Academy of Marketing Studies Journal (Print ISSN: 1095-6298; Online ISSN: 1528-2678)

Research Article: 2019 Vol: 23 Issue: 1

When Love-Becomes-Hate Effect Happens: An Empirical Study of the Impact of Brand Failure Severity Upon Consumers Negative Responses

Kongkidakarn Sakulsinlapakorn, Huazhong University of Science and Technology

Jing Zhang, Huazhong University of Science and Technology

Abstract

The present study investigates four factors (aggressive personality, brand trust, blame attribution, and perceived fairness) leading to Love-becomes-hate effect through the moderating effect of brand love on the relationship between failure severity and consumer’s negative emotions. This paper empirically examines the factors leading to Love-becomes-hate effect based on a questionnaire survey among 532 Chinese respondents. This study found that at high level of Aggressive personality, low level of Brand trust, high level of Blame attribution, and low level of Perceived fairness are considered as factors leading to Love-becomes-hate effect. Consumers with the above traits decide to vent negative emotions and pursue retaliation actions against the focal firms. This study contributes to the theory development of brand failure and consumer retaliation literature. This study also provides suggestions for private and public companies on how to properly deal with consumer’s negative responses to product or service failure. Research findings can be guideline for managers to take the quick decision and prompt action when product or service failure occurs.

Keywords

Love-Becomes-Hate Effect, Failure Severity, Consumer's Negative Emotions, Consumer Retaliation, Brand Failure.

Introduction

Companies are often encountered with the problems of customer’s negative behaviors: when customers face with unfavorable events of product or service failure, they are likely to turn hostile and may cause damage to firms (Kahr et al., 2016). Followings are two typical cases of consumer’s negative responses towards company’s product and service failures. Jeremy Dorosin was an angry customer who bought Starbucks espresso maker. He found that this espresso maker was defective. He has wasted a lot of time complaining to Starbucks. Later on, he decided to run a campaign against Starbucks on Wall Street Journal. This scandal has been reported thoroughly on television and published in newspapers and this caused serious harm to Starbuck’s reputation (Flinn, 1995). Dave Carroll, a Canadian musician, claimed his Taylor guitar was broken during a trip on United Airlines in 2008. He had been trying to negotiate with United Airlines about this matter and it lasted nine months without any satisfactory result. Finally, he wrote a song “United Breaks Guitars” describing his negative experience and this song quickly spread on the Internet (Kahr et al., 2016). This scandal caused up to 180 million dollars in damage to United Airlines (The Economist, 2009). These two cases indicate that nowadays customers are empowered by availability of media and technologies to adopt retaliation behavior against firms (Labrecque et al., 2013). In general, customers usually vent their negative emotions against firms by engaging in negative activities, including vindictive complaining to firms, third-party complaining (e.g. internet, mass media, agency), and negative word-of-mouth (Bonifield & Cole, 2007), which will result in substantial losses of financial assets, brand equity, investor confidence, and corporate reputation.

The study of the mediating role of emotion in the product/service failure and consumer negative emotions link has not been extensively researched in extant marketing literature. In addition, the research findings about the moderation role of brand love in the relationship between brand failure and consumer retaliation are consistent with previous studies. Emotional affection toward brand would buffer negative effects of a product or service failure because consumer is more tolerant with regard to brand transgressions (Tax et al., 1998), making retaliation behavior less probable. In contrast, consumers with lower levels of brand love do not have such tolerance and they will be more likely respond with CBS (Consumer Brand Sabotage: this term is from a study of Kahr et al. (2016). However, it is also possible that consumers with high expectations of brand would perceive a brand failure as betrayal (Thompson et al., 2006), leading to a “love-becomes-hate” effect (Grégoire et al., 2009) and thereby increasing the likelihood of consumer retaliation responses. Therefore, researchers further investigate the contingent factors that impact the brand love’s moderating role in product or service failure and consumer’s negative response link.

This study will: (1) illustrate how failure severity leads to consumer retaliation via consumer’s negative emotion as a mediator, and (2) examine when love-becomes-hate effect happens (when brand love moderates failure severity and consumer’s negative responses link). We will examine how aggressive personality, brand trust, blame attribution, and perceived fairness impact the moderation role of brand love in the relationship between failure severity and consumer’s negative responses. Therefore, academics and marketing managers can gain important insights from our research findings on how to manage product/service failure and how to deal with customer negative behaviors properly.

The remainder of the paper proceeds as follows. First, we review literature, define research constructs, establish conceptual framework and develop hypotheses regarding mediating role of emotion and moderating role of brand love in the relationship between failure severity and consumer retaliation. This is followed by a description of the methods used to test the framework and hypotheses. Subsequently, the research findings are reported. Finally, the conclusions and implications of the study are discussed, and limitations and future research direction are presented.

Literature Review And Research Hypotheses

The Link of Failure Severity and Consumer’s Negative Emotions

Severity issue is discussed in the context of product or service failure and it is related to scope of brand failure and criticality (Weun et al., 2004). Service failure can be defined as service performance of a service provider or a firm that fails to meet customer’s expectations (Kelley & Davis, 1994). In service failure situation, customers may immediately perceive that it causes inconvenience and aggravation to them (Zourrig et al., 2009). In addition, when minor product or service failure with mild inconvenience occurs, a customer may elicit low levels of negative behaviors (Folkes, 1984). However, product or service failure that causes a big inconvenience to customer, it can result in serious consequences such as severe loss to company or vindictive behaviors against firms (Bechwati & Maureen, 2003). According to descriptive approach in marketing, customers may have different emotional reactions depending on their experiences (Oliver, 1997). Negative affect is a broad concept relating to attitudes which commonly associated with negative emotions (i.e. sadness, anger, and hostility) (Diener et al., 1995). Therefore, understanding of the complexity of emotions in negative service experience helps firm to keep unexpected damage from hostile customers away (Bearden & Teel, 1983). In this study, we attempt to capture three of negative emotions including anger, dissatisfaction, and perceived betrayal.

