“The Dark Side of Customer Engaging Behaviour”: How the Dark Side Impacts a Brand’s Image.

“The Dark Side of Customer Engaging Behaviour”:

How the Dark Side Impacts a Brand’s Image.

by

Booy van Hees

Master thesis

MSc Marketing

How the Dark Side Impacts a Brand’s Image.

by

Booy van Hees

Papengang 2

9711 PA Groningen

06-24407704

booyvanhees@hotmail.com

Student Number: 1740792

Supervisors:

Prof. dr. P.C. Verhoef

Prof. dr. B. M. Fennis

Groningen 2013

University of Groningen

Management Summary

Companies, brands and their marketing managers are facing a fast changing environment. These changes in the environment are driven by changes in technology, the evolution of the internet and the emerging importance of social media. This has led to a shift in control of marketing decisions to customers (Sashi, 2012), and thereby increased peoples interest in customer engaging behaviour. Examples of customer engaging behaviour are word-of-mouth, writing reviews, blogging, writing recommendations, complaining to firms etc. (Van Doorn et al, 2010). The rapid growth of social media usage has extended consumers options to gather information about products, before making a purchase decision. Customer engaging behaviour can be either positive or negative. This study tries to discover what the effects are of negative customer engaging behaviour, which is called ‘the dark side’. The focus is on bash actions. Hence, the following problem statement is formulated:

How does a bash action impact a brand’s image, and what role might a company’s response, source identification and brand identification hereby play?

A bash action is executed when a social media platform like Facebook or Twitter is exploited by customers, to express their complaints and bad experiences with the firm. As stated by Jindal & Liu (2008) these bash actions might be executed to damage a firm’s reputation. Thus, bash actions might have a negative effect on brand image. An independent t-test and multiple

regression has indeed found significant evidence for this negative relation. So, the level of brand image decreases when their customers are exposed to a bash action. In addition, three moderating effects are studied. These are (1) the companies’ response, (2) identification with the source, and (3) identification with the brand. Only for the third moderator significant evidence were found. The negative effect of the bash action appears to be weakened, when the customer that is

exposed to the bash action has a high identification with the brand that the bash action is targeted on. For the other 2 moderators no significant evidence was found.

Keywords: customer engaging behaviour, social media, bashes actions, companies’ response,

Preface

With this thesis, I completed my master Marketing Management at the University of Groningen. After six months of hard work, it is now finalized. My life as a student will be ended. However, a working class hero is something to be.

I would like to thank my supervisor prof. dr. Peter Verhoef, for the fun meetings, and for his very useful feedback and guidance during the writing of my thesis.

Additionally, I would like to thanks my roommates for their support during the writing of this thesis. Last but not least, a big thanks goes out to my friends and family. In specialto my parents, for their unconditional support and help during this thesis.

Table of Contents

1. Introduction ...7

1.1 Background ...8

1.2 Theoretical and Managerial Relevance...8

1.3 Research Question ...9 1.4 Structure of Thesis ...9 2. Theoretical Framework...10 2.1 Conceptual Model...10 2.2 Social Media...11 2.3 Bash Actions...12 2.4 Brand Image...14 2.5 Moderating Effects...16 3. Methodology...21 3.1 Experimental Design ...21 3.2 Survey Design...22 3.3 Participants ...25 3.4Stimuli ...25 3.5 Procedure...26 3.6 Plan of Analysis ...26 4 Results...29 4.1 Demographics ...29 4.2 Factor Analysis ...31 4.3 Descriptives...35 4.4 Manipulation Check ...37

4.5 Testing for Multicollinearity...38

4.6 Hypothesis Testing...40

5 Discussion...44

5.1 Conclusion Main Effect ...45

5.2 Conclusions Moderating Effects ...45

5.3 Limitations and Future Research ...47

6 References ...49

7 Appendix ...55

Appendix A: Questionnaire...55

1. Introduction

“People love to talk. People love to talk about products and services. People talk about hair color, cars, computers, sandwiches, TV shows, and floor cleaners. The stuff they use every day. People talk about what you sell right now. It might be a causal mention. It might be a scathing attack. It might be a scathing attack posted on Amazon, where 20 million people will read it before deciding whether to buy your stuff. Or – it might be something really nice.”

A striking start of the book ‘Word of Mouth Marketing’, written by Andy Sernovitz (2012). Striking because it exactly indicates how much people love to talk about their experiences with products, and how massive the impact could be for a company’s sales. It also declares why customer engagement is such a hot topic these days. The increasing interest in customer engagement has been driven by the continued evolution of the internet and emerging importance of social media. Social media has ensured a shift in control of marketing decisions to customers (Sashi, 2012). The saying “customer is king”, might have never been as true as nowadays. In fact, they are probably even more then king. Nowadays customers are the judge, jury and executer of a company’s reputation and success.

Because of this rising power of the customer, customer engagement became an important construct. Customer engagement can be defined as a behavioral manifestation, beyond purchase, from a customer towards a brand / firm (van Doorn, Lemon, Mittal, Nass, Pick, &Verhoef, 2010). This results in customer engaging behavior (hereafter CEB), which includes complaints to firm, word-to-mouth activity, writing reviews, blogging, recommendations, etc.

CEB can be both positive (i.e., posting a positive brand message on a blog) and negative (i.e., organizing public actions against a firm) (Van Doorn et al., 2010). Earlier literature has predominantly focused on the positive expressions of engagement. In this study, the focus is on the ‘dark side’ of CEB and what the consequences for a company’s brand image could be.

1.1 Background

There are many examples that indicate what is meant by the dark side of CEB. A company starts a platform, for example a twitter account or Facebook page, to enable the customer to engage. In contrast to what the company expects, the customer exploits this platform to start express their complaints and bad experiences, or to make a parody public. In this research these forms of the dark side of CEB are named ‘bash actions’.

The range of these bash actions is enormous, and can be reached fast. To illustrate this Intel has researched what happens on social media in one internet minute (Intel, 2012). Each minute 100.000 tweets are published and 320 people make a Twitter account, 277.000 people Login op Facebook leading to 6 million Facebook views, furthermore 1.3 million YouTube videos are watched, and 30 hours of video is uploaded. As a result word-of-mouth has increased enormously: consumers can access these social media from all over the world, so it reaches far beyond a local community (Chen & Xie, 2008).

According to Jindal & Liu (2008) people can either write positive reviews to promote the target objects, or write malicious negative reviews for the target objects to damage their reputation. The latter might be the reason why people take part in negative engaging behavior. But does it indeed damage a firm’s brand image?

1.2Theoretical and Managerial Relevance

This thesis is relevant for marketers, which focus on online word-of-mouth and how this can damage a brand’s image. The expanding online environment increasingly enables customers to influence larger groups of other customers with their positive or negative word of mouth.

