Which rebuild strategy is the most effective for luxury brands to use? A quantitative research concerning different response strategies that yield different effects on brand image

Which rebuild strategy is the most effective for

luxury brands to use?

A quantitative research concerning different response strategies that yield different effects on brand image

Master Thesis

Sjue jing Yao - 12404845

Graduate School of Communication Master Track Corporate Communication Supervisor: mw. dr. S.C. de Bakker Wordcount: 9245

1

Abstract

Corporate crises and crisis response strategies have been researched abundantly in the past. The reason to conduct another research concerning these related topics is because

stakeholders and clients expect a response from the corporation, and certainly even more if the CEOs are directly responsible for the crisis. The unfortunate event that happened to the brand Dolce and Gabbana is one of the few organizational crises that happened in the past two years. This research will examine which component(s) of Coombs rebuild strategy would work the best when a brand is tangled himself in a racial crisis. By means of an experimental study, response elements are tested and determined what would benefit a luxury brand the most after a media crisis. Establishing and maintaining a strong and positive brand image is a continuous journey for corporations and it is highly dependent on the current assessment of its internal and external relations. This research focusses on the different elements of the rebuild strategy and its effect on brand image. Prior research state that brand identification is an important indicator for how clients and customers react on certain crises and crisis

response strategies and therefore the decision has been made to include brand identification as a moderator. A strong level of brand identification can lead to a more positive evaluation of brand image after a crisis hits the brand than when the brand identification is low. A consumer that strongly identifies with a certain brand will have positive brand associations and feelings about it. Results show that there is no difference in brand image between the different conditions (apology, regret, apology and regret, and no response). However, brand identification does influence the evaluation of brand image after a racial crisis.

2

Introduction

On November 18, 2018, a crisis hit the high-end fashion brand Dolce and Gabbana when they released a controversial campaign video on Western social media platforms (Instagram and Twitter ), as well as the Chinese social media platforms (Weibo and WeChat) (Ritschel, 2018). The campaign initially was intended to bring an ode to the Chinese culture and their upcoming fashion show in Shanghai (Raleigh, 2018). However, the campaign received the opposite response, as many outraged consumers, fans and followers posted messages that reflected their fury on various social media platforms (Tashjian, 2018). Hashtags such as #boycotdolce and #boycotdolceandgabbana had been a popular topic trending on the Chinese social media platforms for days (Kong, 2018).

The campaign video was a series of episodes showing and introducing Chinese people with Italian food. The video content of the campaign showed a Chinese female wearing a Dolce and Gabbana dress and holding a pair of chopsticks (Raleigh, 2018). A mandarin male voice narrates the video and instructs the Chinese model how to eat pizza, spaghetti, and cannoli with chopsticks (Xu, 2018). The connotation and the language use of the male narrative are characterized as condescending and insulting. To illustrate, phrases such as “Finally we have a food with a size that won’t make you not know how to handle” and “How to use these little sticks to eat our traditional margherita pizza?” were used in the video (South China Morning Post, 2018, 0:06). Some viewers identified the content of the episodes as mocking and prejudiced, as comical pronunciations of certain Italian words were

purposely used and the Chinese model showed submissive and foolish behaviour in the video (Zhang, 2018). The highlight of the crisis was when private messages of Stefano Gabbana were made public in which he described China with poop emojis and called Chinese people “Ignorant, dirty smelling mafia” (Gabbana, 2018, removed Instagram post). These messages appeared after people criticized the media campaign of the brand. The brand’s initial plan for

3 the mediacampaign was that it was supposed to stimulate the audience’s interest in the big fashion show in Shanghai. The fashion show was planned to be held on November 21, 2018, but it was cancelled by China’s cultural and tourism department just a few hours before the event took place (Zheng & Pan, 2018). Where it started as a perhaps genuine effective media campaign, the video and messages backfired the brand’s goal.

The unique factor, in this case, is that the brand is still run by the designers Domenico Dolce and Stefano Gabbana. In many corporate crises in the past where the Chief Executive Officer (CEO) played a significant part in creating the crisis, resignation is often the solution. It depends on several factors such as: the organization, the type of crisis and the severity of crisis whether the CEO resigns voluntarily or as a collective decision made by the board (Bouaine, Charfedinne, Arouri & Teulon, 2015). Examples are the resignation of Martin Winterkorn (Volkswagen crisis in 2015) and Kwon Oh-hyun (Samsung bribery crisis in 2017) (Ewing, 2015; Grothaus, 2017). If the CEO is perceived to be strongly connected with the crisis, a resignation will help the brand to recover faster (Agnihotri & Bhattacharya, 2019). The brand Dolce and Gabbana is a private brand and not owned by a parent company. Domenico Dolce and Stefano Gabbana own more than 80% of the stake and they initiated the crisis, which means that they are held accountable for the racial media campaign (Szalai, 2013). The designers of the company will still be in charge of the company and therefore one has to realize that the company will not easily be forgiven for its mistake(s). This context requires careful consideration of options and solutions to minimize the harm that is caused by the controversial media campaign. According to Coombs and Holladay (2001), content that includes discriminatory information is an intentional misdeed, which most often results in a reputational threat. Because the designers will not resign and the crisis is categorized as self-initiated, the rebuild strategy is the most appropriate to be implemented to target the crisis. Rebuild strategy proves to be most effective for organizations when the crisis could have

4 been prevented and thus holds a high level of responsibility (Coombs, 2007; van der Meer & Verhoeven, 2014).

Crisis response strategies have been researched in an abundant and comprehensive manner in the past by communication and media scholars (Benoit, 1997; Coombs, 2007, 2008, 2010, 2015, 2018; Coombs & Holladay, 2008; Heath, 1998; Kim, Avery & Lariscy, 2009). Nevertheless, the specific elements of the rebuild strategy have not been applied to analyse the aftermath of a crisis concerning a racial media campaign. Hence, this research will demarcate the different elements and analyse whether the separate components yield different levels of brand image.

This study focusses on the responses of consumers towards their favourite brand involved in a racial crisis. To this date, racial crisis in corporations have been investigated, but only concerning discrimination and prejudice within organizations and the workplace

Deitch, Barsky, Butz, Chan, Brief & Bradley, 2003; Green, 2003). Moreover, campaigns concerning anti-racism have been researched, but racial media campaigns have not been explored in depth (Nakamura, 2012). Although students have examined the racial crisis of Hennes and Maurits (H&M) in 2018, when the fast fashion chain had a picture on their webshop featuring a dark-skinned little boy wearing a t-shirt with ‘the coolest monkey in town’ (Beyer, Bierling & Öjerbrant, 2018; Mensah & Osman, 2018).

It is important to emphasize that these type of organizational crisis (campaigns and advertisements errors) are not life-threatening dangerous, nor unfamiliar with communication and media professionals. Nevertheless, these mistakes happen and people expect a certain response from the board, managers or CEO’s regarding the undesirable situation. According to Coombs (2015), every crisis is unique and therefore requires a tailored crisis response strategy. By analysing and replicating the racial crisis, brands and corporations who find themselves in similar situations in the future can use the findings of this research to provide a

5 guiding framework for their unique response to racial crises. Following the introduction, a research question is formed:

RQ: After the release of a culturally inappropriate media campaign, what element(s) of the

rebuild strategy has the most positive effect on a luxury brand’s image when influenced by brand identification?

