consumer ethical decision-making it is reasonable to expect that guilt will play a significant role. It is also an important consideration for practitioners who might wish to design CRM campaigns to influence consumer choices.

Managerial implications
The findings of the present study suggest important managerial implications. First, results confirm that charity appeals work better with hedonic products than utilitarian products (Strahilevitz 1999;Strahilevitz and Myers, 1998). Whenever possible it is better for marketers to link CRM campaigns with offerings that have a hedonic nature rather than more functional alternatives. Furthermore, as guilt reduction plays a key role in the process analysed, effective campaigns should be able to communicate credibly the outcomes of the philanthropic donation. Research shows that guilt reduction will be stronger when consumers feel that the donation associated with the product is really effective in tackling the social cause presented (Coulter and Pinto, 1995). Overall, a well-designed CRM strategy can make the purchasing experience more gratifying and can garner more support for the cause promoted owing to a reduction of the feelings of guilt felt by consumers when buying hedonic alternatives. EJM 51,11/12 Moreover, such a solution is effective only when there is a strong association between the CRM program and the product. One of the main drivers of CRM effectiveness is the selection of charitable partners (Bignè-Alcaniz et al., 2012) that provide a good fit with the brand/ product (Pracejus and Olsen, 2004;Simmons and Becker-Olsen, 2006). Present findings give support to this assumption and suggest that practitioners should pay attention to the nature of the fit between the product and the charitable partners to improve CRM effectiveness in soothing the negative emotions experienced during hedonic purchases.
As CRM programs are becoming increasingly popular, managers must design campaigns able to stand out from the competition. During hedonic shopping experiences, people look for information that will allow them to justify their purchase through the support for a CRM campaign. This means that, when promoting CRM, the hedonic aspects of a product or service should be stressed to enhance the appeal of the social cause. If utilitarian elements are also present, they should be downplayed in the communication.
Managers of charities and non-profit organizations wishing to launch CRM campaigns should also consider the crucial role of product-cause fit. This information can aid them in searching for potential partners that are likely to have a high functional or image-based fit. Collaborations between high fit partners are more likely to offer win-win situations for both the social cause and the for-profit corporation.

Limitations and areas for further research
The study presents a few limitations that offer interesting avenues for future research. The analysis focuses on products rather than brands. The availability of branded information might influence the results since it might introduce other variables in the analysis (e.g. the brand heritage, positioning and country of origin). Nonetheless, the findings are important in demonstrating how consumer reactions vary systematically across categories, and in relation to different types of products that are generally perceived as hedonic. Future research can examine the model presented in relation to CRM campaigns associated with existing brands and explore potential variables that might extend the theorizing presented here (e.g. pre-existing beliefs about a brand, brand relationships). An additional limitation of the study involves the use of a student sample. Student samples have been criticised mostly because of the limitations in terms of generalizability of the results (Henrich et al., 2010). Although we recognize that our sample has unique demographic characteristics, there is no evidence in the literature that leads us to suggest that the psychological processes examined here should depend on income, age or education. On the contrary, much past research on CRM and hedonic consumption was conducted using student samples and is therefore consistent with the approach adopted here (Botti and McGill, 2011;Hagtvedt and Patrick, 2015). Nonetheless, we do see value in further examinations aimed at investigating our theory on nationally representative samples and in a cross-cultural context.
In our analysis, we have argued that the role of fit could be dependent on two potential mechanisms. It is possible that low fit causes activate consumer skepticism toward the campaign (Drumwright, 1996) and that they might simply be less persuasive because less congruent with the product (Folkes and Kamins, 1999). Our studies, however, did not test explicitly these mechanisms. Future research can examine these two explanations empirically to extend further our understanding how the reasons that explain the role of fit in reactions to CRM.
The results show that guilt plays a key role in explaining why it is effective to link hedonic products with CRM campaigns. Further research should explore the potential role that positive (rather than negative) emotions play in this mechanism. Pride is a positive emotion that shares important similarities with guilt (Antonetti and Maklan, 2014) and Hedonic consumption through guilt reduction might be considered in future research. Findings from past research show that people feel pride when supporting responsible consumption choices (Antonetti and Maklan, 2014) and at the same time that luxury brands are able to elicit strong feelings of pride (McFerran et al., 2014). It is possible therefore that pride will be increased by CRM initiatives linked with luxury brands. Future research might test this proposition because it would open the possibility of using CRM to enhance the positive emotions participants experience when purchasing luxury offerings. This is a possibility that promises to make CRM a compelling sponsorship tool for luxury brand managers. Finally, future research should examine explicitly the potential role of guilt appeals in the promotion of CRM campaigns linked to hedonic products. Past research shows that the effectiveness of guilt appeals depends on a number of factors (Antonetti and Baines, 2015), and in some cases, guilt-based messages can backfire (Coulter and Pinto, 1995). The analysis presented here suggests that CRM initiatives should offer an ideal situation for the successful implementation of guilt-based messaging. Future studies however should test this possibility directly and determine the best conditions for the communication of guilt in CRM initiatives. Finally, future research might address the choice of the "best" fit between the product and the cause taking into account different levels of perceived similarity (e.g. functional vs imagebased). Future investigations might focus also on a plausible negative effects of the fit (Yoon et al., 2006) between the partners involved in a CRM initiative that might cause the consumers to doubt the sincerity of the company's motives.
Notes 1. To check whether different price levels might have somewhat affected the results, we decided to include alternatives that varied in terms of their relative price (e.g. massage coupon e100 and icecream sundae e5). To examined this possibility, we conducted all relevant analyses using only cheap products (120 evaluations) or only expensive products (120 evaluations). The results are robust and consistent suggesting that differences in the monetary value of the products do not affect the results.
2. All participants in the no fit condition saw two products, but the same cause (i.e. aid for victims of environmental disaster). This is different from the fit condition where the products and the cause varied. To check whether this difference might have somewhat affected the results, we also conducted all relevant analyses using only one product evaluation for each participant (120 evaluations). The results are consistent suggesting that using the same cause twice did not affect the results. We thank an anonymous reviewer for this observation.