1.3 Research Question

When we take the prior in consideration, the following questions remain: what is the chance that an intended positive social media campaign backfires into a negative one? Should companies react to this dark side of CEB? Are consumers influenced by these bash actions? So the main question of this thesis is:

How does a bash action impact a brand’s image, and what role might a company’s response, source identification and brand identification hereby play?

To cope with a clear picture on what subjects the study is about, this problem statement is split into several underlying research questions:

1. What is the definition of social media? 2. What is meant by bash actions? 3. What is the definition of brand image?

4. What is the moderating role of the response of the company on bash action’s effect on

brand image?

5. What is the moderating role of identification with the source of the bash action’s effect on

brand image?

6. What is the moderating role of identification with the brand towards the bash action is

directed on brand image?

1.4 Structure of Thesis

2. Theoretical Framework

In order to answer the above outlined research question, a theoretical framework is needed. This chapter will start with outlining the conceptual model used for the thesis. Followed by an exploration of Social Media in general. In section 2.3 examples will clarify how the use of Social Media might backfire into certain bash actions, and in section 2.4 the relationship with brand image is defined. Section 2.5 will illustrate what factors might have a moderating effect on this relationship.

2.1 Conceptual Model

This research examines how bash actions may affect brand equity, and how a company’s response, and the identification with the source of action, or with the brand itself moderates this relationship. Figure 2.1 explains these relations in the form of a conceptual model.

2.2 Social Media

The introduction of the internet can be seen as the major change driver in the ways people think about marketing. The last decennia of the 20th century became known as the decade of ‘e-commerce’, which has developed into the era of ‘social commerce’ in the 21stcentury. (Leeflang, Verhoef, Dahlström, and Freundt, 2013). Consumers are no longer passive receivers of information from marketers, instead they interact with each other and the company to influence consumer purchasing and company decision making (Fader & Winer, 2012). These interactions take the form of the earlier mentioned CEB. Because of social media non-transactional customer behavior is likely to become more important in the near future (Verhoef, Werner, and Krafft, 2010).

Social media can be defined as online tools where content, opinions, perspectives, insights, and media can be shared (Parsons, 2011), and at its core, social media is about relationships and connections between people and organizations (Nair, 2011). The best-known sites of social media include YouTube, Twitter, and Facebook. These media can carry content in the form of words, text, pictures and videos, and is generated by millions of consumers around the world. The social influences through these social networks can be tremendous (Majumdar, & Krishna, 2012). By communicating in their peer network, individuals influence each other, which extends to other people of other networks.

The dynamic and interconnected environment that companies are facing has resulted in two major effects. Companies do not fully control the messaging they use to create brand strategies anymore (Fader & Winer, 2012), and therefore the focus of power shifts away from the company to the customer (Berthon et al, 2012). This may lead to certain risks for companies. Where consumers once voiced their dissatisfaction with a firm to a few family members or friends, they now take their complaints to the mass media, which makes them available to the public (Ward & Ostrom, 2006). Hence, customers who are dissatisfied by the company’s offering may decide to engage in virtual complaints, which could potentially be damaging for a firm (Kaplan & Haenlein, 2010).

2.3 Bash Actions

The social media environment has led to certain success, but also disaster stories for companies. It could become a disaster story when for example virtual complaints or parodies, take the form of bash actions against a firm. Reasons for customers to start bash actions might be a feeling of betrayal. Fitness (2001) defines this as when a firm sends out a signal about how little it cares about, or values its relationship with, the consumer. Another reason might be when consumers’ complaints are repeatedly ignored or inadequately addressed (Ward & Ostrom, 2006). However, these bash actions have one thing in common: it's a result of the shift in power away from the firm towards the consumer (Berthon et al, 2012). This is a consequence of the fact that social media is a hybrid element: marketing managers cannot control the content and frequency of the information spread (Mangold & Faulds, 2009).

commercials for the brand and show the best five on television. However, H.J. Heinz received a lot of submissions that they did not want to associate their brand with. These submissions were rejected, but did find entrance on to YouTube and became a source of parodies. Furthermore, visitors to Heinz’s website wrote messages that they were looking for cheap labor, or that they are lazy to ask consumers to do its marketing work.

2007

2008

Canadian musician Dave Carroll experienced how his guitar was broken during a trip with

United Airlines

Greenpeace decided to go viral with their campaign against

Nestlé

2010

BP

2010

McDonald's

These examples make clear that social media is not a channel for just distributing corporate information: it allows consumers to interact and participate with companies, and allows them to share their opinions with others which influence corporate reputations (Parsons, 2011).

2.4 Brand Image

A consequence for organizations of such bash actions is that it might have lasting and strong effects on brand equity (Fournier & Avery, 2011); therefore it is very relevant to research if these effects really exist. Brand equity can be defined as the differential effect of brand knowledge on consumer response to the marketing of the brand (Keller, 1993). Brand knowledge consists of two components: brand awareness and brand image. This study will focus only on the effects of bash actions on brand image. Brand Image refers to the set of associations linked to the brand that consumers hold in memory (Keller, 1993). The founders of the concept of brand image are Gardner and levy. They were the first to suggest that products have both a social and psychological nature, as well as a physical one (Gardner & Levy, 1955). So in addition to the product itself, peoples purchase choice can also be based on the image of the brand, and the sets of feelings and attitudes that consumers possess towards the brand.

Ever since, brand image became an important concept for marketers. A more modern definition is developed by Cretu and Brodie (2007). According to them brand image can be defined as a mental picture of the product/service which is offered, andit includes symbolic meanings which are associated with the specific attributes of the product or service. However, they are not just symbolic meanings. Brand image consists also of the kinds of users that are typically associated with the product, its employee, packaging details, product category associations, advertising message and style, price, distribution channel, and so forth (Batra & Homer, 2004).

2.4.1 Outcomes Bash Actions on Brand Image

Social media has become a major factor in influencing various aspects of consumer behavior, including awareness, opinions and attitudes (Mangold & Faulds, 2009). Different types of social media might possibly achieve different marketing goals. Social network sites like Facebook and YouTube, and micro blogging sites like Twitter are effective for influencing consumer beliefs and attitudes (Parsons, 2011). The actions of an organization on these media also affect organizational image and identity (Majumdar & Krishna, 2012). Van Reijmersdal, Neijens and Smit(2007) explain the possible effects on brand image by the human association’s theory of Anderson and Bauer (1973). This theory states that people’s specific associations and evaluations can change over time because people are connected and influence each other through networks. As a bash action is executed online through a social network, this theory might indicate that it may have severe effects on a brand’s image.

In addition, online shoppers put twelve times more trust in peers’ opinions than in marketer-initiated sources (Ludwig, de Ruyter, Friedman, Brüggen, Wetzels, and Pfann, 2013). Therefore bash actions might have an even greater negative impact on a brand’s image when it is executed by a source where people can identify with.