Theoretical framework Racial crisis

Coombs (2016) designed three crisis clusters that can help organizations identify the nature of the crisis. The first cluster is the ‘victim cluster’ where the organization is a victim as well in the crisis. A crisis can be identified in the accidental cluster if the organization had no intentional motifs concerning the crisis. The preventable cluster is when the organization is aware of their actions and violations leading up to a crisis.

The campaign video’s released by Dolce and Gabbana was meant to function as a perhaps genuine effective media campaign. Ultimately, the video and messages backfired the brand’s goal, as people started to speak up on several media platforms and express their opinion. The crisis was (un)consciously created by the brand, as the media campaign contained racial and discriminating content. The assumption is that the dynamic designers wanted to create a controversial campaign that would gain peoples full attention. By

analysing their past campaigns (Appendix C), it becomes apparent that the designers love to look up the extremes and create provoking campaigns and fashion items (Demopoulos, 2018). Some individuals are convinced that the designer’s “bratty bad boys” attitude are enhancing their brand’s overall image (Givhan, 2018; Ginsberg, 2018). Nevertheless, this specific campaign turned out to be too controversial and denigrating, as the video campaign received criticism the moment it was online.

6 Considering the discriminatory content of the #DGlovesChina campaign, this crisis is identified as an intentional organizational misdeed (Coombs and Holladay, 2001). No

external or internal relation is physically hurt, but the content does contain hurtful language towards the Chinese culture and population. The private messages of Domenico Dolce (Appendix A), show that he and Stefano Gabbana did not care about how they made Chinese people feel with their campaign. The knowledge of the designers’ provocative past, combined with the content of Dolce’s private messages ensured that this crisis is not categorized in the accidental cluster, but in the preventable cluster.

Brand image

The brand image is the direct meaning, impression, and association that appears to mind when the stakeholders are exposed to the brand. The image of a brand is dynamic of nature and can be altered at any moment. An established (positive) brand image can be used to influence the consumer of rating the brand with a higher likeability and stimulate the purchasing behaviour (Sasmita & Mohd Suki, 2015).

This research will apply the conceptualization of brand image according to Abratt and Kleyn (2012). Brand image in their opinion is construed from brand experiences, brand relationships and brand communities. Brand experiences are the sensory feelings and

behavioural responses that a consumer experiences when exposed to brand-related products, services and information (Brakus, Schmitt & Zarantonello, 2009). These brand-related stimuli can range from characteristics such as the font type and colour to more significant characteristics such as slogans and marketing communications (Abratt & Kleyn, 2012). Brand relationship can be defined as the relationship between the consumer and the overall brand and its products (Chang & Chieng, 2006). It is the connection between a consumer and a brand that has come into existence after one or more encounters between consumers and the

7 brand or brand-related stimuli. The consumer experience plays a significant role in the

relationship, as consumers reflect on their experience with the brand and as a result of the reflection a relationship is defined (Aggarwal, 2004). The level of relationship also depends on how consumers evaluate the brand’s reputation, overall corporate image, and its culture.

According to Abratt and Kleyn (2012), to build and maintain a strong relationship, corporations need to be consistent, strategic and strong in their communication. Not only communication but also promises that have been communicated by the organization need to be fulfilled in order to keep a strong relationship (Abratt & Kleyn, 2012). Brand communities include individuals who form a ‘community’ by sharing to some extent the same set of feelings, emotions, values, and standards about a brand (Schouten, McAlexander & Koenig, 2007). The love, admiration, and enjoyment that consumers feel for the brand and brand-related stimuli, are fundamental for brand communities (Abratt & Kleyn, 2012). Moreover, individuals that are part of a brand community share a mutual interest in other community members and experience a high level of belonging to this specific group (Ouwersloot & Odekerken-Schröder, 2008).

Rebuild strategy

The situational crisis communication theory by Coombs (2007) is used when designing the fictional response from the Off-white. Coombs created the Situational Crisis Communication Theory in 1995, in order to provide crisis managers the fitting crisis response for the possible threat (Coombs, 2010). He proposed that there are three different types of responses that a brand can give after a crisis occurs, which is part of the SCCT. The three options a brand can decide on are: (a) to go on denial, (b) to diminish, or/and (c) to rebuild (van Zoonen & van der Meer, 2015). Organizations may decide to approach the crisis with the denial strategy when they do not see the connection between them and the crisis, or if they do not want to be

8 related to the crisis. This could be effective if the crisis revolves around a rumour or perhaps a misleading news report (Coombs, 2007). Secondly, diminish strategies are applied by corporations if they do acknowledge the crisis but argue that the crisis is not as bad as the public think (Claeys & Cauberghe, 2012). Moreover, organizations apply the diminish strategies when they are of the opinion that the crisis is not fully their responsibility. The crisis could be partly started by factors such as accidents and unforeseen mistakes. Lastly, rebuild strategies are practiced when organizations realize that the crisis is (partly) created by intentional errors and misconducts. When corporations apply the rebuild strategy, they need to understand that it depends on the nature of the crisis whether to provide material

compensation or symbolic aid or even both (van Zoonen & van der Meer, 2015). Relating back to the case of Dolce and Gabbana, one can state that the #D&GlovesChina campaign is categorized in the preventable cluster, as the brand

intentionally created a campaign that violated cultural and social norms, which means that the brand carries a high level of crisis responsibility. Crises that are categorized in the

preventable cluster is advised to implement the rebuild strategy in their crisis response (Coombs, 2006, 2007, 2010)

This research will examine the elements of the rebuild strategy more in-depth. The rebuild strategy contains the elements apology, regret, and compensation (Coombs 2006; van der Meer & Verhoeven, 2014). The element compensation is not included, because Dolce and Gabbana’s media crisis is categorized as a moral crisis. The crisis did not concern product

harm, and no consumer was physically hurt by the crisis. Moreover, because it was a racial crisis, it is not appropriate to offer a specific person of group of individuals a monetary gift (Baker, 2001). Therefore, the decision has been made to focus on the elements apology and

9 Elements of rebuild strategy

Apology

According to Coombs (2002), an apology is when an organization expresses its responsibility and request forgiveness from the public. Kellerman (2006) state that organizations and brands should acknowledge the importance of official apologies coming from CEOs. The CEO’s are head of the organization and therefore ultimately responsible for the crisis that occurred. Moreover, the internal and external relations will feel a sense of sincerity, as the head of the organization recognizes the mistakes.