First, according appraisal theory, anger is a basic human emotion that is caused by external attribution (Roseman, 1991). Anger may expose when customers condemn a firm for deterring them from completing their goals (Kahr et al., 2016). Second, the study of dissatisfaction has been widely investigated (Souca, 2014); it is generally examined in contrast with satisfaction (Mittal et al., 1999). Theoretical framework of satisfaction can be applied to acquire concept of dissatisfaction and to classify its key components as affective response (Giese & Cote, 2000). Besides, affective response describes the dissatisfied reaction of customers towards unfavorable experiences with strong emotion and feelings (e.g. angry, disappointed, and cheated) (Giese & Cote, 2000). Third, perceived betrayal is a customer’s perception of firm’s norm violation in the relationships between them (Bechwati & Morrin, 2003). Perceived betrayal is associated with product or service failure because when stronger relationship customers experience unfair failure situations, it leads to perception of betrayal (Grégoire & Fisher, 2008). Lee et al. (2013) argue that in service failure context: when normative standard in the relationship between customers and service providers is violated, then they are likely to perceive betrayal.

Based on the above mentioned understanding, we put forward the following hypotheses (Figure 1):

academy-entrepreneurship-Conceptual-Model

Figure 1:Indicates The Conceptual Model, Which Illustrates All The Research Hypotheses In This Paper.

The Link of Consumer’s Negative Emotions and Consumer Retaliation

The Link of Consumer’s Negative Emotions and Consumer Retaliation

A number of studies in service research indicated the linkage between negative emotions and customer behavior. The recognition of the negative experience specifically leads customers to take an action upon unfavorable service experience or similar events (Oliver & Westbrook, 1993). Ward & Ostrom (2006) posit that betrayal is the main driver for consumer motivation to participate in online consumer protest movements. Occasionally, victims of betrayal incidents are labeled as grudge holders that can drive them to get involved with aggressive behaviors (Koehler & Gershoff, 2003). Retaliatory behavior refers to an action taken by a customer as a coping strategy in response to the tension and frustration that caused negative experience by firms (Porath et al., 2010). It is a kind of customer’s negative behavior intending to punish and cause difficulties to firms for something harmful that firms have done (Grégoire & Fisher, 2008). The equity theory is considered as a significant basis of revenge and retaliation studies (Funches et al., 2009). Retaliation associate with customer’s intention to restore equity or cope with misbehavior because in some cases customers just want to protect themselves and other customers from misbehavior that would occur in the future (Tripp et al., 2002). In this study, by following Gelbrich (2010) and Johnson et al. (2011) retaliation consists of three kinds of behaviors, namely vindictive complaining to the firm, third-party complaining for negative publicity, and negative word-of-mouth.

Vindictive complaining is a direct form of customer retaliatory behavior which related to the action that a customer aims to complain directly to firm’s frontline employees or representatives about a product or service problem with purpose of seeking revenge (McColl-Kennedy et al., 2009). Third party complaining is an indirect form of customer retaliatory behavior. In this study, it associates with the action of a customer aiming to complain directly to online or offline third party (e.g. a media, an agency, complaint website) with the purpose of spreading the wrongdoings of a firm to public and making it go viral (Grégoire et al., 2010). Negative word-of-mouth is another indirect form of customer retaliatory behavior. It is related to the action that a customer aims to speak out his or her negative experience to friends and family or other people in order to warn them to stay away from firms and reduce future patronage (Grégoire & Fisher, 2006).

Based on the above mentioned arguments, we propose:

H2: Consumer’s negative emotions have a positive effect on consumer retaliation.

The Moderating Effect of Brand Love

“Brand love” is defined as “the degree of passionate emotional attachment a satisfied consumer has for a particular trade name” (Carroll & Ahuvia, 2006). It is associated with a strong bond of affection between a customer and a product of service of a brand which is similar to “interpersonal love” (Albert et al., 2008; Langner et al., 2015). Love for a brand can be found when a customer’s strong feeling of wanting to have a specific product reaches its aim (Ahuvia, 2005). Customers are not willing to separate from a brand or change to other brands (Fournier, 1998). Literature holds contradictory viewpoint about the moderating roles of brand love in the linkage of failure severity and negative emotions. Some literature believes that love involves tolerance about the mistake of other parties. When customers encounter with brand failure situation, brand love can relieve the levels of aggressions or hostile thoughts. In addition, when a consumer has a positive brand attitude and positive past experiences with beloved brands (Joji & Ashwin, 2012), a customer is able to integrate part of his/her self-expressiveness by demonstrating love toward it (Huber et al., 2015). In other words, when an unfavorable situation happens, brand love can eliminate levels of negative emotions and reduce the likelihood of a retaliatory behavior. The concept of love-becomes-hate effect in business research has been recently studied by Grégoire & Fisher (2006), Grégoire et al. (2009), and Kahr et al. (2016). For example, a strong relationship customer tends to feel more betrayed and get involved with aggressive behaviors when he or she encounters with a product or service failure because he or she perceives that firms owe him or her much more than the weak relationship customer (Grégoire et al., 2005). Furthermore, Grégoire et al. (2009) state that strong relationship customers usually have stronger desire for revenge and tend to hold a grudge for longer period of time than weak relationship customers; and this phenomenon of love-becomes-hate usually takes place over time. In this paper, we admit the second viewpoint about the love-becomes-hate effect and present the following hypothesis:

H3: Brand love positively moderates the relationship between failure severity and consumer’s negative emotions.

Four Contingent Factors of Love-Becomes-Hate Effect

Researchers believe that the moderation role of brand love in the link between product/service failure and consumer’s negative response depends on contingent factors. For instance, Grégoire & Fisher (2008) examine high and low relationship quality customer’s response to a poor recovery after a service failure, and find that high relationship quality customers feel more betrayed when they perceive a low level of both distributive fairness and process fairness, and they are more likely to take the negative actions against firms. Next, we will explore the four contingent factors, as second order moderating variables that may impact the moderating effect of brand love in failure severity and consumer’s negative emotions link.