Attackers can manipulate a reputation system to disseminate damaging contents (Braun et al, 2012). These contents might be extra damaging because of informational cascade. Informational cascade is a phenomenon that occurs when it is optimal for an individual, having observed the actions of those ahead of him, to follow the behavior of the preceding individual without regard to his own information (Bikhchandan et al, 1992). Because bash actions use social media as a channel to make their damaging content public, informational cascade ensures that a potential enormous amount of people might see the information as true and thereby follow the action, which damages the brand’s image.

In general, bash actions lead to electronic word-of-mouth. This has a critical impact on a brand’s image (Jansen et al, 2009). Based on the studies above we might assume that it has a negative impact on image. Therefore, the following hypothesis is suggested:

2.5 Moderating Effects

As mentioned above, the assumption can be made that bash actions have a negative effect on brand image. The impact of these bash actions might be strengthened or weakened by certain moderating effects. These effects are the response of the company and identification with the brand and source.

2.5.1 Response of the Company

Blodgett,Wakefield and Barnes (1995) showed that the way a company handles complaints has an effect on loyalty of the customer. Companiesthat develop a reputation for not being willing to remedy their customers’complaints may slowly lose many of their customers. Although bash actions go further than just complaining, it might be likeable that the way a company responses to these actions impact its effects on their brand image.

Bash actions might require externally visible actions and reactions from a firm as well (Van Doorn et al, 2010). In responding, a company can choose to give in to the bash action, or it can actively fight back (Fournier, 2012). When we position these possible responses on the continuum developed by Siomkos & Kurzbard (1994), ‘fighting back’ can be placed on the ‘denial’ end, and ‘giving in can be placed at the super-effort end. Thus, companies might benefit from simply attempting recoveries (Voorhees, Brady, and Horowitz,2006), and in that way they might turn negative CEB into positive CEB.

An example of a case that illustrates different responses to a bash action is of Mentos and Diet Coke. In 2006 a video of two guys appeared on YouTube where they demonstrated what happened when both are mixed. Lots of people copied this, and as a result funny videos of people shooting soda were passed on from consumer to consumer. Diet Coke could not appreciate the humor of the videos, and fought back because it did not fit with its brand’s personality. Mentos, however, saw the positive side of this free publicity and decided to reach out and sponsor the two creators (Young, 2009).

DellHell is a great example of how you can turn negative CEB into something positive. Blogger Jeff Jarvis posted a series of complaints, coined “Dell Hell”, about his Dell laptop. His posts caught the attention of others, who followed his example and posted their own negative experiences with Dell’s customer service. Despite the bad press and negative blogging, Dell remained silent, and even closed its online consumer forum. This resulted in droppings of their sales and reputation. After a year, Dell decided to work together with Jarvis and launched its own blog: Direct2Dell. As a result, after a year 23% of online conservation about Dell was negative, instead of earlier 50% (Williams, 2009).

These examples, and the research performed by Siomkos & Kurzbard (1994) illustrate that the way a company reacts to a bash action might have an impact on how damaging the action is on the brand’s image. Therefore, the following hypotheses are suggested:

H2b: Fighting back will strengthen the negative effect of a bash action on brand image

H2c: Giving in will weaken the negative effect of a bash action on brand image

2.5.2 Identification with the Source and Brand

As mentioned earlier, bash actions might have a greater negative impact on a brand’s image when it’s executed by a source where people can identify with. The source of a bash action might differ per situation. The earlier mentioned Nestlé bash was executed by the environmental activists group Greenpeace. The United Airlines example on the other hand, was executed by a regular customer. The assumption might be made that these different sources impact the effect of the bash action on brand image in a different manner. It has been demonstrated that trustworthy and expert spokespeople induce a greater positive attitude towards the position they advocate (Sternthal, Dholakia, and Leavitt, 1978). Being trustworthy is evaluated by two concepts (Wang & Vassileva, 2003). Firstly by trust, this is defined as a peer’s belief in another peer’s capabilities, honesty and reliability based on its own direct experiences. Secondly by reputation, this is defined as a peer’s belief in another peer’s capabilities, honesty and reliability based on

recommendations received from other peers.

This might indicate that a bash action has more effect when people see the source as trustworthy or expert, and therefore can identify themselves with the source. Support can be found in the theory of homophily. The theory of homophily is rooted in antiquity as the ancient Greeks already stated that people love those who are like themselves (McPherson, -Lovin, and, 2001). This means that “birds of feather, flock together”, i.e. homophily is the tendency of individuals to associate and bond with similar others. So an assumption can be made that when an individual has started the bash action, like the United Airlines example, the matter in which the receiver thinks this individual is like him or her will influence the impact of the action on the perceived image of the brand.

matter of collective self-construal, ‘we’ and ‘us’, against ‘them’ (Hogg, 2006). This stresses that being member of a group seems sufficient for them to act as a group. On the basis of the group membership of the social identity theory, an assumption can be made that when a person identifies him- or herself as member of a group, he or she might engage with the thoughts and actions where the group stands for. On the other hand, if a person does not identify himself as being a member of a group as Greenpeace for example, then he will probably have little belief in their capabilities, honesty and reliability, and will not think they are trustworthy. For this reason the bash action started by the group will probably have little impact on how he thinks about a certain company.

Thus, concluding: the impact of the action might be larger when it is started by a regular customer(s), expert or group, with whom the receiver can identify and therefore think is trustworthy. Therefore, the following hypothesisis suggested:

H3: Identification with the source will strengthen the negative effect of a bash action on brand image

Next to identification with the source, identification with the brand may also play a moderating role in the relation between bash actions and brand image. Keller (2007) defines this as brand resonance: the level of identification that a customer has with a brand and the extent to which the customers feel that they are “in sync” with the brand. Brand resonance can be divided into four categories: behavioral loyalty, attitudinal attachment, sense of community and active engagement. There are certain brands (e.g. Coca-Cola, Apple) that possess high brand equity, resulting in higher market shares then their competitors (Badenhausen, 1996), this is due to the fact that they have high customer loyalty and other assets (Aaker, 1991). On the other hand, firms often fail to build strong customer loyalty because they are unable to create strong emotional bonds with their customers (Yim, Tse and Chan, 2008).

more resilient to negative information, and forgiving of the company’s mistakes (Bhattacharya & Sen, 2003). Customers reject any negative information they may receive about a company (or its products) they identify with. In the case of bash action this means we might suggest that identification with the brand leads to rejection of these actions.

Next to developing loyalty, trust plays another important role in developing identification. Relating to brands, trust is defined as a feeling of security held by the consumer that the brand will meet his/her consumer expectation (Delgado & Munuera, 2001). Its role is important as it is a key variable for the development of brand loyalty (Delgado & Munuera, 2005), which again, is necessary for identifying with the brand. One of the main marketing advantages of brand loyalty is a greater resistance among loyal consumers to competitive strategies (Chaudhuri & Holbrook, 2001). It is likely to assume that this greater resistance also applies to bash actions. Therefore, the following hypothesis is suggested:

3. Methodology

To test the hypotheses stated in chapter two, quantitative analysis has been performed. Quantitative research can give a good insight in the opinion or attitudes of large amounts of people (Malhotra, 2009). For this study this implicates a good understanding into opinions and attitudes of the respondents towards the source of the bash action and towards the brand which it is directed. In the following chapter, the experimental and survey design, and the procedure of the two studies conducted for this paper will be discussed.