Apologies are meant to restore the injustice done by a person or a group (Exline, Deshea & Holeman, 2007). It is characterized as one of the most effective strategies to use after a corporate crisis, as the CEO exhibit feelings of guilt, embarrassment, and

responsibility (Brocato, Peterson & Crittenden, 2012). Individuals value apology because it demonstrates that an organization cares about the mistakes that they have made. Previous inquiries suggest that apology can be conceptualized as showing regret and accepting full responsibility for the crisis that happened (Frandsen & Johansen, 2010). It is vital to know that there is a difference between regret and apology. An apology can contain expressions of regret about the situation, but it does not work the other way around. Regret is a state of wishful thinking; if one could go back in time, one would have made different decisions. Given that studies have proven that the use of apology is the most effective for preventable crises, this study contends the following hypothesis:

H1a: An official statement given by the CEO providing an apology generates a better

10 Regret

Prior studies indicate that organizations and their spokespersons should try to avoid an

official apology, as apology is used as liable evidence in lawsuits (Coombs, 2006; Coombs & Holladay, 2008). Therefore the role of emotional communication is essential during post-crisis communication because it allows the public to view the organization as a dynamic construct that makes mistakes. Corporations can show empathy and regret in a response, without risking the chance to face considerable financial loss during a lawsuit.

Moreover, a response that contains emotional cues is perceived to be more genuine and empathic than when signs of emotions are absent. Showing emotions also expresses concern for the crisis, the victims, employees, and stakeholders. For every crisis cluster, there is a selection of emotions that are the most effective to apply in post-crisis communication strategies. Lee (2005) illustrates that a mere apology is insufficient for an organization to recover from a crisis. Stakeholders will hold more positive and stronger brand associations when a brand addresses and recognize their role and responsibility regarding the crisis in a public apology. Related to this, Van der Meer and Verhoeven (2014) claim that providing an apology without emotions (regret) is experienced as less efficient than when an organization provides an apology containing the element regret. Namely, forgiveness and empathy play important roles in how an organization will be perceived by the public after the crisis.

Additionally, being sincere in a response message is highly recommended, as it will attribute in conveying the message as well as making an emotional connection with the audience (Kim, 2011; Kim & Cameron, 2011). Hence, the aim is to find out whether the dimension regret from the rebuild strategy is needed in order to generate the most positive brand image after a preventable crisis.

By using emotions in crisis response, the aftermath of organizational crises is

11 to the prior findings concerning emotional communication, the following hypothesis is

formed:

H2a: An official statement given by the CEO containing only elements of regret generates a

better post-crisis brand image, than a statement without regret.

Apology and regret

According to Cooley and Cooley (2011), in situations where the management does not resign after a crisis, which is the case of Dolce and Gabbana, the situation requires both a statement containing apology and regret in order to restore the brand’s reputation. The reason is that they need to convince their internal and external relations that they indeed acknowledge the mistake and that they are responsible for it. Moreover, people who are affected by the crises need to experience genuine empathy coming from the organization and an official apology without empathy or signs of regret will not be as effective (Claeys, Cauberghe & Leysen, 2013).

Additionally, forgiveness is more likely to be achieved when emotions are put into an apology than when it is absent from an apology. By expressing regret in an apology, the receiver of the message will be more likely to accept it and the individual will experience the message as more truthful and authentic. The combination of regret (emotions) and apology indicate sincerity, which shows that the wrongdoer feels bad about the error (Hareli &

Eisikovits, 2006).The element regret shows signs of empathy and compassion, which triggers the behaviour of forgiveness. Crisis response containing empathy and apology will generate a higher level of acceptance and forgiveness will lead to a more positive brand image than when clients are defensive about the response (Howell, Turowski & Buro, 2012). Given that studies have proven that the combination of emotions (regret) and apology is the most effective for preventable crises, this study contends the following hypothesis:

12 H3a: An official statement given by the CEO containing both the elements regret and apology

generates a better post-crisis brand image, than when the CEO provides a statement with one

of these elements or provides no response

Brand identification

Additionally, the success of the strategy by the consumer is also partly dependent on the identification of the consumer with the brand. Studies have shown that the stronger the brand identification, the more ‘insurance’ the brand has in a crisis situation (Johar, Birk &

Einwiller, 2010; Lisjak, Lee & Gardner, 2012; Lin & Sung, 2014). Thus, it could be theorized that the negative information of a crisis will not immediately affect the consumers who have a high brand identification. Therefore, it is expected that brand identification will act as a moderator in this research.

Mishra (2017) established a pattern and recognized three types of brand consumers, namely devotees, doubters, and dropouts. Devotees are individuals that identify with the brand and have a strong connection with the brand and its products. These individuals will more likely show low reluctance towards the brand after a crisis. Doubters are moderately enthusiastic about the brand but do have a positive attitude towards the brand. These

individuals show little reluctance of purchasing a product from the brand after the crisis. The last category are dropouts, this group of customers have a weak emotional bond with the brand and lost their trust in the brand after the crisis. Products of the brand will not be purchased by this group anymore. Nevertheless, the accuracy of this statement depends on the severity of the crisis.

A consumer that strongly identifies with a certain brand will have positive brand associations and feelings about it. Research proves that loyal consumers show more biased behaviour towards the brand than non-loyal consumers (Cleeren, Dekimpe & Helsen, 2008).

13 The consumers who strongly identify with a brand tend to dismiss the negative information concerning the brand and its products. They show resilient behaviour towards the negative information or interpret the information more favourable than the people who weakly identify with the brand (Einwiller, Fedorikhin, Johnson & Kamins, 2006; He & Li, 2011). This could result in a more positive attitude toward the brand, its reputation, and image. Moreover, Stokburger-Sauer, Ratneshwar & Sen (2012) indicate that brand identification has a stronger influence in luxury product categories than in corporations that sell convenience goods. In other words, brand identification could be a crucial element in how consumers evaluate the brand’s overall image, as a luxury brand is involved.

On another note, Einwiller and colleagues (2006) prove that the tolerance level of strong brand identifiers is higher in comparison to weak brand identifiers, but it is not limitless. They argue that the role of loyal and strong identifying consumers can work counterproductive. To illustrate, negative information that shapes the crisis may directly attack the consumer’s basis of brand identification and thereby indirectly violates a part of the values and believes that they hold. So, in this case, such a crisis will lead to more negative reactions from identifiers than non-identifiers (Bhattacharya & Sen, 2003; Einwiller, Fedorikhin, Johnson & Kamins, 2006 Schmalz & Orth, 2012).

This research, however, will postulate that a strong brand identification will lead to a more positive post-crisis brand image than when individuals have a weak brand identification with Off-White. The decision to follow this line of reasoning is based on the fact that the participants are exposed to one negative news report about the crisis. The participants are not exposed to any external sources (news, social media and peers) that discusses this crisis, which means that they base their emotions and opinions on a single news report provided in the experiment. The lack of quantity of negative news releases and social media posts,

14 lead to a more positive post-crisis brand image, and therefore the following hypotheses are formed:

H1b: An official statement given by the CEO providing an apology generates a better

post-crisis brand image than when no apology is provided, and this effect is stronger for high

brand identification

H2b: An official statement given by the CEO containing elements of regret generates

a better post-crisis brand image than a statement without regret, and this effect is stronger

for high brand identification

H3b: An official statement given by the CEO containing both the element regret and

apology generates a better post-crisis brand image than a statement with one of these

elements, and this effect is stronger for high brand identification

Methods

Research design

A quantitative approach is used in this research to test the various rebuild strategies for a luxury brand to apply. In this experimental research, 4 conditions have been tested in order to answer the research question. The conditions represent the elements of the rebuild theory by Coombs (2006). The conditions are as follows: A response containing the element apology (A), a response with words of regret (B) and response with both an apology as well as

elements of regret (C). The control group will receive no response in this experimental set-up. This inquiry depended on an experiment executed by means of surveys on Qualtrics. Different factors will be exposed to 4 different groups of participants, in order to examine the possible difference in outcome. The experiment consisted of 3 forms of rebuild response that companies and crises experts can implement in case of crisis and a control condition (no response).