Aggressive personality

People with aggressive personality in nature are more likely to exhibit high levels of negative emotions and aggression in difficult situations than people with low levels of aggression (Anderson & Bushman, 2002). According to consumer behavior study, a customer with an aggressive personality has higher tendency to perceive emotional-provoking situations by failure of a brand and behaves aggressively against a firm (Kahr et al., 2016). In this case, a customer is more sensitive to situational provocation, as a consequence; he/she may easily engage in aggressive behaviors upon an unfavorable situation (Marshall & Brown, 2006). Likewise, when an aggressive customer encounters with brand failure, he/she would take the action against a firm in the form of hostility, as well as, possess negative emotions (Anderson et al., 2008). In our study, we develop an understanding aggressive behavior in a consumer-brand relationship context which regards as one of our second-order variables that moderates brand love in order to propose theory development. This variable also controls and manages customer love. In this case, consumers could have expressed the levels of aggressions along with the levels of brand love in the same direction. We also developed the hypotheses to test Love-becomes-hate effect as well. Thus, we predict that consumers with high levels of aggressive personality have the different reaction from consumers with low levels of aggressive personality. Specifically, we put forward the following hypotheses:

H4: Aggressive personality positively moderates the brand love’s moderation effect in the relationship between failure severity and consumer’s negative emotions. To be more specific.

H4a: When aggressive personality is high, brand love positively moderates the relationship between failure severity and consumer’s negative emotions (love-becomes-hate effect will happen).

H4b: When aggressive personality is low, brand love does not moderate the relationship between failure severity and consumer’s negative emotions (love-becomes-hate effect will not happen).

Brand trust

Brand trust is built from past experience, and it will also reflect future experience of a brand (Drennan et al., 2015). Hence, a consumer’s trust in a brand usually results in positive outcomes consisting of positive attitudes, commitment, and faithfulness (Albert et al., 2008). For example, when consumers encounter unpleasant consumption experience, brand trust can help reduce their uncertainty and vulnerability feeling (Coulter & Coulter, 2002). In this case, consumers may hold the belief that the brand will correct the mistake appropriately and still keep its promise in future offering provision (Coulter & Coulter, 2002). As a result, they are less likely to turn their brand love into hatred. In contrast, Robinson (1996) addresses that product or service failure causes a strong relationship customer to obsess the feeling of lack of trust (low level of trust); as a result, it leads to negative emotions and retaliatory behaviors. Thus, we put forward the following hypotheses:

H5: Brand trust negatively moderates the brand love’s moderation effect in the relationship between failure severity and consumer’s negative emotions. To be more specific.

H5a: When brand trust is high, brand love does not moderate the relationship between failure severity and consumer’s negative emotions (love-becomes-hate effect will not happen).

H5b: When brand trust is low, brand love positively moderates the relationship between failure severity and consumer’s negative emotions (love-becomes-hate effect will happen).

Blame attribution

Blame attribution is defined as the perception of customers about firms to be responsible for social irresponsibility and failed recovery, and this mental action can be perceived after failed recovery (Grégoire et al., 2010; Zourrig et al., 2009). According to attribution theory (Kelley, 1967), blame can be classified into three dimensions consisting of locus, stability, and controllability (Folkes, 1984; Weiner, 1980). Specifically, locus of behavior refers to customer’s perception of where responsibility for a brand failure should be placed including of internal or external situation that causes the crisis (Iglesias, 2009; Klein & Dawar, 2004). Stability of the behavior refers to customer’s perception of whether a brand failure situation remains the same or temporary (Swanson & Kelley, 2001; Weiner, 1980). Controllability of the behavior refers to customer’s perception of brand failure that occurs under the control of the firm or outside of the firm's control (Weiner, 1980; Wirtz & Mattila, 2004). Previous studies addressed that customers blame the firms for negative behaviors which cause the negative outcomes, such as customer anger, negative word-of-mouth, requirement of customer compensation (e.g. refund, apology), and so on (Folkes, 1984). Moreover, when customers perceive that product failure is caused by the firms (under the firm’s control), they tend to have the higher levels of anger and have the desire to hurt firms (Folkes, 1984). In this case, it is more likely that love-becomes-hate effect happens, for the reason that consumers are not able to shift responsibilities of product/service failure to other parties and only the firm can be the target for consumers to release their negative emotions. Thus, we propose the following hypotheses:

H6: Blame attribution positively moderates the brand love’s moderation effect in the relationship between failure severity and consumer’s negative emotions. To be more specific.

H6a: When blame attribution is high, brand love positively moderates the relationship between failure severity and consumer’s negative emotions (love-becomes-hate effect will happen).

H6b: When blame attribution is low, brand love does not moderate the relationship between failure severity and consumer’s negative emotions (love-becomes-hate effect will not happen).

Perceived fairness

Fairness judgment can be classified into three elements including; distributive fairness, interactional fairness, and procedural fairness (Smith et al., 1999). Firstly, distributive fairness refers to the situation that an individual receives what he or she expects to get from the firm, such as a positive outcome (Kim et al., 2010). Secondly, interactional fairness refers to the situation that an individual assumes to be treated with the politeness and respect by firms (Cropanzano et al., 2001). It is related to the ways and the manner of service staff treating customers during period of service recovery which can lead to the levels of fairness perception (Blodgett et al., 1997). Thirdly, procedural fairness refers to the situation that an individual views decision-making processes of a firm to reach an outcome in a dispute as fair (Tax et al., 1998).

In general, customers may form different perceptions and behaviors on fairness during product or service failure and recovery (Magnini & Ford, 2004). Previous studies of Grégoire et al. (2010) and McColl-Kennedy et al. (2009) found lower levels of perceived fairness in brand failure situation lead to higher levels of anger and negative responses. Likewise, customers who suffer from severe brand failure usually perceive lower levels of fairness leading them to keep in mind that firm merely cares for its own interests; as a consequence, it triggers customers to possess negative emotions (Crossley, 2009), and a desire for punishment the unfair firms (Ambrose & Schminke, 2009). On the other hand, when they believe they are treated by the firm in a fair way, they will be more tolerant and hold a less negative attitude towards the unpleasant consumption experience. Therefore, it’s less likely for them to convert brand love into hostility. Thus, we propose the following hypotheses (Figure 1):

H7: Perceived fairness negatively moderates the brand love’s moderation effect in the relationship between failure severity and consumer’s negative emotions. To be more specific.