3.1 Experimental Design

According to Malhotra (2009), an experiment is formed when the researcher manipulates one or more independent variables and measures their effect on one or more dependent variables, while controlling for the effect of extraneous variables. For this study the independent variables ‘Identification’ and ‘Companies Response’ are manipulated to measure their effect on the dependent variable ‘brand image’.

In the first study, the source of the bash action is PowNed. In the second study, the source is a group of students of the RUG. For both studies the bash action is directed towards McDonalds. Therefore for both studies, one and two, there is a three (no response, McDonalds fighting back, McDonalds giving in to bash action) versus two (bash action initiated by PowNed, initiated by group of students) experimental design performed to test the hypotheses. This has resulted in six experimental groups, see table 3.1. Besides, a control group is included in which the respondents are not exposed to any bash action.

Table 3.1: Experimental Groups

3.2 Survey Design

Prior to being exposed to the different scenarios, each group has to answer some statements in a questionnaire. By answering the statements their identification with the brand McDonalds is tested. These statements are based on the potential measures on customer-company identification prepared by Bhattachary and Sen (2003) and Ahearne., Bhattacharya, and Gruen (2005).

1. Has good products/services. 2. Is well managed.

3. Only wants to make money. 4. Is involved in the community. 5. Responds to consumer needs. 6. Is a good company to work for.

For all statements about brand identification, identification with source and brand image, a seven-point Likert scale is used, ranging from “strongly disagree” (1) to “strongly agree” (7). There is chosen for a seven-point as this is best for optimal reliability (Matell & Jacoby, 1971). The respondents are also confronted with a few manipulation-check questions. To assess how people thought about the way McDonalds reacted to the several bash actions, they are asked to rate to what extend McDonalds reaction was weak or strong. Thereby it ispossible to check whether people see giving in indeed as a good reaction, and fighting back as a bad reaction. To assess how personally relevant the bash actionis people are asked to rate to what extend they feel to join the bash action. Finally, the survey includes some demographical questions, like age, gender and education, and respondents are asked about their social media usage and fast-food consumption.

3.3 Participants

The experiment will be extracting both male and female adults. For each experimental group a questionnaire is developed which is distributed via social media and direct mail, and can be completed online. The sample size consists of at least 140 respondents. Each respondent will be randomly assigned to the different groups, each consisting of at least 20 persons, whereas six groups function as the experimental groups and the seventh group is assigned as the control group. The experiment can be completed in approximately five to ten minutes. However, respondents will be allowed to process the bash action for as long as they wish for, while they are filling in the questionnaire.

3.4 Stimuli

3.4.1 Brand

For testing the hypotheses, it is necessary that the brand image of a certain brand is evaluated. For both studies this bash action is directed towardsMcDonalds, who is accused that their hamburgers contain horsemeat. The brand McDonalds is chosen because almost everyone has experiences with their products, or at least heard about it and so has formed an opinion about McDonalds. Furthermore, the brand is used by both males as females, which makes it adequate for this research.

3.4.2 Source

A choice has been made to conduct two studies as this can indicate the influential differences of the source of a certain bash action. The two sources chosen are television broadcast PowNed and a group of students. PowNed is chosen as it is a broadcast with a very clear image, and strong opinions. People know where PowNed stands for, and therefore we might assume that they can easily identify themselves as a member of this source. Students are chosen because they can more easily be seen as a ‘regular customer’ and therefore as a peer. We may expect that most people identify more with students then with PowNed, as the latter is probably perceived as less honest and reliable, and therefore less trustworthy.

3.4.3 Social Media Network

‘bashtag’ is named #crackdonalds. Respondents of all studies are asked if they would like to join the bash action, by posting tweets on this bashtag.

Furthermore, both studies make use of parodies on social media. The group of students parodies a commercial of McDonalds and publishes the video on YouTube. PowNed has got their own social media for sharing of pictures and video’s, called Dumpert and GeenStijl, and they publish their parodies on those networks.

3.4.4 Scenario’s

Each of the six experiment groups are exposed to a certain scenario. The respondents of study 1 are exposed to a bash action performed by the students, and the respondents of study 2 by PowNed. Furthermore, for the three groups of each study McDonalds responds in a different way (giving in, fighting back or no reaction). See Appendix B for the six scenarios.

3.5 Procedure

Before distribution of the questionnaires, a group of ten people performed a pre-test. After this, some little adoptions were made and the final questionnaires were published online. People received a link to the online questionnaire by e-mail or social media. As mentioned there are 6 different versions, each with their own scenario, and there is one control group. People are randomly assigned to a scenario or control group, which is selected by the website www.thesistools.com. After filling in some statements about brand identification, the people are asked to pay close attention to the scenario, for as long as they wish. Then some statements will follow about source identification, and brand image, and it ends with some demographical questions. The control group is not exposed to any scenario, and they do not have to rate statements about source identification.

3.6 Plan of Analysis

Next, independent t-tests and ANOVA tests are performed to compare the means for the descriptives. These tests are also used for the manipulation check.

In order to find prove for the hypothesesmultiple tests are performed. To test H1, an independent t-test is performed. Hereby the different means on brand image are compared, and significant differences might be found to see if a bash action is indeed damaging for a brand’s image. Furthermore a linear regression is performed to determine the strength of the effect.

In order to test the other hypotheses, a multiple regression analysis is performed to see what the main effects of the model are, and to compare the strength of these effects with the main effect of H1. A multiple regression is used to explore a relationship between a set of independent variables, and one dependent variable. Hereby the dependent variable is: brand image. For each respondent the mean of their valuation on the six statements about brand image is calculated and applied for the remainder of this research. Two dummies are made for the three possible ways McDonalds responses. One for giving in and one for fighting back, giving no reaction is hereby the base category. Both dummies are added as an independent variable in the multiple regression model. Furthermore, both identification with source as with brand are included as independent variables. For these variables the mean is applied for the remainder of the research as well. The equation of the multiple regression line is as follows:

BI = β0 + β1GI + β2FB + β3B + β4S +

ε

Furthermore, an extra multiple regression analysis is performed in order to test hypotheses 4 better. As brand identification is also measured for the respondents in the control group, the interaction term with the variable bash action can here be taken into account: whether or not the respondent was exposed to a bash action. The equation of the multiple regression line is as follows (where BA is bash action):

BI = β0 + β1B + β2BA + β3B*BA +

ε

Additionally, an explorative multiple regression analysis is performed to see if there are any interaction effects. Four new variables are created: the interaction terms. These interaction effects are not part of the hypotheses, but the results might provide us some useful information. Again, an extra multiple regression is performed including a dummy for no reaction, instead of fighting back Hereby, the equation of the multiple regression line is as follows:

4 Results

This chapter gives an overview of the results of the study. First of all, the demographics will be outlined, which will provide general information and sample characteristics. Then a factor analysis is presented for data preparation. Thereafter, a manipulation check is performed, followed by a multicollinearity analysis. Finally the hypotheses are tested.