15 Stimuli

The brand that is applied for the experimental design is Off-White by Virgil Abloh. The reason to use this brand in the experiment was because Off-White has been named the most popular brand targeted at the age of 16 – 37 (Newbold, 2019). The aim was to investigate which element of the rebuild strategy ensures the most positive brand image after a brand faces danger during a media crisis. Off-White has not faced a media crisis yet concerning cultural misunderstanding. The decision has been made to not use the exact Dolce and Gabbana crisis in the experiment, as it has been on the news and on social media platforms for a considerable period of time. To limit prior knowledge and emotions about the crisis as much as possible, it was better to recreate a similar situation instead of using the real crisis. In order to remove any familiarity with the case, a fictitious crisis had been created for the brand Off-White.

The experiment was based on a true news report written by Wilkinson (2018) for CNN business. He wrote about the crisis that had hit Dolce and Gabbana after the release of the “DGlovesChina” campaign described in the introduction of this research. This news report has been altered to fit in with the experiment. Both the original as well as the altered news report are provided in appendix D and E. Every participant was exposed to the news release, in which it states that Off-White made an intense cultural inappropriate mistake in their new summer campaign. In the campaign, Caucasian models were photographed with cultural foods from Mexico. The context of the model posing with the food and wearing an Off-White shirt with the text: “Not black but mostly white, Off-White” (wink face) just created negative comments from a lot of people. To guarantee anonymity, names in the original news report are removed or altered in a way that commentators, spokespersons and other actors are unrecognizable.

16 report informed the participants about Off-White’s racial crisis and images were created to visually support the story. In relation to this news release, one of the 4 response scenarios were randomly presented to the participants. The three crisis responses were all designed according to the elements from the rebuild strategy of Coombs (2007). Response A consisted of the element apology, response B had the element regret, response C contained both an apology as well as signs of regret and response D was no response (control group). The 4 crisis responses had the same similar design and format, which is heavily influenced by the original response that Dolce and Gabbana provided after the crisis.

As described in a previous section, the difference between apology and regret is particularly small. A manipulation check was implemented in order to examine whether the response containing the elements apology, regret, apology and regret and no response were different to the participants. The manipulation question proposed to the participants was: ‘What response did you just read?’. The answer options were: apology, regret, apology and regret, and no response.

After the exposal of the condition, participants in each condition were asked to answer a few questions related to Off-White and Off-White’s brand image. In order to check whether the manipulation was successful, a Chi-square test of independence has been run with the variables ‘Crisis response manipulation’ and ‘manipulation check’. The result of the

manipulation test shows a significant difference between the 4 types of responses, (p < 0.01, Cramer’s V = 0.66). This means that the participants were able to identify the correct

17 TABLE 1

Manipulation check elements of rebuild strategy (N = 151)

Participants and procedure

To reach a wide variety of individuals, the link of the survey on Qualtrics had been shared on Facebook groups concerning fashion such as The New Fashion Society and student

association platforms. Moreover, it was shared on LinkedIn, as it is a non-commercial and fashion related platform, to reach individuals from all kinds of social and educational

backgrounds, who may know little or much about the brand. Moreover, for 3 days (Thursday, Friday and Saturday), the link to the experiment has been distributed in a large shopping mall in Amsterdam (the Bijenkorf). The official link to the experiment has been converted to a shorter link, which was easier for the warehouse visitors to use. The warehouse visitors could choose to fill in their survey on their own phones or they were able to fill in the survey by

18 using the portable notebook or iPad from the researcher. The reason to do the distribution in this shopping mall was that the Bijenkorf is the most popular warehouse of the Netherlands and they recently started to sell Off-White’s merchandise in stores (“De Bijenkorf

introduceert,” 2018). Another reason why the link has been distributed in a shopping mall is because of the ratio between social media participants (online) and ‘real-life’ participants (offline).

By entering the link in chrome or safari, participants were able to view the consent form that described a few important aspects such as: guaranteed Anonymity, the goal of the research and the auteur’s contact information. The participants who agreed on the terms described in the consent form were able to continue with the experiment. After the page that asked general questions about the participants’ age, country of origin, etc., a description followed that highlight the importance of the information and text provided during the experiment. Participants were asked whether they know the brand Off-White. Questions concerning brand identification were only visible for participants who indicated that they know Off-White. As mentioned before, participants were randomly assigned to the 3

experimental conditions (apology, regret, apology and regret) or the control group (Table 2). A thank you note appeared when participants finished the online experiment. The

19 TABLE 2

The distribution of conditions (N = 151)

By the time that the data was collected, a total of 189 participants cooperated in the experimental survey over a period of 3 weeks. After careful evaluation, some participants had to be excluded from this experiment. From the 189 responds, 8 of them were previews, 21 participants did not finish the questionnaire in its entirety and 9 participants were not qualified (age) for the research, which means a sample of 151 participants remained. After the removal of outliers, the youngest participant was 18 years old and the oldest participant was 36 years old. Ultimately, the response of 151 participants was used (N = 151).

Participants had a mean age of 23 years (SD = 3.23). More than half of the participants were female, 67.5% (n = 102) and 32.5% were male participants (n = 49). Although this research has been distributed on many international social media pages, many participants were from the Netherlands (n = 89). Moreover, Off-White’s target market is focussed on the millennial generation. Therefore the decision has been made to focus on participants aged 18 – 37 (Serafino, 2018). In addition to the demographic ‘gender’, ‘level of education’ was also taken into consideration. Individuals that hold a high school degree accounted for 18.8% (n = 28), 9.4% of the participants have an intermediate vocational education degree (n = 14), people

20 with a degree for University of Applied Sciences accounted for 14.8% (n = 22), the most responses came from individuals with a Bachelor degree, namely: 40.9% (n = 61), and 16.1% of the participants hold a graduate degree (n = 24).

First of all, a randomization check has been implemented for the demographics ‘gender’ and ‘level of education’, in order to validify whether the several conditions were

distributed equally around the groups. The Chi-square test was used for the variable ‘gender’ and ‘level of education’ and One-way ANOVA was conducted for the variable ‘age’. The conditions (apology, regret, apology and regret, and no response) were the independent variable, and gender, level of education and age were the depended variable. For gender (male, female and other) the Chi-square test showed there is no statistical significant difference between the conditions and the participants’ gender, X(3) = 2.342, p = .504. The Chi-square test was also conducted with the variable ‘level of education’, and showed no statistically significance between the conditions and the participants’ level of education, X(12) = 15.729, p = 0.204. For the variable ‘age’, a one-way ANOVA was conducted and the results show that there no statistically significant difference is found between the participants’ age and the conditions (F(3, 147) = .259, p = .855). The results from the randomization test show that the variables ‘gender’, ‘level of education’ and ‘age’ do not have to be controlled for.