H7a: When perceived fairness is high, brand love does not moderate the relationship failure severity and consumer’s negative emotions (love-becomes-hate effect will not happen).

H7b: When perceived fairness is low, brand love positively moderates the relationship between failure severity and consumer’s negative emotions (love-becomes-hate effect will happen).

Research Methodology

Research Design

To study of the mediating role of emotion in the brand failure and consumer negative emotions link. It is possible that consumers with high expectations of brand would perceive a brand failure as betrayal (Thompson et al., 2006), leading to a “love-becomes-hate” effect (Grégoire et al., 2009) and thereby increasing the likelihood of consumer retaliation responses. Therefore, researchers further investigate four contingent factors (aggressive personality, brand trust, blame attribution, and perceived fairness) that impact the brand love’s moderating role in product or service failure and consumer’s negative response link.

This study will: (1) illustrate how failure severity leads to consumer retaliation via consumer’s negative emotion as a mediator, and (2) examine when love-becomes-hate effect happens (when brand love moderates failure severity and consumer’s negative responses link). We will examine how aggressive personality, brand trust, blame attribution, and perceived fairness impact the moderation role of brand love in the relationship between failure severity and consumer’s negative responses. Hierarchical moderated regression analysis was employed to analyze the hypothesis in this study.

Questionnaire Design

This study has developed the survey questionnaire (close-ended-question) to acquire the responses from Chinese respondents. The questionnaire was made in Chinese and English versions for the ease of respondents to answer. It was divided into three parts, including brand failure details (2 items), demographic profile (6 items), and measurement scales (32 items). Respondents were asked to indicate their level of agreement toward each statement (a five-point Likert scale), from 1=strongly disagree to 5=strongly agree.

Data Collection

This study employed the purposive sampling method; it is practically synonymous with quantitative research. It is a non-representative subset of some larger population, and constructed to serve for the exclusive need or purpose. The data was collected by delivering one by one questionnaire survey in the department stores to Chinese people in Wuhan city, China. This survey was conducted around three months from 1st of July, 2017 to 1st of October, 2017. Random sampling technique was employed and a total of 550 respondents were asked to participate the survey. After deleting low quality ones, we were left 532 valid responses. In order to guarantee high response quality, at the beginning of the survey session, researchers professionally asked every single respondent whether he or she had encountered product or service failure or not. Then, respondents were asked to fill in the questionnaires. Respondents were promised that their answers were kept strictly confidential by the authors, therefore no individual information was disclosed, and only collective data analysis was used.

Data Analysis Procedure

In this study, Lisrel version 8.8 will be used to analyze the reliabilities and validities of measurement scales. In order to get the results from the hypothesis tests, the SPSS version 21.0 will be used to analyze the data. The data analysis procedures are conducted by the following methods;

Structural Equation Modeling (SEM)

Structural Equation Modeling (SEM) was conducted to analyze the reliabilities and validities of measurement scales.

Hierarchical moderated regression analysis

Hierarchical moderated regression analysis was conducted, according to the procedure delineated in Cohen & Cohen (1983), to examine the moderating effect of brand love on the relationship between failure severity and consumer’s negative emotions at different levels of aggressive personality, brand trust, blame attribution, and perceived fairness. The significance of interaction effects was assessed after controlling main effects. The moderating role of brand love in the relationship between failure severity and consumer’s negative emotions at different levels of aggressive personality, brand trust, blame attribution, and perceived fairness. Gender and age were entered first as control variables, while predictor variable (failure severity) was entered in the second step. Moderator variable (brand love) was entered in the third step. Lastly, interaction term was entered in the fourth step. In order to avoid multicollinearity problems, the predictor and moderator variables were centered and the standardized scores were used in the regression analysis (Aiken & West, 1991).

Results

Characteristics of Respondents

The basic characteristics of the 532 respondents, including gender, marital status, age, education level, occupation, and income per month. 52.44% of respondents were male, and 47.56% of respondents were female. Most of the respondents were single (51.69%), followed by in partnership (28.38%), married (18.80%), and divorced (1.13%). Besides, most of the respondent’s age was between 18 to 25 years old (56.20%), followed by 26 to 35 years old (25.94%), 36 to 45 years old (10.34%), 46 to 55 years old (3.20%), less than 18 years old (3.01%), and more than 55 years old (1.31%). 54.14% of the respondents held Bachelor’s Degree, Master’s Degree (30.45%), Doctoral Degree (12.03%), and high school or lower (3.38%). The largest demographic group was students (51.88%), followed by management & professional (17.86%), freelance/part time (15.60%), self-employed (13.16%), housewife (0.94%), and unemployed (0.56%). 33.08% of respondents had an income of 86-430 USD, followed by 431-860 USD per month (28.38%), 861-1,715 USD per month (15.98%), Less than 85 USD (12.41%), and above 1,716 USD (10.15%).

Reliabilities and Validities of Measurement Scales

Descriptive analysis and confirmatory factor analysis are used to assess all scale’s reliabilities and validities, as Tables 1 and 2 indicate. As shown in Table 1, all the construct’s Cronbach’s alpha coefficients (ranging from 0.708 to 0.912) and the Composite Reliabilities (CR) (ranging from 0.834 to 0.928) indicate that each exceeds the accepted reliability threshold of 0.70. In addition, all of the Average Variance Extracted (AVEs) is greater than 0.5 cut-off (ranging from 0.614 to 0.748). Thus, all the measures demonstrate adequate reliabilities.