4.1 Demographics

In this part, the sample will be outlined. The study is targeted on a population consisting of Dutch males and females. The questionnaires were filled in by 148 people, of which 145 could be used for hypotheses testing. The other 3 questionnaires were not fully completed, and therefore excluded for further research.

4.1.1 Gender, Age, Education

Table 4.1 shows the distribution of males and females in the sample. The proportion of males and females is surprisingly equal, as 51% are male, and 49% are female.

Table 4.1: Male / Female

Table 4.2 shows the age of the sample population. The minimum age of the respondents is 18. The sample is quite young, as more than half (56.6%) is aged between 18 and 25 years. Followed by 31% of the sample who have an age between the 26 and 45 years. 11.7% is aged between 46 and 65 years, and just 0.7% is older than 65.

Table 4.2: Age

Table 4.3: Education Level 4.1.2 Fast Food Usage

How much fast-food the respondents consume is shown in table 4.4. It does not show any extraordinary percentages, as 75.1% consumes fast-food every once in a while.

Table 4.4: Fast Food Usage

4.1.3 Social Media Usage

respondents. The top three is completed by LinkedIn (109 respondents), and Twitter (52 respondents).

Facebook is also the most frequently used, as 39.2% of the respondent’s uses it multiple times a day. This is a big difference when compared with the other social media, as Twitter is multiple times used a day by only 5.5%, and LinkedIn by 2.1%.

Table 4.5: Social Media Usage

4.2 Factor Analysis

A factor analysis can be performed to derive a set of uncorrelated variables for further analysis as the use of highly correlated variables may yield misleading results in regression analysis (Kim & Mueller, 1978). Thus, it is used for data reduction and data summarization and therefore to reduce the number of variables needed for the regression equation. The purpose is to reduce a large quantity of data by finding a common variance. Therefore the variables must be combined into a factor to get parsimony and less multicollinearity (Janssens,Wijnen, de Pelsmacker, P. and van Kenhove, 2008).

they are related to each other (Malhotra, 2009). At the same time the cronbach’s alpha is measured if an item is deleted. If this cronbach’s alpha is higher the item should be deleted for further analysis.

The closer to 1 the cronbach’s alpha value is, the more reliable (Janssens et al, 2008). Thus, the more likeable that these items measure the same construct. Above 0.6 is considered to be adequate. Above 0.8 is optimal. Table 4.6 shows that all constructs have a cronbach’s alpha which are considered optimal, and therefore reliable for further analysis. However, for the construct brand identification it appears that the cronbach’s alpha is higher when an item is deleted. This item is the variable ‘Trust McDonalds’, which was allocated to the fourth factor in the factor analysis. For that reason this item is deleted for further analysis, and is not used for the continuation of the research.

Table 4.6: Cronbach's Alpha

Table 4.7: KMO and Bartlett's Test

Next, it is necessary to determine the number of factors that can be used for further analysis. Only those factors can be used of which the total eigenvalue is higher than 1. Secondly those factors must explain 5% of the variance each, and finally the cumulative explained variance must be higher than 60% (Janssens et al, 2008). Table 4.8 shows that 3 factors should be used for further analysis.

The rotated component matrix in table 4.9 shows all factor loadings. All variables have loadings above 0.5 at the component that they belong to (Janssens, 2008). Indeed, all are loaded to the right component.

Conclusion

After deleting the variable ‘Trust McDonalds’, the 17 remaining variables used in the survey are downsized into three factors instead of four. The factors are recoded into the variables: brand identification, source identification and brand image.

4.3 Descriptives

Before checking whether the manipulation checks are successful and testing the hypotheses, this part provides some descriptives. By running several ANOVA-tests the mean values on brand image of the different groups are compared, to see if there are any significant differences. Table 4.10 shows that the source of the bash action, and the way the company responds did not lead to significant differences in the respondents mean value on brand image.

Table 4.10: Means on Brand Image

Figure 4.1: Graph of Mean Values on Brand Image

3.57

3.71

3.79

3.57 _3.56

Students PowNed Giving In Fight Back No Reaction

4.4 Manipulation Check

In order to check whether or not the respondents see the manipulated variables as it was meant to be seen, a manipulation check is done. For this research, two constructs have been manipulated: the source of the bash action and the way the company responds to the action.

4.4.1 Source of the Bash Action

As this manipulation contains two different groups of people, one group was exposed to a bash action performed by PowNed and the other to a bash action performed by a group of students, an independent sample T-test is good to use (Janssens et al, 2008). By using this test, the mean scores of the two groups are compared. If there is a significant difference between the group, the manipulation is successful. As can be seen in table 4.11 the p-value is 0.000, which is strongly significant. Therefore, the conclusion can be made that the manipulation check succeeded. In other words, respondents who were exposed to a bash action by a group of students identified themselves significantly more with that source than when they were exposed to a bash action performed by PowNed.

*p<0.05

Table 4.11: Manipulation Check Source of Bash Action 4.4.2 Company’s response

The difference between the groups is strongly significant with a value of p=0.000. This means that the manipulation check was successful. The posthoc test results show us the differences between the groups. The results are shown in table 4.12, and tell us that:

 The difference between a company giving in and fighting back is strongly significant (p=0.000)

 The difference between a company giving in and giving no reaction is strongly significant (p=0.000).

 There is no significant difference between a company fighting back or giving no reaction (p=0.548).

* p<0.05

Table 4.12: Manipulation Check Company’s Response

4.5 Testing for Multicollinearity

When a regression analysis is performed, the risk of having multicollinearity will rise. Multicollinearity means that there is a high degree of correlation between the independent variables, this is not permitted because it has negative effects (Janssens et al 2008).

A possible way to encounter this problem of multicollinearity is to create standardized variables. This means that the mean is subtracted from each of the independent variables. Thus, for this research all the independent variables which are included in the multiple regression analysis are standardized.

that all VIF scores are lower than 3.525 and tolerance scores are higher than 0.1. Therefore, we might conclude that the variables have no multicollinearity.

4.6 Hypothesis Testing

In this section, the results of the independent t-test and the various multiple regression models are presented and discussed to see if the hypotheses can be supported or not.

4.6.1 Main Effect

In order to reject or accept hypothesis 1, the results of the independent t-test and the linear regression are discussed. The relation between the independent variable ‘Bash Action’ with the dependent variable ‘Brand Image’ is tested. The independent t-test compares the means to see if there is a significant difference. Table 4.14 shows the results.