Operationalization

In this experimental research, 3 hypothesis and 3 sub hypotheses (moderators) were formed and tested in order to answer the research question. The variables that need to be measured are also described in the literature section. The terminologies are explained comprehensively in the theory section and these formed a guiding principle throughout the questionnaires and the research.

21 As previously mentioned, brand image was supposed to be measured through three sub-variables, namely: brand experience, brand relationship, and brand community. However, the exploratory factor analysis indicated that for the variable ‘brand image’ (brand

experience, brand relationship, and brand community), the scale was two dimensional. The items of brand relationship and brand community loaded on one factor, which has been renamed to brand connection. Therefore, the variable ‘brand image’ will consist of two variables in this research, namely: brand experience and brand connection. Two factors had an Eigenvalue above 1.00. Factor 1 had an Eigenvalue of 11.04, which explained 64.92% of the variance in twelve items. Factor 2 had an Eigenvalue of 1.33, which explained 7.82% of the variance in five items. Important to note is that eventually factor 1 has 12 items and factor 2 exist of 5 items. The item: ‘I really feel Off-White items go with my way of life’ originally belongs to factor 2, but has a slightly higher loading for factor 1. Therefore, the decision has been made to include this item under factor 1. The variable ‘brand connection’ had 1 item that could increase the reliability. However, the difference in score was small and

theoretically, there was no reason to remove the item, so it has been decided to keep the item. Conforming to the factor and reliability analysis, brand experience and brand connection were the two components that were used in the analysis.

Brand experience. The dimension brand experience consist out of 5 items: ‘I find this

brand visually interesting’, ‘This brand makes strong impression on me’, ‘I have positive feelings towards this brand’, ‘I truly enjoy Off-White’s products’, ‘This brand intrigues me’.

These items are acquired from Brakus, Schmitt and Zarantonello (2009). The five items scale (brand experience) also proved to be reliable as indicated by a Cronbach’s Alpha of .90 (M = 3.63, SD = 1.33).

Brand connection. Brand connection consist out of 12 items: ‘I really feel Off-White

22 and my self-image are similar in a lot of ways’, ‘This brand says a lot about the kind of person I am or want to be’, ‘This brand reminds me of who I am’, ‘This brand is part of me’, ‘I love the collections of Off-White’, ‘I am proud of my Off-White items that I own’, ‘My White items are one of my favourite fashion possessions’, ‘I would recommend Off-White to my friends’, ‘Off-Off-White cares about my opinion’, ‘I feel a sense of kinship with

other people who own Off-White items’. The twelve items scale (brand connection) does meet the standard reliability of 0.60, as indicated by the Cronbach’s Alpha of .96 (M = 2.89, SD = 1.26). These scale items are acquired from Swaminathan, Page and Gürhan-Canli

(2007) and Schouten, McAlexander and Koenig (2007).

Brand identification. The variable brand identification is measured according to the

scale designed by Podnar (2004) and Stokburger-Sauer, Ratneshwar and Sen (2012). In this experimental research, the consumer’s identification scale will be a combination of the abovementioned researchers. Tuškej, Golob & Podnar (2013) applied Podnar’s scale items as well and they reported a Cronbach’s alpha of 0.82. The items from Stokburger-Sauer,

Ratneshwar and Sen (2012) and had an alpha of .94. The brand identification of participants were measured with the following scale items: 1. (‘I feel that my personality and the

personality of this brand are very similar’), 2. (‘I have a lot in common with other people using this brand’), 3. (‘I feel that my values and the values of this brand are very similar’), 4. (‘I feel a strong sense of belonging to the brand White’), 5. (‘I identify strongly with Off-White’), 6. (‘Off-white embodies what I believe in’), 7. (‘Off white is like a part of me’) and 8. (‘Off-White has a great deal of personal meaning for me’). The 7-point Likert scale is used with 1 (strongly disagree) and 7 (strongly agree). The levels 1, 2 and 3 are low identification levels and 4, 5, 6 and 7 are individuals who identify strongly with the brand “Off-White”. For this variable, an exploratory factor analysis indicated that the scale was unidimensional, as only one component had an Eigenvalue above 1.00. The component which had an

23 Eigenvalue of 4.41, explained 88.21% of the variance in the five items. The 5-items scale also proved to be reliable as indicated by a Cronbach’s α = .97 (M = 3.29, SD = 1.66).

Data analysis

A one-way analysis of variance (ANOVA) was used for this research, as to determine whether the 4 different conditions (apology, regret, apology and regret, and no response) yield different outcomes for a brand’s overall image (brand experience and brand

connection). The variable ‘brand identification’ was incorporated as a covariate, as it may influence the outcome of this experimental research. This was performed by means of Process (Hayes, 2013). In order to check and perform the assumptions for the analysis, mean scores are created for three variables: brand identification, brand experience, and brand connection.

Several assumptions had to be checked and ran before conducting the analysis. For the variable ‘brand experience’, there was 1 moderate univariate outlier found (case 12) and there were 2 moderate univariate outliers found for the variable ‘brand connection’ (case 64) and (case 139), as these scores are considered a possible outcome, these outliers will be retained.

Residuals were extracted in order to test the multivariate normality. The plots show that for conditions 3 and 4 (brand experience), there is a small violation. For brand

connection, condition 3 also show a small violation. Shapiro-Wilk test of normality showed small violations, however, the normal Q-Q plots show that the depended variables were normally distributed.

Homogeneity of variances is required and testing the variable brand experience with the Levene’s test for homogeneity of variances show that it meets the assumption, F(3, 147) = 1.09, p = .355. For the variable brand connection, F(3, 147) = 3.05, p = .031, the

24 homogeneity of variances violated. Accordingly, a Welch F test has been conducted, F(3, 80) = 1.90, p = .136

Results

To identify which crisis response from the rebuild strategy is the most effective for a luxury brand to use after a racial crisis have hit them, two analysis of variance have been conducted for H1a, H2a and H3a.

Brand experience. The means scores of the variables brand experience and brand

connection are displayed in Table 3. According to the results, it has been identified that there

is not a significant difference between the 4 conditions. The Analysis of Variance indicated that the 4 conditions (apology, regret, apology and regret and no response) did not have a significant effect on brand experience, F(3, 147) = 1.86, p = .140, η2= .04. Thus, a Bonferroni post hoc test is left out.

Brand connection. The mean scores of the variable brand connection are displayed in

Table 3. According to the results, it has been identified that there is not a significant

difference between the 4 conditions. The Analysis of Variance indicated that the 4 conditions (apology, regret, apology and regret and no response) did not have a significant effect on brand connection, F(3, 147) = 1.98, p =.119, η2= .04. Therefore, a Bonferroni post hoc test was not conducted.

The analysis show that there are no significant differences between the 4 conditions and how participants evaluate the brand image. In other words, a response containing apology does not generate more positive brand image than when other rebuild elements are used. This also applies to regret, apology and regret, and no response. Thus, hypothesis H1a, H2a and H3a are rejected.