Table 1: Descriptive Statistics, Correlations, Reliabilities And Discriminant Validities Of Measurements
Variables 1 2 3 4 5 6 7 8
Failure severity 0.864a              
Negative motions 0.274**b 0.832            
Brand love 0.683** 0.189** 0.799          
Aggressive personality -0.162** -0.012 -0.168** 0.791        
Brand trust 0.137* 0.119** 0.097* -0.338** 0.840      
Blame attribution 0.022 -0.031 0.009 0.152** -0.255** 0.829    
Perceived fairness 0.027 -0.111* 0.094* -0.245** -0.031 -0.164** 0.784  
Retaliation 0.764** 0.245** 0.772** -0.133** 0.068 0.094* 0.045 0.810
Mean 3.640 2.220 3.587 3.226 2.293 3.432 2.351 3.758
S.D. 0.927 0.972 0.909 0.974 0.988 1.008 0.863 0.901
Cronbach’s alpha 0.839 0.857 0.819 0.708 0.861 0.849 0.793 0.912
Composite reliability 0.899 0.926 0.903 0.834 0.928 0.925 0.898 0.919
AVE 0.748 0.693 0.638 0.626 0.706 0.687 0.614 0.656
Table 2: Measurement Scale Items And Cfa Results
Blame attribution
(Gregoire et al., 2010; Gregoire & Fisher, 2008)
1. The firm was totally responsible for the failures (1)-not at all responsible for the failure (5). 0.95 (20.89)
2. The brand failure was completely the firm’s fault. 0.93 (19.49)
3. From the brand failure situation, I wasted a lot of time and effort dealing with this issue. 0.96 (19.96)
4. To what extent do you blame the firm for what happened? Not at all (1)-completely (5). 0.88 (17.61)
Perceived fairness
(Gregoire et al., 2009; Joireman et al., 2013)
1. The employee(s) who interacted with me treated me with. .empathy. 0.75 (16.40)
2. Overall, the outcomes I received from the firm were fair. 0.77 (16.49)
3. Given the time, money, and hassle, I got fair outcomes. 0.84 (17.73)
4. The firm gave me an opportunity to have a say in the handling of the problem. 0.72 (15.33)
Consumer retaliation 1. I complained to firm to be unpleasant with the representative of the company. 0.86 (23.86)
(Gregoire et al., 2009; Joireman et al., 2013). 2. I complained to firm to make someone from the organization suffer for their services. 0.86 (21.90)
3. I complained to a social media to have it reported my experience to other consumers. 0.81 (18.48)
  4. I complained to a social media so that my experience with the firm would be known. 0.81 (21.62)
  5. I spread negative word-of-mouth about the firm. 0.94 (22.91)
  6. When my friends were looking for a similar product or service, I told them not to buy from this firm. 0.90 (22.49)

As shown in Table 2, CFA yields a model that fits the data well with NNFI and CFI all exceeding 0.90, GFI exceeding 0.80, and RMSEA not exceeding 0.08. All item loadings ranging from 0.72 to 0.96 are significant at the 0.01 level, which indicates convergent validities of all the measures are acceptable. Finally, according to Tables 1 and 2 all diagonal elements representing the square root of the AVEs are larger than any other corresponding row or column entry, which means that each construct sufficiently differs from other constructs and, therefore, the discriminant validities of all measures are established.

Regression Analysis

Relationship between failure severity and consumer’s negative emotions

Regression analysis was used to test H1, which predicts failure severity has a positive effect on consumer’s negative emotions. The results, as shown in Table 3, indicate that standardized regression coefficient of failure severity upon consumer’s negative emotions is significantly positive at 0.001 level (β=0.276, p<0.001), R2 is 0.085 with p-value of 0.000, and F value is greater than 4. Therefore, H1 is supported.

Relationship between consumer’s negative emotions and consumer retaliation

Regression analysis was used to test H2, which predicts consumer’s negative emotions have a positive effect on consumer retaliation. The results, as shown in Table 3, indicate that standardized regression coefficient of consumer’s negative emotions upon consumer retaliation is significantly positive at 0.001 level (β=0.242, p<0.001), R2 is 0.062 with p-value of 0.000, and F value is greater than 4. Therefore, H2 is supported.

Table 3: Regression Of Failure Severity-Consumer’s Negative Emotions (H1) And Regression Of Consumer’s Negative Emotions And Consumer Retaliation (H2)
Variables Failure severity-Consumer’s negative emotions Consumer’s negative emotions-
Consumer Retaliation
Model 1a Model 2a Model 1b Model 2b
1.Control variables
Gender 0.096 (2.169*) 0.095 (2.248*) 0.039 (0.894) 0.016 (0.379)
Age 0.008 (0.174) 0.019 (0.456) 0.043 (0.964) 0.041 (0.949)
2.Independent variable
Failure severity - 0.276 (6.614***)   0.242 (5.724***)
VIF (≤) 1.035 1.002 1.035 1.010
R2 0.009 0.085 0.004 0.062
Adjusted R2 0.006 0.080 0.000 0.057
F value 2.523+ 16.398*** 1.060 11.670***
∆R2 0.009 0.076 0.004 0.058
∆F value 2.523 43.739 1.060 32.764
Sig. ∆F value 0.081 0.000 0.347 0.000

Love-becomes-hate effect

Hierarchical moderated regression analysis was conducted, according to the procedure delineated in Cohen & Cohen (1983), to examine the moderating effect of brand love on the relationship between failure severity and consumer’s negative emotions at different levels of aggressive personality, brand trust, blame attribution, and perceived fairness. The significance of interaction effects was assessed after controlling main effects.

We test the H3 with the whole sample. The left half of Table 4 indicates the results about moderating role of brand love in the relationship between failure severity and consumer’s negative emotions. Gender and age were entered first as control variables (Model 1a) while predictor variable (failure severity) was entered in the second step (Model 2a). Moderator variable (brand love) was entered in the third step (Model 3a). Lastly, interaction term was entered in the fourth step (Model 4a). In order to avoid multicollinearity problems, the predictor and moderator variables were centered and the standardized scores were used in the regression analysis (Aiken & West, 1991). As can be seen in Model 4a results from Table 4, the interaction effect for failure severity and brand love has a positive effect on consumer’s negative emotions (β=0.158, p<0.01), and F value is greater than 4. Therefore, H3 is supported.

