Music and the Dark Side of Marketing -Music in the Media Markt-

(1)

Music and the Dark Side of Marketing

-Music in the Media Markt-

(2)

Music and the Dark Side of Marketing -Music in the Media Markt-

Author: Linda Rorijs 1612549

Msc. Marketing Management Supervisor 1: Liane Voerman

Supervisor 2: Debra Trampe

Topic: The dark side of marketing, music congruency, music tempo, approach/ avoidance behaviour, consumer feelings, Media Markt, ambient factors, retail environment, computer simulation, factor analysis, ANOVA, MANOVA

Hand- in date first draft: 29-12-2010 Hand- in final version: 07-03-2011 Number of pages: 62

(3)

Management summary

Goal of this research is to get a better understanding of the dark side of marketing. The dark side of marketing, as coined by d’Astous (2000), aims to identify the environmental factors that “create negative consumer feelings during shopping” (d’Astous, 2000; p149). Argued is that only after the negative aspects of an environment have been fixed, aspects specifically for triggering positive behaviour can have effect. The main research question of this research is:

How does music in the retail environment cause negative feelings and avoidance behaviour among customers in a retail setting?

The central research question is divided in several sub questions, to guide the structure of this research. The sub questions, and their answers, can be found below.

1) How does the retail environment, and more specifically ambient factors, influence customers?

(4)

2) What factors can be considered as creating the dark side of marketing in ambient factors?

The literature and research on the dark side of marketing is very limited. While there is a fair amount of research spent on approach/avoidance behaviour, the majority of research is still focused on the bright side of marketing. The dark side of marketing is underdeveloped.

Volume, familiarity, congruency, and tempo are pointed out as factors in music that are able to create avoidance behaviour and negative consumer feelings. Logically, one could assume that volume is the largest determinant of the dark side of marketing, as customers notice incongruent, unfamiliar, or unpreferred music more when it is played loudly. However, Smith and Curnow (1965) found that both low and high volumes could create avoidance behaviour. Music familiarity and music preference will not have priority in this research, due to the strong interaction between these variables. Music congruency formed the focus of this research together with music tempo. People often have a strong preference for tempo, and might thus create avoidance behaviour and negative feelings when the preferred music tempo and the actual/played music tempo do not match. Congruency is already firmly established in the scent literature, but received little attention in music literature. Expected is that congruency also relates to the ambient factor music (Areni & Kim,1993).

3) What type of avoidance behaviour and negative feelings are caused by ambient factors, and more specifically music?

(5)

The results of the questionnaire indicated that all individual elements within the approach/ avoidance construct are significantly influenced by one or more independent variables. The same applies to consumer feelings. This variable is also significantly influenced by multiple independent variables. Below, all significant relations found will be discussed shortly.

Music congruency significantly influenced consumer feelings. Congruency thus also plays a role in the ambient factor music (while before this variable was only established in scent literature). Other variables that significantly influence the dark side of marketing, although not set up in hypotheses, include music liking (significantly influences consumer feelings), song preference (significantly influences perceived time spent in the retail environment), shopping enjoyment (significantly influences perceived time spent, GAPM, and intention to revisit), brand preference (significantly influences purchase intention, GAE, GAPP, and intention to revisit), and remembering stimulus (significantly influences consumer feelings). These variables only influence one component of the dark side of marketing at a time, signaling the importance of including both consumer feelings and approach/ avoidance behaviour under the header of the dark side of marketing.

(6)

4) What can retailers do to prevent avoidance behaviour and negative feelings among customers by using music?

Retailers should avoid using incongruent music. Incongruent music increases negative feelings among customers. Congruency, in this research, is related to congruency with the store, not the products in the assortment (see section 3.5.2). Congruency related to a store concept could present the retailer with many opportunities, as stores selling multiple categories of products, can still have congruent music without having to play a different type of music in each department.

Next to music congruency, music preference and music liking are also important factors that retailers should focus on. Music should be liked and preferred by customers, but this can be a challenging task as music preferences are highly individual. Media Markt should use consumer panels to determine a range of appropriate music that is congruent, preferred and liked by its customers to decrease avoidance behaviour. This consumer panel, which should consist out of a representative sample of Media Markt customers, can also be used to judge other ambient factors, design factors, and social factors (Baker, 1986).

(7)

Preface

In the period of April 2010 to March 2011 I have worked on my master thesis as a graduation project from the study Marketing Management at the Marketing Department of the Rijksuniversiteit Groningen. This thesis is based on a literature research and an experiment.