* p<0.05

Table 4.14: Results Independent t-test

There is a significant difference (p=0.016) between how a respondent values the brand image when he or she is exposed to the bash action and when it is not. The results indicate that brand image is valued much lower when the respondent was exposed to the bash action, this tells us that a bash action indeed negatively affects brand image. Furthermore a linear regression is performed in order to determine the strength of the main effect of the bash action. See table 4.15. The linear regression model is significant with a F-value of 5.889 (p=0.016). Just 4% of the total variance is explained by the model as a whole (R²). As predicted, a bash action appears to have a negative effect on brand image with a strength of -.0199.

a. Dependent variable: Brand Image b. All the independent variables are standardized *p<0.05

4.6.2 Multiple Regression

In order to reject or accept hypothesis 2a, 2b, 2c, 3 and 4, the results of the various multiple regression analysis are discussed. First, a multiple regression analysis is performed in order to compare the strengths of the independent variables. Hereby the relation between the independent variables company response (GI, FB and NR), brand identification (B) and source identification (S) with the dependent variable brand image (BI) is tested. Preliminary analysis was performed to make sure that there was no violation of the assumptions of normality, linearity, and homoscedasticity.Table 4.16 shows the results. First of all it indicates that the model is significant with an F-value of 6.297 (p=0.000) and 17.3% of the total variance is explained by the model as a whole (R²).

a. Dependent variable: Brand Image b. All the independent variables are standardized *p<0.05

Table 4.16

The Beta shows the standardized coefficient of the variable. These values are converted to the same scale for each variable so they can be compared with each other and with the Beta of the bash action itself. As expected, giving in to a bash action seems to have a positive effect for a brand’s image. In contrary, although weaker than the bash action itself, fighting back has indeed a negative relation with brand image.

An extra multiple regression analysis is performed in order to make sure that hypothesis 4 can be accepted. The results are shown in table 4.17. The model is significant with an F-value of 11.977 (p=0.000) and 20.3% of the total variance is explained by the model as a whole. The results confirm that a bash action has a negative effect on a brand’s image, which is significant. The interaction term has a positive effect on brand image as well. This indicates that the negative effect of the bash action is weakened when the respondent identifies him- or herself with the brand. However, this effect is not significant. The main effect between identification with brand and brand image is again highly positive and significant.

a. Dependent variable: Brand Image

b. All the independent variables are standardized *p<0.05

Table 4.17

a. Dependent variable: Brand Image b. All the independent variables are standardized *p<0.05

5 Discussion

The study applied two bash actions performed by either PowNed or a group of students towards McDonalds. The goal of this study was to examine the effect of the bash actions, executed by Social Media, on brand image. Hereby, three constructs (1) a company’s response, (2)

identification with the source, and (3) identification with the brand were proposed to strengthen or weaken the relation between the bash action and a brand’s image. This chapter will provide the conclusions regarding the problem statement. Furthermore, the limitations and implications of the study, and recommendations for further research are given.

The study started with an extensive theoretical research to find out what was written about the variables and moderators in earlier literature. As a result of these findings six hypotheses were formulated and tested later on. Table 5.1 gives a complete overview of the six hypotheses and its results.

Table 5.1: Hypotheses

The following problem statement was formulated:

 How does a bash action impact a brand’s image, and what role might a company’s

The problem statement can be separated into two parts. The first part is about the direct relation, the main effect, between the independent (bash action) and the dependent variable (brand image), the second part describes the moderating effects.

5.1 Conclusion Main Effect

The first part of the problem statement is related to H1. The results of the independent t-test and linear regression analysis showed that a bash action negatively influences brand image, and that there is a significant difference between how the respondents valuated a brand’s image if they were exposed to a bash action and when they are not. These outcomes are in line with the

expectations prior to the study. And therefore H1 is accepted. The main reason that a bash action is so damaging for a brand may lay in the fact that it is executed by social media. The downside risk to a brand’s image from media attention is, in general, greater than the upside potential for image enhancement (Gray & Balmer, 1998). In the era of social media in particular, it is very hard for managers to control the content and frequency of information spread by the users (Mangold & Faulds, 2009). However, when a firm is involved in an image damaging event, the primary challenge to repair their image is to minimize the diffusion of news about that event (Rhee & Valdez, 2009). Minimizing the diffusion of the damaging news that the bash action includes seems impossible, and therefore the image can’t easily be repaired. Findings of Trusov, Bodapati, and Bucklin (2010) contradicts the negative relation. They state that negative postings at corporate Facebook pages have no effect. According to their theory, consumers do not

experience any difference between the posts whether it is negative or positive.However, we might assume that a bash action as used for this research has a larger negative effect then just negative Facebook postings, because of the rapid diffusion of its content.

5.2 Conclusions Moderating Effects

super-effort. This super-effort is comparable to giving in to a bash action. An explanation why there are no significant results might be that giving in to the bash action is perceived as positive, and therefore the negative effect on brand image is minimalized.

Fighting back has, as expected, a negative effect on brand image. Results of a study by Siomkos and Shrivastava (1993) showed that companies must avoid denying their responsibility for a crisis, as this does not convince the consumers of their innocence. However, the results for fighting back were far from significant so H2b is rejected. A reason might be that it only has a very small negative effect on brand image, and therefore the differences are probably too small to show significant results.

The results of the one-way ANOVA and themultiple regression model show that giving no reaction has a negative effect on brand image, just as fighting back. H2c is rejected, but the results are striking as the prediction was that giving no reaction would have no effect. The outcomes however indicate that giving no reaction is damaging for brand image. An explanation could be that consumers form more negative opinions if they find out that a company does not take action while they are aware of the problem (Siomkos & Shrivastava, 1993).

study might have led to different results when another company than McDonalds was chosen, as McDonalds is a company that has a very clear and established reputation.

5.3 Limitations and Future Research

This research has some limitations that will be discussed in this section. Furthermore, the limitations provide directions for future research. These directions will be provided in this section as well.

Firstly, the relative small sample size of 145 respondents is a limitation. The sample used for this research might be a bad representation of the total population. This might be strengthened by the fact that the demographical factors of the sample have large differences with the general Dutch population. The sample is for a large part high educated, and is furthermore relatively young. As a result, this might cause biased results. So, because of the small sample size and composition, it is possible that the findings cannot be generalized. Hence, a larger sample size is recommended for future research in order to generalize the findings.

Secondly, a limitation might be the choice to direct the bash actions used for this research only at the brand McDonalds. As a large part of the sample did not identify itself with this brand, we might say that this is an unlikeable brand. Therefore, there might have been different results when the bash actions were directed towards a more likeable brand. So future research might investigate the effect of bash actions towards for example Apple as this is probably a more brand that more people identify itself with. The choice to direct the bash actions toward McDonalds might also be a limitation as it is a very well-reputed company. As mentioned earlier the research might have led to different results when a less known company was used.