25 TABLE 3

Mean scores brand identification, brand experience and brand connection after manipulation

(apology, regret, apology and regret, and no response) (N = 151)

The moderation effect of brand identification. Initially, it was expected that the relationship

between the conditions and post crisis brand image was moderated by brand identification. By analysing the moderation effect with Process macro in SPSS (model 1), the exact moderation effect was found (Hayes, 2013). The confidence interval of 95% and the 5000 bootstrap model was applied. For the independent variables (conditions), dummy variables were created.

Brand experience. First, the regression analysis was conducted for brand experience.

The overall model (F(7, 86) = 4.28, p < .001, R² = .26), demonstrate that 26% of the variance can be explained by the 4 different conditions (apology, regret, apology and regret, and no response) and brand identification.

26 (condition 2) was not a significant predictor, b = -.09, t(86) = -.11, p = .912, 95% CI [-1.63, 1.45]. Secondly, results indicate that the difference in apology (condition 1) and apology and regret (condition 3) was not a significant predictor, b = -1.38, t(86) = -1.67, p = .099, 95% CI [-3.03, 0.27]. Thirdly, results show that the difference in apology (condition 1) and no

response (condition 4) was not a significant predictor, b = -.30, t(86) = -.38, p = .710, 95% CI [-1.89, 1.29].

When analysing the interaction effect, it is evident that for interaction 1 (apology – regret – brand identification), no significance was found, b = .16, t(86) = .79, p = .431, 95% CI [-0.25, 0.57]. However, interaction 2 (apology and regret – apology – brand

identification), significance is found, b = .59, t(86) = 2.47, p = .016, 95% CI [0.11, 1.06]. Interaction 3 (no response – apology – brand identification) was found to be insignificant, b = .06, t(86) = .28, p = .778, 95% CI [-0.36, 0.48].

Interaction 2 (apology – apology and regret – brand identification) was found to be significant and thus further analysis was done. By analysing the participants with low levels of brand identification (1.6), comparison between apology and apology and regret, the effect is insignificant, b = -.44, t(86) = -.85, p = .395, 95% CI [-1.46, 0.58]. The participants with moderate levels of brand identification (2.9), also within the comparison between apology and apology and regret, the effect is insignificant, b = .32, t(86) = .88, p = .384, 95% CI [-0.41, 1.06]. The effect is significant for participants who had high level of brand

identification (5.56), within the comparison between apology and apology and regret , b = 1.89, t(86) = 2.75, p = .007, 95% CI [0.52, 3.25]. This means that people in condition 3 (apology and regret) compared to condition 1 (apology), who identify themselves high with the brand, reported a higher and thus a more positive brand experience.

Brand connection. The overall model F(7, 86) = 6.98, p < .001, R² = .36, demonstrate that 36% of the variance can be explained by the 4 conditions (apology, regret, apology and

27 regret, and no response) and brand identification.

Essentially, results show that the difference in apology (condition 1) and regret (condition 2) was not a significant predictor, b = -.30, t(86) = -.45, p = .656, 95% CI [-1.63, 1.03]. Secondly, results indicate that the difference in apology (condition 1) and apology and regret (condition 3) was not a significant predictor, b = -.92, t(86) = -1.29, p = .201, 95% CI [-2.35, 0.50]. Thirdly, results show that the difference in apology (condition 1) and no

response (condition 4) was not a significant predictor, b = -.29, t(86) = -.42, p = .678, 95% CI [-1.67, 1.09].

When analysing the interaction effect, it is evident that for interaction 1 (apology – regret – brand identification), no significance was found, b = .19, t(86) = 1.04, p = .298, 95% CI [-0.17, 0.54]. However, interaction 2 (apology and regret – apology – brand

identification), significance is found, b = .49, t(86) = 2.39, p = .019, 95% CI [0.08, 0.90]. The final interaction (no response – apology – brand identification) was found to be insignificant, b = .06, t(86) = .31, p = .760, 95% CI[-0.31, 0.42].

Interaction 2 (apology – apology and regret – brand identification) indicate

significant results and therefore further analysis was conducted. When looking at participants with low levels of brand identification (1.6), comparison between apology and apology and regret. The effect is insignificant, b = -.13, t(86) = -.30, p = .765, 95% CI[-1.02, 0.75]. The

participants with moderate levels of brand identification (2.9), the effect is also insignificant, b = .51, t(86) = 1.59, p = .116, 95% CI[-0.13, 1.14]. The participants who had high level of

brand identification (5.56), within the comparison between apology and apology and regret. The effect is significant, b = 1.82, t(86) = 3.06, p = .003, 95% CI[0.64, 3.00], which means that people in condition 3 (apology and regret) compared to condition 1 (apology), who identify themselves high with the brand, reported a higher and thus a more positive brand connection.

28 Ultimately, the different elements of the rebuild strategy are not a significant predictor and brand identification only show moderating effects between the groups apology and apology and regret. Hence, H1b and H2b are rejected, but H3b is partly accepted.

Conclusion and discussion

In response to the racial media crisis of Dolce and Gabbana in 2018, a research is conducted concerning the different elements of the rebuild strategy. The main aim of this research was to identify the differences between the elements of the rebuild strategy and to detect which element is the most important to include when a brand wants to reach the best possible brand image after a preventable media crisis.

First, the results show that the evaluation of brand image does not differ between the 4 different rebuild elements. This finding is not supported by Van der Meer and Verhoeven (2014), as their research established the fact that crisis response with emotions (regret) is experienced more effective than a response without regret. The insignificant findings could partly be related to the conceptualization of regret and apology. Previous inquiries indicated that apology can be conceptualized as showing regret and accepting full responsibility for the crisis that happened (Frandsen & Johansen, 2010). In other words, some scholars

conceptualize regret and apology as a similar concept, which could explain the insignificant findings, as the difference is difficult to detect. It could be concluded that the difference between apology and regret is rather a nuance than an actual difference, and therefore could be applied interchangeably in crisis response.

Coombs & Holladay (2008) proved that organizations try to avoid an official apology and instead provides signs of regret, as an apology is used as liable evidence in courtrooms and lawsuits. Organizations can, according to the results, provide crisis response expressing signs of shame and regret, and it will yield the same evaluation for brand image as when they

29 would provide an apology, without having the risk to lose lawsuits.

Second, in line with Frandsen and Johansen (2010), Coombs’ (2006) research

concludes that expressions of concern, compassion, or regret are rated equal to an apology. In other words, people experience an organization’s expressions of regret and guilt the same as when they provide an apology, and thus the evaluation of brand image does not differ when people are exposed to regret or apology, which explains the insignificant finding between the conditions apology and regret.

In addition, Lee’s finding (2005) is another perspective that could also explain the lack of differences between the apology and no response, as she declares that a mere apology is insufficient for an organization to recover from a crisis. This may explain why the element apology (response) yield the same evaluation of a brand’s image as a non-response, because an apology as response is not enough for a brand to recover its brand image from a crisis.