Table 4: Test Results About Moderating Effects Of Brand Love In The Relationship Between Failure Severity And Consumer’s Negative Emotions (Whole Sample) In Love-Becomes-Hate Effect: Standardized Coefficients (T Value)
Variables Moderating effects of brand love (whole sample)
Model 1a Model 2a Model 3a Model 4a
1.Control variables
Gender 0.096 (2.169*) 0.095 (2.248*) 0.096 (2.258*) 0.105 (2.477*)
Age 0.008 (0.174) 0.019 (0.456) 0.020 (0.465) 0.006 (0.141)
2.Independent variable
Failure severity   0.276 (6.614***) 0.266 (4.653***) 0.298 (5.173***)
3.Moderating variable
Brand love     0.014 (0.253) 0.081 (1.338)
4.Interaction variable
Failure severity×Brand love     0.158 (3.065**)
VIF (≤) 1.035 1.002 1.887 1.548
R2 0.009 0.085 0.085 0.101
Adjusted R2 0.006 0.080 0.078 0.093
F value 2.523+ 16.398*** 12.293*** 11.870***
∆R2 0.009 0.076 0.000 0.016
∆F value 2.523 43.739 0.064 9.396
Sig. ∆F value 0.081 0.000 0.801 0.002

Moderating role of aggressive personality in love-becomes-hate effect

To test the H4 (H4a and H4b), we first divide the whole sample into two groups with high and low level of aggressive personality by using mean value as cutoff. The left half of Table 5 indicates the results about moderating role of brand love in high level of aggressive personality. Following the same procedure as indicated in H3 to get the results. As can be seen in Model 4a results from Table 5, the interaction effect for failure severity and brand love has a positive effect on consumer’s negative emotions (β=0.156, p<0.05), and F value is greater than 4. Therefore, H4a is supported.

Table 5: Test Results About Moderating Effects Of Aggressive Personality In Love-Becomes-Hate Effect: Standardized Coefficients (T Value)
Variables High level of Aggressive personality Low level of Aggressive personality
Model 1a Model 2a Model 3a Model 4a Model 1b Model 2b Model 3b Model 4b
1.Control variables
Gender 0.124 (2.322*) 0.115 (2.246*) 0.112 (2.172*) 0.123 (2.392*) 0.017 (0.193) 0.034 (0.383) 0.040 (0.462) 0.039 (0.449)
Age 0.021(0.396) 0.026 (0.503) 0.025 (0.482) 0.009 (0.167) 0.014 (0.158) 0.023 (0.267) 0.046 (0.525) 0.041 (0.462)
2.Independent variable
Failure severity   0.286 (5.618***) 0.335 (4.097***) 0.344 (4.231***)   0.180 (2.116*) 0.120 (1.333) 0.131 (1.436)
3.Moderating variable
Brand love     -0.063 (-0.767) 0.029 (0.319)     0.171 (1.882+) 0.180 (1.963+)
4.Interaction variable
Failure severity×Brand love     0.156 (2.365*)       0.068 (0.781)
VIF (≤) 1.014 1.001 2.582 1.708 1.086 1.014 1.183 1.080
R2 0.016 0.098 0.100 0.114 0.001 0.032 0.057 0.061
Adjusted R2 0.011 0.090 0.089 0.101 -0.014 0.011 0.029 0.026
F value 2.919+ 12.636*** 9.613*** 8.910*** 0.043 1.521 2.048+ 1.756
∆R2 0.016 0.082 0.002 0.014 0.001 0.032 0.025 0.004
∆F value 2.919 31.538 0.588 5.591 0.043 4.476 3.542 0.610
Sig. ∆F value 0.055 0.000 0.444 0.019 0.958 0.036 0.062 0.436

Following the same procedure, we test H4b in the subsample of low level of Aggressive personality. As indicated in Model 4b of the right half of Table 5, the interaction effect for failure severity and brand love has no effect on consumer’s negative emotions. Therefore, H4b is also supported.

Moderating role of brand trust in love-becomes-hate effect

To test the H5 (H5a and H5b), we first divide the whole sample into two groups with high and low level of brand trust by using mean value as cutoff. The left half of Table 6 indicates the results about moderating role of brand love in high level of brand trust. Following the same procedure as indicated in H3 to get the results. As can be seen in Model 4a results from Table 6, the interaction effect for failure severity and brand love has no effect on consumer’s negative emotions. Therefore, H5a is supported.

Table 6: Test Results About Moderating Effects Of Brand Trust In Love-Becomes-Hate Effect: Standardized Coefficients (T Value)
Variables High level of Brand trust Low level of Brand trust
Model 1a Model 2a Model 3a Model 4a Model 1b Model 2b Model 3b Model 4b
1.Control variables
Gender 0.095 (1.120) 0.123 (1.398) 0.129 (1.467) 0.130 (1.473) 0.103 (1.921+) 0.078 (1.525) 0.076 (1.481) 0.081 (1.610)
Age -0.066 (-0.778) -0.052 (-0.604) -0.039(-0.452) -0.039(-0.450) 0.057 (1.065) 0.044 (0.876) 0.045 (0.877) 0.006 (0.124)
2.Independent variable
Failure severity   0.108 (1.239) 0.057 (0.537) 0.047 (0.431)   0.318 (6.327***) 0.345 (5.038***) 0.390 (5.694***)
3.Moderating variable
Brand love     0.088 (0.815) 0.069 (0.597)     -0.040 (-0.581) 0.052 (0.727)
4.Interaction variable
Failure severity×Brand love     -0.045 (-0.439)       0.222(3.649***)
VIF (≤) 1.060 1.110 1.709 1.530 1.028 1.009 1.855 1.527
R2 0.010 0.021 0.025 0.027 0.016 0.116 0.117 0.149
Adjusted R2 -0.003 0.000 -0.002 -0.008 0.010 0.108 0.107 0.137
F value 0.766 1.024 0.933 0.781 2.823+ 15.431*** 11.636*** 12.298***
∆R2 0.010 0.010 0.005 0.001 0.016 0.100 0.001 0.032
∆F value 0.766 1.536 0.665 0.193 2.823 40.025 0.338 13.317
Sig. ∆F value 0.467 0.217 0.416 0.661 0.061 0.000 0.561 0.000

Following the same procedure, we test H5b in the subsample of low level of brand trust. As indicated in Model 4b of the right half of Table 7, the interaction effect for failure severity and brand love has a positive effect on consumer’s negative emotions (β=0.222, p<0.001), and F value is greater than 4. Therefore, H5b is supported.