Thirdly, this research did not measure any difference in brand image. Respondents were only asked to rate statements at the end of the questionnaire. Future research might want to start with these statements as well. By doing so the effect of the bash actions on brand image might be measured more precise.

5.4 Implications

This study provides insight in what the exact implications of a bash action are, the way a company should respond to these actions and what role identification with the source and brand plays. As a consequence, strategies and communication tactics to deal with the dark side of this kind of customer engaging behaviour can be determined.

6 References

Aaker, D. A. (1991). ‘Managing brand equity’. New York: Free Press

Ahearne, M., Bhattacharya, C.B., and Gruen, T. (2005), ‘Antecedents and Consequences of Customer–Company Identification: Expanding the Role of Relationship Marketing’, Journal of

Applied Psychology, Vol. 90, No. 3, p574-585

Anderson, J. R., & Bower, G. H. (1973). Human associative memory. New York: Halstead. Badenhausen, K. (1996),‘Blind faith’, Financial World, July 8, pp. 50–55

Batra, R., and Homer, P.M. (2004), ‘The Situational Impact of Brand Image Beliefs’, Journal of

Consumer Psychology, 14(3), p318-330

Bhattacharya, C.B., and Sen, S. (2003), ‘Consumer-Company Identification: A Framework for Understanding Consumers' Relationships with Companies’,Journal of Marketing, Vol. 67, No. 2 (Apr., 2003), pp. 76-88

Bikhchandani, S., Hirshleifer, D., Welch, I. (1992), ‘A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades’, Journal of Political Economy, Vol. 100, No. 5, pp. 992-1026

Blodgett, J.G., Wakefield, K.L, and Barnes, J.H.,(1995), ‘The effects of customer service on consumer complaining behavior’, Journal of Services Marketing, vol. 9, no 4, pp31-42

Brodie, R.J., Hollebeek, L.D., Juric, B., Braun, A.I., & Esswein, W. (2012), ‘Corporate Risks in Social Networks- towards a risk management framework’, Journal of Service Research, 2011 14: 252 originally published online 6 July 2011

Braun, R., and Esswein, W. (2012), ‘Corporate Risks in Social Networks- towards a risk management framework’, AMCIS 2012 Proceedings.

Cretu, Anca E. and Brodie, Roderick J. (2007), "The influence of brand image and company reputation where manufacturers market to small firms: A customer value perspective,"

Industrial Marketing Management, 36 (2), 230-240.

Dawar, N., & Pillutla, M. (2000). ‘The impact of product–harm crises on brand equity: The moderating role of consumer expectations.’, Journal of Marketing Research, XXXVII, p215 – 226.

Delgado-Ballester, E., and Munuera-Alemán, J.L. (2001), ‘Brand Trust in the Context of Consumer Loyalty’, European Journal of Marketing, Vol. 35 No 11/12, 2001, pp 1238-1258 Delgado-Ballester, E., and Munuera-Alemán, J.L. (2005), ‘Does brand trust matter to brand equity?’, Journal of Product & Brand Management, 14/3/2005, p187-196

van Doorn, J., K.N. Lemon, V. Mittal, S. Nass, D. Pick, & P.C. Verhoef (2010), ‘Customer Engagement Behavior: Theoretical Foundations and Research Directions’, Journal of Service

Research, 2010 13: 253

Fader, P.S., and Russell, S.W. (2012), ‘Introduction to the Special Issue on the Emergence and Impact of User-Generated Content’, ‘Marketing Science’, Vol. 31, No.3, May-June 2012, pp 369-371.

Fitness, J. (2001), ‘Betrayal, Rejection, Revenge, and Forgiveness: An Interpersonal Script Approach’, Interpersonal rejection, Ed. Leary, M, New York: Oxford University Press, pp. 73-103.

Fournier, S., Avery, J., (2011), ‘The Uninvited Brand’, Business Horizons, 54, p193-207 Gardner, B.G., and Levy, S.J. (1955), ‘The Product and the Brand’, Harvard Business Review, March-April, p33-39

Gray, E.R., and Balmer, J.M.T. (1998), ‘Managing Corporate Image and Corporate Reputation’,

Elsevier Science Ltd, Vol. 31, No. 5, pp. 695 to 702

Hennig-Thurau T., K.P. Gwinner, G. Walsh and D.D. Gremler (2004), ‘Electronic Word-of_{‐Mouth Via Consumer-Opinion Platforms: What Motivates Consumers to Articulate} Themselves on the Internet?, ‘Journal ofInteractiveMarketing,’ 8 (1), 38-52.

Hogg, M.A. (2006), ‘Contemporary Social Psychological Theories’.

Javalgi, R.G., Traylor, M.B., Gross, A.C. and Lampman, E.(1994), ‘Awareness of Sponsorship and Corporate Image: An Empirical Investigation’, Journal of Advertising, Vol. 23, No. 4 (Dec., 1994), pp. 47-58

Janssens, W., Wijnen, K., de Pelsmacker, P. and van Kenhove, P. (2008), ‘Marketing Research with SPSS’

Kaplan, A.M., Haenlein, M. (2010), ‘Users of the world, unite! The challenges and opportunities of Social Media’, Business Horizons 53, 59—68

Keller, K.L., (1993), ‘Conceptualizing, Measuring, and Managing Customer-Based Brand Equity’, Journal of Marketing, Vol. 57, No. 1 (Jan., 1993), pp. 1-22

Keller, K.L. (2007), ‘Strategic Brand Management:Building, Measuring, And Managing Brand Equity’

Kim, J.O., and Mueller, C.W. (1978), ‘Factor Analysis: Statistical Methods and Practical Issues’. Klein. J., and Dawar, N. (2003), ‘Corporate social responsibility and consumers’ attributions and brand evaluations in a product–harm crisis’, Intern. J. of Research in Marketing, 21 (2004) p203 – 217

Ludwig, S., de Ruyter, K., Friedman, M., Brüggen, E.C., Wetzels, M., and Pfann, G. (2013), ‘More than Words: The Influence of Affective Content and Linguistic Style Matches in Online Reviews on Conversion Rates’, Journal of Marketing, Vol. 77 (January 2013), p87-103

Mael, F.A., and Tetrick, L.E. (1992), ‘Identifying Organizational Identification’, Educational

and Psychological Measurement, 1992 52: 813

Majumdar, A., and Krishna, S. (2012), ‘Organizational Identity, Image and Role of the Influencers in Social Media: A Framework for Analyzing their Interactions’.