During this period, I had help from a few people whom I would like to thank for a variety of reasons. First and foremost, I would like to thank Liane Voerman and Debra Trampe for their help, guidance, and structural feedback throughout this master thesis project, which allowed me to finish this project to my satisfaction. Secondly, I would like to thank Remy Tuithof from Media Markt Groningen for allowing me to take photographs of the retail environment, which formed the basis of my computer simulation. Thirdly, I would like to thank Harm de Weerd from the Rijksuniversiteit Groningen for creating the computer simulation. Without Harm de Weerd it would not have been possible to gather the data needed for this research in such an original manner. Fourthly, I would like to thank Philip Pruim for helping me take the pictures in the Media Markt, which formed the input for my computer simulation. I would also like to thank Bram Liemburg, for helping me create a website and integrate the computer simulation on this website. Lastly, I would like to thank my father, mother, and sister, who motivated me to achieve the best result possible.

Without all these people I would not have been able to create my master thesis as it is now. I believe my master thesis stands out positively due to its innovative angle. A topic is researched that did not receive a lot of attention until now, namely the dark side of marketing. Furthermore, a computer simulation is used for gathering data, and extensive literature research is done. The past period has been intense, but I enjoyed exploring literature and data, in search of new insights in the topic of the dark side of marketing.

(8)

Chapter 1 Introduction

Throughout history, competition between stores has increased tremendously. Where retailers mostly competed on price and advertisements before, there is now a focus on creating pleasant retail environments for customers (Levy & Weitz, 2009). Aim is to retain customers, increase spending and increase customer loyalty. What better way to attain this, than by creating a retail environment that customers feel happy in? Retail environments have been the focus of numerous studies (e.g. Yalch & Spangenberg, 2000; Baker et al, 2002; Grewal & Baker, 1994). Still, gaps exist in the literature. Where literature mostly focuses on how to create positive reactions to a retail environment, less attention has gone to how negative reactions towards the retail environment can be prevented. While it is important to figure out what attracts customers to a retail environment, it might be equally, or even more important, to understand what drives them away. This dark side of marketing, as coined by d’Astous (2000), aims to identify the environmental factors that “create negative consumer feelings during shopping” (d’Astous, 2000; p149). When understanding what aspects of a store can cause negative consumer feelings, retail owners can anticipate on these aspects to prevent customers from feeling disappointed, frustrated, or sad. Studying the dark side of marketing is thus a first step in creating a more pleasant retail environment. Only after the negative aspects of an environment have been fixed, aspects specifically for triggering positive behaviour can have effect.

1.1 The retail environment

(12)

Mehrabian and Russel (1974) developed a framework for “studying the effects of store atmosphere on shopping behaviour”. This framework, named S-O-R (or Stimuli, Organism, and Response), argues that the environment stimulates the customer, affects the customer’s internal evaluations, and ultimately leads to the customer’s approach or avoidance behaviour. Stimuli are the environmental factors in the retail environment, which include ambient factors, design factors and social factors (Baker, 1986). In their research, Spangenberg et al (2005) conclude that the ambient factor music can significantly influence customers in retail environments. The focus of this research and the following experiment will therefore be on music.

1.2 Research questions

In short, this research tries to deepen the knowledge about the dark side of marketing by looking at the retail environment, ambient factors in general, and specifically music. Hypotheses will be tested using a computer simulation in a real experiment. The research question that forms the basis of this research is:

How does music in the retail environment cause negative feelings and avoidance behaviour among customers in a retail setting?

The central research question is divided in several sub questions, to guide the structure of this research. These sub questions are as follows:

1) How does the retail environment, and more specifically ambient factors, influence customers? (Section 2.1)

2) What factors can be considered as creating the dark side of marketing in ambient factors? (Section 2.2 and 2.3)

3) What type of avoidance behaviour and negative feelings are caused by ambient factors, and more specifically music? (Section 4.4/ section 5.1)

(13)

1.3 Structure of the research

(14)

Chapter 2 The retail environment

The goal of this chapter is to present the reader with a thorough overview of both the retail environment and the music literature. Chapter 2 discusses the retail environment, interactions between customers and the retail environment, ambient factors in general, and the ambient factor music. This chapter forms the basis of the research and experiment that will be discussed later in this research.