Moderating role of blame attribution in love-becomes-hate effect

To test the H6 (H6a and H6b), we first divide the whole sample into two groups with high and low level of blame attribution by using mean value as cut-off. The left half of Table 7 indicates the results about moderating role of brand love in high level of blame attribution. Following the same procedure as indicated in H3 to get the results. As can be seen in Model 4a results from Table 7, the interaction effect for failure severity and brand love has a positive effect on consumer’s negative emotions (β=0.200, p<0.001), and F value is greater than 4. Therefore, H6a is supported.

Table 7: Test Results About Moderating Effects Of Blame Attribution In Love-Becomes-Hate Effect: Standardized Coefficients (T Value)
Variables High level of Blame attribution Low level of Blame attribution
Model 1a Model 2a Model 3a Model 4a Model 1b Model 2b Model 3b Model 4b
1.Control variables
Gender 0.118 (2.362*) 0.093 (1.941+) 0.092 (1.919+) 0.102 (2.159*) 0.032 (0.336) 0.081 (0.826) 0.085 (0.858) 0.086 (0.860)
Age 0.059 (1.177) 0.051 (1.064) 0.050 (1.054) 0.033 (0.701) -0.157 (-1.661+) -0.094 (-0.926) -0.090 (-0.885) -0.090 (-0.885)
2.Independent variable
Failure severity   0.304 (6.419***) 0.317 (5.082***) 0.357 (5.709***)   0.178 (1.681+) 0.058 (0.352) 0.062 (0.364)
3.Moderating variable
Brand love     -0.019 (-0.309) 0.064 (0.972)     0.150 (0.943) 0.155 (0.924)
4.Interaction variable
Failure severity×Brand love     0.200 (3.527***)       0.013 (0.098)
VIF (≤) 1.024 1.008 1.725 1.487 1.054 1.335 3.014 1.931
R2 0.020 0.111 0.112 0.139 0.023 0.047 0.055 0.055
Adjusted R2 0.015 0.105 0.103 0.128 0.006 0.022 0.021 0.012
F value 4.001* 16.668*** 12.497*** 12.773*** 1.381 1.877 1.629 1.294
∆R2 0.020 0.092 0.000 0.027 0.023 0.024 0.007 0.000
∆F value 4.001 41.199 0.095 12.442 1.381 2.826 0.890 0.010
Sig. ∆F value 0.019 0.000 0.758 0.000 0.256 0.095 0.348 0.922

Following the same procedure, we test H6b in the subsample of low level of Aggressive personality. As indicated in Model 4b of the right half of Table 7, the interaction effect for failure severity and brand love has no effect on consumer’s negative emotions. Therefore, H6b is supported.

Table 8: Test Results About Moderating Effects Of Perceived Fairness In Love-Becomes-Hate Effect:Standardized Coefficients (T Value)
Variables High level of Perceived fairness Low level of Perceived fairness
Model 1a Model 2a Model 3a Model 4a Model 1b Model 2b Model 3b Model 4b
1.Control variables
Gender 0.080 (1.029) 0.081 (1.080) 0.081 (1.073) 0.076 (1.005) 0.102 (1.765+) 0.103 (1.854+) 0.100 (1.769+) 0.121 (2.123*)
Age 0.040 (0.514) 0.073 (0.961) 0.067 (0.878) 0.061 (0.804) -0.059 (-1.015) -0.050 (-0.902) -0.053 (-0.948) -0.064 (-1.145)
2.Independent variable
Failure everity   0.258 (3.533**) 0.228 (2.956**) 0.249 (3.162**)   0.279 (5.048***) 0.318 (3.002**) 0.334 (3.164**)
3.Moderating variable
Brand love     0.089 (1.169) 0.110 (1.412)     -0.046 (0.429) 0.044 (0.385)
4.Interaction variable
Failure severity×Brand love     0.098 (1.274)       0.160 (2.189*)
VIF (≤) 1.087 1.017 1.122 1.149 1.014 1.001 3.719 1.762
R2 0.010 0.075 0.082 0.091 0.012 0.090 0.091 0.105
Adjusted R2 -0.001 0.059 0.061 0.065 0.006 0.081 0.079 0.090
F value 0.882 4.786** 3.939** 3.487** 1.890 9.858*** 7.419*** 6.969***
∆R2 0.010 0.065 0.007 0.008 0.012 0.078 0.001 0.014
∆F value 0.882 12.481 1.366 1.623 1.890 26.483 0.184 4.792
Sig. ∆F value 0.416 0.001 0.244 0.204 0.153 0.000 0.669 0.029

Moderating role of perceived fairness in love-becomes-hate effect

To test the H7 (H7a and H7b), we first divide the whole sample into two groups with high and low level of perceived fairness by using mean value as cut-off. The left half of Table 8 indicates the results about moderating role of brand love in high level of perceived fairness. Following the same procedure as indicated in H3 to get the results. As can be seen in Model 4a results from Table 8, the interaction effect for failure severity and brand love has no effect on consumer’s negative emotions. Therefore, H7a is supported.

Following the same procedure, we test H7b in the subsample of low level of perceived fairness. As indicated in Model 4b of the right half of Table 8, the interaction effect for failure severity and brand love has a positive effect on consumer’s negative emotions (β=0.160, p<0.05), and F value is greater than 4. Therefore, H7b is supported.