Results indicate that the moderating variable (brand identification) had a significant effect between apology and apology and regret. Brand identification plays a role in the evaluation of brand image, as it is proved that consumers who identify strongly with a certain brand, have more tendency to neglect negative information concerning the brand, and thus rate the brand image more positive than weak brand identifiers. This is also confirmed in prior research, as devotees of a brand appear to be strongly connected with the brand’s community and experience a more positive brand experience than individuals with a low brand identification (Cleeren, Dekimpe & Helsen, 2008; Einwiller, Fedorikhin, Johnson & Kamins, 2006; He & Li, 2011; Mishra, 2017). Moreover, the more a consumer identify and appreciate a brand, the more positive they are about the brand’s overall image after a media crisis, which confirms the theory that states that strong brand identification provides more insurance for a brand in a crisis situation (Johar, Birk & Einwiller, 2010; Lisjak, Lee & Gardner, 2012). In other words, corporations in crisis situations can use this finding and

30 incorporate it into their post-crisis response strategy.

In conclusion, there are only differences found for the elements apology and apology and regret and the positivity level of brand image, moderated by a high level of brand

identification. This means that consumers who are exposed to a crisis response containing apology and regret will evaluate more positive brand image levels than when an apology is given. This relationship is dependent on the strength of brand identification.

Limitations and future research

Despite the careful and thorough consideration of designing and conducting this experimental research, there are limitations that can influence the research negatively. The sample size of the population (N = 189) could partly explain the insignificant findings of this study. Each of the 4 conditions contained one message. Perhaps, one response per condition is too limited and has little effect on how consumers perceive the brand after a crisis.

The results of the manipulation test show that the conditions were successfully manipulated. However, in Table 1 it becomes apparent that 61.5% of the participants who were exposed to the condition regret, indicated that they were exposed to the condition apology and regret. This error could explain the insignificant test results, as participants were

not able to distinguish the difference between regret and regret and apology in the provided response.

Moreover, the insignificant difference in the level of the brand image when an organization provides a response (apology, regret, apology and regret) or no response, moderated by brand identification, contradicts prior studies. This could be explained by the following; consumer scepticism concerning a brand after a crisis. Future research can look into how scepticism affects the acceptance of a crisis response from an organization. In short, whether it is an apology, regret or apology and regret, it will not yield a better rating of brand

31 image than when no response is provided, because of the high level of scepticism among consumers and clients concerning a brand after a crisis. Moreover, research may look into the correlation of specific emotional speech and emotional words on paper and the scepticism level. Therefore, corporations know what emotions and words of emotions to use when crisis happens.

This research provided an experimental survey in English, which automatically excluded participants who do not understand, read or write English. Although the Dutch population has a high level of understanding the English language, this should be taken into account when designing and conducting a survey. In relation to this, many participants that were approached in the shopping mall indicated that they preferred to fill in the experimental survey on paper. First and foremost, this was not possible due to the randomization of

conditions and secondly it would be a time-consuming process to manually insert the surveys into Qualtrics.

Although prior knowledge and feelings about Off-White was partly controlled for by implementing the variable brand identification as a moderator, it should be acknowledged that by using an existing fashion brand (Off-White) as a stimulus, can evoke biased and not necessarily genuine opinions about a brand. For future research, it is important to note that this particular research used Off-White, which was indicated to be highly popular amongst individuals from the age 16 – 37. It would be highly recommended for scholars to look into brands that already have a bad reputation and brand image and see how the audience will react on their crisis response. In addition, this brand does not have a crisis history (yet). It is interesting for scholars to look into brands who have a history of crises and identify how the audience will respond to their crisis response. Lastly, this research provided a response in a form of an official statement. Therefore, emotional cues could be overlooked. Future research could focus on the difference between providing a response (apology, regret, apology and

32 regret) in an official written statement, on social media, or in a video where nonverbal

communication is emphasized.

Moreover, more than half of the participants had a bachelor degree (university) or higher. The educational background can influence the outcome of the study, as the results from the study cannot be generalized for the average individual between the age of 18 - 37.

Theoretical and practical implications

Reflecting on the research, it appears that brand identification has a considerable effect on brand image. Individuals who experience a strong and positive connection with the brand will evaluate a brand in crisis situation more positively than individuals who have low

identification with a brand. Previous studies confirm the positive relationship between brand identification and brand image, which is a valuable insight for organizations. In crises situations, brands can decide to focus their crisis response on the clients and consumers who have a strong identification with the brand. The strong brand identifiers are more likely to exhibit positive evaluations and could steer the offline and online conversation towards a positive direction.

However, prior research does prove that strong brand identification could also backfire a brand in crisis. It highly depends on the nature of the crisis and whether the

negative information attacks the consumer’s ideals, values and believes, as these are the base for brand identification. Future research can focus on the focal point (the top of the parabola) of the positive relationship of the covariate brand identification and the variable brand image and see to what extent it will reverse and change from a positive relationship to a negative relationship. This could assist organizations and brands to design communication strategies based on the severity of the crises and the level of brand identification of clients and

33

References

Aggarwal, P. (2004). The effects of brand relationship norms on consumer attitudes and behavior. Journal of consumer research, 31(1), 87-101. DOI: 10.1086/383426 Agnihotri, A., & Bhattacharya, S. (2019). Lactalis: The product-harm and product recall

crises. SAGE Publications: SAGE Business Cases Originals.

Astor, M. (2017). Dove drops an ad accused of racism. Retrieved from the New York Times online: https://www.nytimes.com/2017/10/08/business/dove-ad-racist.html

Baker, G. F. (2001). Race and reputation: Restoring image beyond the crisis. Handbook of public relations, 513-520.

Benoit, W. L. (1997). Image repair discourse and crisis communication. Public relations review, 23(2), 177-186. DOI: 10.1016/S0363-8111(97)90023-0

Berthon, P. R., Pitt, L. F., Plangger, K., & Shapiro, D. (2012). Marketing meets Web 2.0, social media, and creative consumers: Implications for international marketing strategy. Business horizons, 55(3), 261-271. DOI: 10.1016/j.bushor.2012.01.007 Beyer, J., Bierling, K., & Öjerbrant, F. (2018). Case H&M-(Un) coolest Monkey in the

Jungle-When Ads Turn Against Their Creators. LBMG Corporate Brand Management and Reputation-Masters Case Series.

Bhattacharya, C. B., & Sen, S. (2003). Consumer-company identification: A framework for understanding consumers’ relationships with companies. Journal of Marketing, 67, 76–88. DOI: 10.1509/jmkg.67.2.76.18609

Bouaine, W., Charfeddine, L., Arouri, M., & Teulon, F. (2015). The influence of CEO departure and board characteristics on firm performance. Journal of Applied Business Research, 31(2), 345.

Brakus, J. J., Schmitt, B. H., & Zarantonello, L. (2009). Brand experience: what is it? How is it measured? Does it affect loyalty?. Journal of marketing, 73(3), 52-68. DOI:

10.1509/jmkg.73.3.052

Brocato, E. D., Peterson, R. A., & Crittenden, V. L. (2012). When things go wrong: Account strategy following a corporate crisis event. Corporate Reputation Review, 15(1), 35 51. DOI: 10.1057/crr.2011.24

Chang, P. L., & Chieng, M. H. (2006). Building consumer–brand relationship: A cross cultural experiential view. Psychology & Marketing, 23(11), 927-959. DOI: 10.1002/mar.20140

34 Claeys, A. S., Cauberghe, V., & Leysen, J. (2013). Implications of stealing thunder for the

impact of expressing emotions in organizational crisis communication. Journal of Applied Communication Research, 41(3), 293-308. DOI:

10.1080/00909882.2013.806991

Cleeren, K., Dekimpe, M. G., & Helsen, K. (2008). Weathering product-harm crises. Journal of the Academy of Marketing Science, 36(2), 262-270. DOI: 10.1007/s11747-007 0022-8

Cooley, S. C., & Cooley, A. B. (2011). An examination of the situational crisis

communication theory through the general motors bankruptcy. Journal of media and communication studies, 3(6), 203-211.