Malhotra, N., 2009, Marketing research: an applied orientation, Pearson education, 6thedition

Mangold, G.W., and Faulds, D.J. (2009), ‘Social media: The new hybrid element of the promotion mix’, Business Horizons 52, 357—365

Martinez, E., and de Chernatony, L. (2004), ‘The Effect of Brand Extension Strategies upon Brand Image’, Journal of Consumer Marketing, Vol. 20, Number 1, pp 39-50

Matell, M.S., and Jacoby, J. (1971), ‘Is There an Optimal Number of Alternatives for Likert Scale Items? StudyI: Reliability and Validity’, Educational and Psychological Measurement, 31: 657

McPherson, M., Smith-Lovin, L., and Cook, J.M. (2001), ‘Birds of a Feather: Homophily in Social Networks’, Annu. Rev. Sociol. 2001. 27:415–44

Mudambi, Susan (2002), "Branding importance in business-to-business markets: Three buyer clusters," Industrial Marketing Management, 31 (6), 525-533.

Nair, M (2011), ‘Understanding and Measuring the Value of Social Media’, Wiley Online

Parsons, A.L. (2011), ‘Social Media from a Corporate Perspective: A Content Analysis of Official Facebook Pages’, Proceedings of the Acadamy Studies, vol. 16, number 2

van Reijmersdal, E.A., Neijens, P.C., and Smit, E.G. (2007), ‘Effects of Television

Brand Placement on Brand Image’, Psychology & Marketing, Vol. 24(5): 403–420 (May 2007) Rhee, M., and Valdez, M.E. (2009), ‘Contextual Factors Surrounding Reputation Damage with Potential Implications for Reputation Repair’, Academy of Management Review, Vol. 34, No. 1, 146–168.

Sashi, C.M. (2012), ‘Customer engagement, buyer-seller relationships, and social media’,

Management Decision, Vol. 50 No. 2, 2012, pp. 253-272

Sawhney, M., G. Verona and E. Prandelli (2005), ‘Collaborating to Create: The Internet as a Platform for Customer Engagement in Product Innovation’, Journal of Interactive Marketing, vol 19, issue 4

Sinclaire, J.K., and Vogus C.E. (2011), ‘Adoption of Social Networking Sites: an Exploratory Adaptive Structuration Perspective for Global Organizations’ Information Technology and

Management, 12(4): p293-314

Siomkos, G.J., and Kurzbard, G. (1994), ‘The Hidden Crisis in Product-Harm Crisis Management’, European Journal of Marketing, Vol. 28 No. 2, pp. 30-41.

Siomkos, G.J., and Shrivastava, P. (1993), ‘Responding to Product Liability Crises’, Long Range Planning, Vol. 26, No. 5, pp. 72 to 79

Trusov, Bodapati, and Bucklin (2010), “Determining Influential Users in Internet Social Networks,” Journal of Marketing Research, 47 (4), 643-658.

Verhoef, P.C., Werner, J.R., and Krafft, M. (2010), ‘Customer Engagement as a New Perspective in Customer Management’, Journal of Service Research, 2010 13: 247

Voorhees, C.M., Brady, M.K., and Horowitz, D.M., (2006), ‘A Voice From the Silent Masses: An Exploratory and Comparative Analysis of Noncomplainers’,Journal of the Academy of

Marketing Science, 2006 34: 514

Wang, Y., and Vassileve, J. (2003), ‘Trust and Reputation Model in Peer-to-Peer Networks’, Proc. of IEEE Conference on P2P Computing, Linkoeping, Sweden, September 2003.

Williams, K.T. (2009), ‘Case Study: Dell Hell. The impact of social media on corporate communication’

Yim, C. K., Tse, D. K., & Chan, K. W. (2008). Strengthening customer loyalty through intimacy and passion: roles of customer-firm affection and customer-staff relationships in

services. Journal of Marketing Research, 45(6), 741-756.

Young, T. (2009), ‘Flashback 2006: When Mentos 'Got It' & Coke Didn't. What Would the Story be Today?’

7 Appendix

Appendix A: Questionnaire

Allereerst wil ik u bij deze hartelijk danken voor het deelnemen aan dit onderzoek!!

Deze enquête wordt gehouden voor het afstudeeronderzoek van Booy van Hees, met betrekking tot zijn Master Marketing aan de Rijksuniversiteit Groningen. De vragen gaan over uw attitude tegenover het merk McDonalds, en hoe externe factoren hierop van invloed kunnen zijn. De vragenlijst duurt tussen de 5 en 10 minuten, en de resultaten zullen uiteraard anoniem worden verwerkt.

Belangrijk: het scenario met betrekking tot McDonalds is volstrekt fictief.

1. De volgende vragen hebben betrekking tot McDonalds. Vult u alstublieft in tot welke mate u het eens bent met de volgende stellingen, op een schaal van 1 tot 7 (waarin 1 totaal oneens is, en 7 totaal eens).

a. Ik eet vaak Fastfood

b. Waar McDonalds voor staat bevalt me c. Ik herken me in McDonalds

d. Ik vertrouw McDonalds e. Ik denk vaak aan McDonalds

f. Ik probeer de nieuwe producten die McDonalds introduceert graag uit

g. Ik praat graag positief over McDonalds en haar producten tegen mijn vrienden en/of familie

Na deze vragen wordt de ondervraagde blootgesteld aan een van de scenario’s, te vinden in Appendix B.

2. Na het lezen van het scenario hebben de volgende vragen betrekking tot PowNed / Groep Studenten. Vult u alstublieft in tot welke mate u het eens bent met de volgende stellingen, op een schaal van 1 tot 7 (waarin 1 totaal oneens is, en 7 totaal eens).

a. Waar PowNed / Groep Studenten voor staat bevalt me b. Ik herken me in PowNed / Groep Studenten

c. Ik vertrouw PowNed / Groep Studenten

d. Ik praat graag positief over PowNed / Groep Studenten tegen mijn vrienden en/of familie

3. Na het lezen van het scenario hebben de volgende vragen wederom betrekking tot

McDonalds. Vult u alstublieft in tot welke mate u het eens bent met de volgende stellingen, op een schaal van 1 tot 7 (waarin 1 totaal oneens is, en 7 totaal eens).

a. Ik beschouw de reactie van McDonalds als zijnde sterk b. Ik beschouw de reactie van McDonalds als zijnde zwak

c. Ik zou deel willen nemen aan de haat campagne, bijvoorbeeld door tweets te posten op #CrackDonalds

d. Ik denk dat McDonalds betrouwbaar is

e. Ik vind dat McDonalds haar producten van goede kwaliteit zijn f. Ik vind dat McDonalds goede service levert

g. Ik vind dat McDonalds goed gemanaged wordt

h. Ik vind dat McDonalds betrokken is bij de maatschappij

j. Ik vind dat McDonalds haar klanten respecteert

4. De volgende vragen gaan over uw social media gebruik.

a. Van welk van de volgende social media maakt u gebruik?

meerdere antwoorden mogelijk

i. Facebook ii. Hyves iii. MySpace iv. Google + v. LinkedIn vi. Twitter

b. Hoe vaak maakt u gebruik van de volgende social media?