2.1 Interactions between customers and the retail environment1

Mehrabian and Russel (1974) developed a framework “for studying effects of store atmosphere on shopping behaviour”.2 The framework, which is better known as the S-O-R framework, contains three horizontally linked stages that describe how customers are influenced by a store atmosphere. The three individual stages are termed ‘Stimuli’, ‘Organism’ and ‘Responses’. The S-O-R framework argues, that “the environment is a stimulus (S), containing cues that combine to affect people’s internal evaluations (O), which in turn create approach or avoidance responses (R)” (Spangenberg et al, 1996; p68). Formally stated, this framework relates “features of the environment (S) to approach/ avoidance behaviours (R) within the environment, mediated by the individual's emotional states (O) aroused by the environment” (Donovan et al, 1994; p284).

Figure 1: The S-O-R model (Mehrabian & Russel, 1974)

1_{Lazarus (1991) and Kotler (1974) also proposed a framework discussing how retail environments} influence customers. These frameworks received little attention in researches until now and will, for that reason, not be discussed further in this study.

(15)

2.1.1 Stimuli (S)

The features of the environment, or stimuli (S), are the elements present in a retail environment. These elements, which can include lighting, scent, music, signage and layout, form the input for the S-O-R model (Fiore & Kim, 2007). The stimuli and retail environment will be further discussed in section 2.2.

2.1.2 Mediating factors (O)

According to the framework of Mehrabian and Russel (1974), three basic emotional states mediate the relationship between the stimuli and the approach/ avoidance behaviours of customers. These emotional states are summarized as PAD, or Pleasure, Arousal, and Dominance (Donovan et al, 1994). Pleasure is “the degree to which a person feels good, joyful, or happy” (Fiore & Kim, 2007; p430). Arousal is “the degree to which a person feels excited, stimulated, alert, or active” (Fiore & Kim, 2007; p430). Arousal is used most often in literature as a mediating factor in research related to the retail environment. Optimal- arousal theory discusses that every individual has a different optimal arousal point (Spangenberg et al, 2006). When the arousal experienced lies above this level of optimal arousal, avoidance behaviours are likely to follow (see Responses (R)). Avoidance behaviour might also occur when the arousal that is experienced by customers is too low, as this creates a boring store with little stimulation.

(16)

2.1.3 Responses (R)

The resulting responses (R) in the model are restricted to two opposite behaviours; approach or avoidance behaviour3. Approach behaviours are positive behaviours towards the retail environment and the merchandise within that environment, while avoidance behaviours represent the opposite. Avoidance behaviours reflect “a desire to leave a store, limited intention to revisit, or a failure to spend money” (Spangenberg et al, 2006; p.1282).

Stimuli, arousal and approach/avoidance behaviour will feature in the upcoming literature discussion and the following experiment.

2.2 Defining the retail environment

Retail environments have been defined as the space that “distinguishes the person from the physical features surrounding him or her” (d’Astous, 2000, p.149). Considering physical features only, suggests that concrete elements alone define the retail environment. Research, however, proposes that the retail environment also includes intangible elements, such as lighting, music or scents (also known as ambient factors) (e.g. Areni & Kim, 1993; Bone & Ellen, 1999). Nevertheless, naming up all the elements that comprise a retail environment would not be possible, as these are simply ‘too numerous’ (d’Astous, 2000). Several researchers (e.g. Massara & Pelloso, 2006; Baker, 1986; Kotler, 1974; Bitner, 1992) recognized this shortcoming and created categorizations of elements that comprise the retail environment.4 These categorizations aim to “classify environmental features into a finite set of categories” (Hunt, 1991).

3_{According to Donovan and Rossiter (1982) there are four general aspects to approach/ avoidance}

behaviour. These general aspects are: physical approach/avoidance behaviour, exploratory approach/avoidance behaviour, communication approach/ avoidance behaviour, and performance and satisfaction approach/avoidance behaviour.

4_{The categorizations of Bitner (1992) and Kotler (1974) will not be further taken into account, as limited}

(17)

Massara & Pelloso (2006) defined the retail environment as consisting out of micro, meso and macro environments (Massara & Pelloso, 2006). When a customer enters the store, he/she will first see the macro environment, than pass through the mesa environment, and stops to buy something in the micro- environment. The macro environment includes the external environment or exterior of the store (e.g. size of building, colour of a building, surrounding stores and signs). The mesa environment includes the internal environment of the store, and “all the layout and design variables that determine the structure of the interior” (e.g. orientation signs, floor carpeting, space design)(Massara & Pelloso, 2006; p521). The micro- environment, includes “all the variables that are, in a geographical sense, within close proximity to the customer” (Massara & Pelloso, 2006; p522). These variables can include scent, lighting, product displays, crowding and music (see figure 2).