Discussion

This study examines how failure severity impacts consumer’s negative emotions and consumer retaliation. Empirical findings show that failure severity has a positive effect on consumer’s negative emotions, and consumer’s negative emotions have a positive effect on consumer retaliation. To be more specific, when customers encounter with product or service failure, failure severity usually leads customers to obsess with negative emotions. In addition, gender as one of our control variables (in the moderating role of the failure severity and consumer’s negative emotions link at high levels of aggressive personality, low levels of brand trust, high levels of blame attribution, and low levels of perceived fairness) implies that when female consumers encounter product or service failure situations, they are more likely to possess higher levels of negative emotions than male consumers do. As a result, negative emotions practically drive both male and female consumers to take the actions against firms in terms of retaliation. This research finding reveals product/service failure generates consumer retaliation via the mediating role of negative emotions. Therefore, managing customer’s emotions should be of strategic importance for brand to successfully deal with product/service failure and prevent retaliation.

Also, this study investigates love-becomes-hate effect, which refers to positive moderating role of brand love in the relationship between failure severity and consumer’s negative emotions, and finds empirical evidence for this effect. The previous studies on the moderation of brand love in the failure and negative responses link put forward inconsistent arguments and research findings. Some suggest brand love can make consumers more tolerant and therefore offset their negative emotions (Joji & Ashwin, 2012); while others believe that brand love will reinforce the impact of failure severity and customer’s negative responses (Gregoire & Fisher, 2005). Our research findings support the second viewpoint which is also termed love-becomes-hate effect, indicating that brand love may cause great trouble for firms to handle customer’s emotions in the case of product/service failure.

What’s more, this study examines four contingent factors of love-becomes-hate effect, and finds that brand love plays positive moderating role in failure severity and consumer’s negative emotions link at high levels of aggressive personality, low levels of brand trust, high levels of blame attribution, and low levels of perceived fairness. To the best of our knowledge, the analysis of the contingent factors of love-becomes-hate effect has not been done before by other researchers.

The major findings of this study provide several significant managerial implications for private and public companies on how to properly deal with consumer’s negative responses to product or service failure. The results indicate that product or service failure will lead to consumer retaliation via triggering consumer’s negative emotions such as anger, dissatisfaction and perceived betrayal. Therefore, when a brand makes some mistakes (failure crisis); it should make every effort in managing consumer’s emotions in order to avoid their possible retaliation behaviors (vindictive complaining to the firm, third-party complaining for negative publicity, and negative word-of-mouth). Managers need to develop appropriate failure recovery strategies: pairing an apology with compensation for product or service failure (Joireman et al., 2013). Besides, managers should provide training sessions for employees about dealing with customers with respect, friendly, and empathy in brand failure situation. They should set staff empowerment policy because when problems occur; then, stuff will be able to handle customer’s complaints in time and it enables customers to claim for compensation easily. Our research findings provide following suggestions to brand managers;

First, our qualitative review implies that consumers with aggressive personality have the tendency to engage in aggressive behavior only in response to product or service failure across situations. Also, company should evaluate consumer’s aggressive personality level by tracking their past consumption and claim history. Also, nowadays on the social media platforms, consumers with high degree of aggressive personality can be identified by employing big-data technology and browsing their posts or comments. Then, for those consumers with high level of aggressive personality, managers should take quick actions when product or service failure situation occurs in order to prevent unfavorable love-becomes-hate effect. If the process takes too long, then these highly risky consumers may get involved with retaliatory actions against firms, for example, the retaliation cases of Dorosin (Starbucks Case), and Dave Carroll (United Airlines Case).

Secondly, our findings show that when consumers with low level of brand trust encounter product or service failure, their brand love will strengthen the consumer’s negative emotions. Therefore, company should create the strong brand trust in the mind of consumers by communicating the brand’s values to nurture consumer’s identification with that brand. A brand should have a strong brand image to enable new and existing customers to effectively express themselves to others through the use of the brand. To build brand trust is not a short-term plan; it is a long-term program which needs to try to improve to the positive directions day by day (Chaudhuri & Holbrook, 2001; Sharma & Patterson, 1999).

Thirdly, the study reveals that consumers with high level of blame attribution are more likely to turn their brand love into hatred and then take negative actions against brands. This finding suggests that managers should develop the better strategies to shape consumer’s perception about blame attribution in coping with brand failure. When product or service failure occurs, managers should admit the fault which is really caused by the firms, and also clarify the other stakeholder’s responsibilities in causing failure to consumers, especially those loyal customers, at the earliest time. They should not let consumers to fight for their rights and brands should be responsible for product or service failure and provide consumers with the positive final solutions. As a result, this may prevent consumers with high level of blame attribution from taking hostile actions against firms.

Conclusion

The results of low level of perceived fairness show that when consumers perceive lower levels of fairness (e.g. fairness violation), they are more like to suffer Love-becomes-hate effect. This result recommends managers to establish a clear and fair compensation policy to improve consumer’s perceived procedural fairness when they design failure recovery strategy. Also, firms need to pay extra attention to the fairness outcomes when dealing with customers to guarantee distributive fairness. Besides, managers should provide training sessions for employees about dealing with customers with respect and empathy in failure situation. They should set staff empowerment policy because when the problem occurs; then, stuff will be able to handle customer’s complaints in time and it enables customers to claim for compensation easily. In this way, consumers can perceive higher level of interactional fairness. All of these three kinds of perceived fairness will contributes to offset Love-becomes-hate effect and manage failure recovery strategies in a more effectively way.

Limitation And Future Research Directions

This study has some suggestions for future research. First, future research may identify other factors that would affect “love-becomes-hate effect” or moderate the relationship between failure severity and consumer’s negative emotions, and consumer’s negative emotions and consumer retaliation, such as brand loyalty, Relationship Quality (RQ), and so forth. Second, this study only examines the brand love’s moderating role in the link between failure severity and consumer’s negative emotions. Future research can also probe into the potential moderation effect of brand love in the relationship between consumer’s negative emotions and consumer retaliation.

Besides, there are two limitations of the present study. First, this survey was mainly conducted in China. For future research, researchers may conduct the survey in some other countries or specific regions, or conduct the comparative research between two countries in order to gain insights from the broader population groups. Second, future research may conduct both quantitative and qualitative approaches (in-depth interview) to obtain more effective and high quality results.

References