Coombs, W. T., & Holladay, S. J. (2001). An extended examination of the crisis situations: A fusion of the relational management and symbolic approaches. Journal of public relations research, 13(4), 321-340. DOI: 10.1207/S1532754XJPRR1304_03

Coombs, W. T., & Holladay, S. J. (2002). Helping crisis managers protect reputational assets: Initial tests of the situational crisis communication theory. Management

Communication Quarterly, 16(2), 165-186. DOI: 10.1177/089331802237233 Coombs, W. T. (2006). The protective powers of crisis response strategies: Managing

reputational assets during a crisis. Journal of promotion management, 12(3-4), 241 260. DOI: 10.1300/J057v12n03_13

Coombs, W. T. (2007). Protecting organization reputations during a crisis: The development and application of situational crisis communication theory. Corporate reputation review, 10(3), 163-176. DOI: 10.1057/palgrave.crr.1550049

Coombs, W. T., & Holladay, S. J. (2008). Comparing apology to equivalent crisis response strategies: Clarifying apology's role and value in crisis communication. Public Relations Review, 34(3), 252-257. DOI: 10.1016/j.pubrev.2008.04.001

Coombs, W. T. (2010). Conceptualizing crisis communication. In Handbook of risk and crisis communication (pp. 111-130). Routledge.

Coombs, W. T. (2015). The value of communication during a crisis: Insights from strategic communication research. Business Horizons, 58(2), 141-148. DOI:

10.1016/j.bushor.2014.10.003

Dawar, N., & Pillutla, M. (2000). Impact of product-harm crises on brand equity: The moderating role of consumer expectations. Journal of Marketing Research, 37(2), 215–226. DOI: 10.1509/jmkr.37.2.215.18729

35 De Bijenkorf introduceert nieuwe merken voor Lente/Zomer 2018. (2010) Retrieved from

Bijenkorf online: https://www.debijenkorf.nl/press/2018/introductie-nieuwe-merken lente-zomer-2018.html

Deitch, E. A., Barsky, A., Butz, R. M., Chan, S., Brief, A. P., & Bradley, J. C. (2003). Subtle yet significant: The existence and impact of everyday racial discrimination in the workplace. Human Relations, 56(11), 1299-1324. DOI: 10.1177/00187267035611002 Demopoulos, A. (2018). Dolce & Gabbana never met a controversy they didn’t like – until

now. Retrieved from Daily beast online: https://www.thedailybeast.com/dolce-and gabbana-never-met-a-controversy-they-didnt-like-until-now

Dolce and Gabbana cancels shanghai fashion show amid racism accusations, 2018. Retrieved from BBC online: https://www.bbc.com/news/entertainment-arts-46288884

Einwiller, S. A., Fedorikhin, A., Johnson, A. R., & Kamins, M. A. (2006). Enough is enough! When identification no longer prevents negative corporate associations. Journal of the Academy of Marketing Science, 34(2), 185-194. DOI: 10.1177/0092070305284983 Esch, F. R., Langner, T., Schmitt, B. H., & Geus, P. (2006). Are brands forever? How brand

knowledge and relationships affect current and future purchases. Journal of Product & Brand Management, 15(2), 98-105. DOI: 10.1108/10610420610658938

Ewing, J. (2015). Volkswagen C.E.O Martin Winterkorn resigns amid emissions scandal. Retrieved from:

https://www.nytimes.com/2015/09/24/business/international/volkswagen-chief martin-winterkorn-resigns-amid-emissions-scandal.html

Exline, J. J., Deshea, L., & Holeman, V. T. (2007). Is apology worth the risk? Predictors, outcomes, and ways to avoid regret. Journal of Social and Clinical Psychology, 26(4), 479-504. DOI: 10.1521/jscp.2007.26.4.479

Frandsen, F., & Johansen, W. (2010). Apologizing in a globalizing world: crisis

communication and apologetic ethics. Corporate Communications: An International Journal, 15(4), 350-364. DOI: 10.1108/13563281011085475

Givhan, R. (2018). Dolce and Gabbana are rich, extravagant and politcally incorrect. You got a problem with that? Retrieved from The Washington Post online:

https://www.washingtonpost.com/lifestyle/style/dolce-and-gabbana-are-rich extravagant-and-politically-incorrect-you-got-a-problem-with

that/2018/05/22/63caaa8c-41b0-11e8-ad8f

36 Green, T. K. (2003). Discrimination in workplace dynamics: Toward a structural account of

disparate treatment theory. Harv. CR-CLL Rev., 38, 91.

Grothaus, M. (2017). Samsung’s CEO resigns due to “unprecedented crisis” at company. Retrieved from fast company online:

https://www.fastcompany.com/40481057/samsungs-ceo-resigns-due-to unprecedented-crisis-at-company

Hareli, S., & Eisikovits, Z. (2006). The role of communicating social emotions accompanying apologies in forgiveness. Motivation and Emotion, 30(3), 189-197. DOI:

10.1007/s11031-006-9025-x

Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford Publications.

He, H., & Li, Y. (2011). CSR and service brand: The mediating effect of brand identification and moderating effect of service quality. Journal of Business Ethics, 100(4), 673-688. DOI: 10.1007/s10551-010-0703-y

Heath, R. L. (1998). Crisis Management for Managers and Executives: Business Crises: the Definitive Handbook to Reduction, Readiness, Response and Recovery. Financial Times Management.

Howell, A. J., Turowski, J. B., & Buro, K. (2012). Guilt, empathy, and apology. Personality and Individual Differences, 53(7), 917-922. DOI: 10.1016/j.paid.2012.06.021 Janssen, C., Sen, S., & Bhattacharya, C. B. (2015). Corporate crises in the age of corporate

social responsibility. Business Horizons, 58(2), 183-192. DOI: 10.1016/j.bushor.2014.11.002

Johar, G. V., Birk, M. M., & Einwiller, S. A. (2010). How to save your brand in the face of crisis. MIT Sloan Management Review, 51(4), 57-64.

Kim, H. J., & Cameron, G. T. (2011). Emotions matter in crisis: The role of anger and sadness in the public’s response to crisis news framing and corporate crisis response. Communication Research, 38(6), 826-855. DOI: 10.1177/0093650210385813 Kim, H. S. (2011). A reputational approach examining public’s attributions on corporate

social responsibility motives. Asian Journal of Communication, 21(1), 84-101. DOI: 10.1080/01292986.2010.524230

Kim, S., Avery, E. J., & Lariscy, R. W. (2009). Are crisis communicators practicing what we preach?: An evaluation of crisis response strategy analyzed in public relations

research from 1991 to 2009. Public Relations Review, 35(4), 446-448. DOI: 10.1016/j.pubrev.2009.08.002