Baker (1986) proposed a different kind of definition. She defines the retail environment as consisting out of ambient factors, design factors, and social factors. The overview of these factors can be found in table 1. As mentioned before, ambient factors are the Stimuli (S) from the model of Mehrabian and Russel (1974). The environmental factors that comprise Baker’s categories are used to create a certain atmosphere in a store.5 These atmospheres can, among others, facilitate purchase probability and/or create approach/avoidance behaviour (Kotler, 1974; Spangenberg et al, 2006).

Figure 2: The retail environment according to Massara & Pelloso (2006)

5_{The manipulation of environmental factors to create atmospheres is also called the ‘study of}

(18)

Table 1: The retail environment according to Baker (1986)

2.3 Ambient factors

Ambient factors can be defined as “background conditions that exist below the level of our immediate awareness” (Baker, 1986). These background conditions “may or may not be consciously perceived but affect all human senses” (d’Astous, 2000). Ambient factors were already shortly introduced in the previous section. As the ambient factors mentioned by Baker are limited, further research was done to create a wider variety of ambient factors. An overview of ambient factors can be found in table 2. From these factors, this research will focus on music.7 Before discussing the ambient factor music in more detail, the dark side of marketing will be introduced.

Ambient factors

Ventilation (Baker, 1986) Cleanliness (Baker, 1986) Temperature (Baker, 1986; Bitner, 1992)

Air quality (Baker, 1986) Noise (Baker, 1986; Bitner, 1992) Lighting (Bitner, 1992) Humidity (Baker, 1986) Scent (Baker, 1986; Bitner, 1992) Music (Bitner, 1992)

Table 2: Overview ambient factors

6_{While social factors do have an influence on the behaviour of customers in a retail environment, they may} not be seen as atmospherics as they cannot be manipulated to fit the context or type of store, like can be done with lighting, design or signage. The appearance (e.g. uniforms) of service personnel may be seen as atmospheric, however, this is only part of how Baker defined the term social factors.

7_{While this research will focus on music separately, it is important to note that interactions between}

ambient factors also can have major influences on the approach/avoidance behaviour of customers in a retail environment and subsequently the dark side of marketing. Solomon (1983) states that customers decode meanings and structure their behaviour after looking at the “total collection of cues in the environment”. Music, for example can interact significantly with scent (Spangenberg et al, 2005; Mattila & Wirtz, 2001).

Categorization Environmental factors

Ambient factors Air quality, Noise, Cleanliness

Design factors Aesthetic (materials, texture), Functional (layout, signage)

Baker (1986)

(19)

Bright versus dark side of marketing using ambient factors

Mostly, research has focused on how ambient factors can create approach behaviour. The dark side of marketing, however, is researched to a lesser extent. The dark side of marketing deals with studying the shopping behaviour of customers and identifying which environmental factors might “create negative consumer feelings during shopping” (d’Astous, 2000; p149). Studying the dark side of marketing is important, as “several studies have confirmed the prominence of negative information over positive information in the formation of consumer evaluations” (d’Astous, 2000; p149). In other words, customers may attach more importance to negative information than positive information when, for example, deciding to revisit a retail environment.

(20)

2.3.1 Music

Music is an ambient factor that can be highly controlled by sales personnel. It is also easily adjustable to fit to the preferences of the customers in a retail environment. However, the music of choice has often been determined by intuition or preference of the storeowner or employees, and less so by research (Milliman, 1982; Dubé & Morin, 2001). Choosing music based on research rather than common sense is important, as music has been defined as a “powerful means of influencing customers’ affective responses in retail environments, thereby influencing evaluations of, and behaviours within retail settings” (Spangenberg et al, 2005; p1584). What music should be played in a store to limit avoidance behaviour? The best music for a specific retail environment is determined by looking at the effects of presence, tempo, volume, familiarity, preferences, and congruency.8

The S-O-R model, or Stimulus, Organism, Response model by Mehrabian and Russel (1974), can be specified towards the music literature context. Based on a literature review, Garlin and Owen (2006) found that music could influence affective, financial, attitudinal/perceptional, temporal, and behavioural outcomes. Table 3 provides an overview of the various outcomes mentioned by Garlin and Owen (2006). Furthermore, the link of these outcomes with the S-O-R model is illustrated. Also, examples of studies discussing the outcomes are provided. Table 3 will form the basis of this section. Goal is to relate these outcomes to the underlying facets of the music construct (being tempo, volume, familiarity, preferences, and congruency), in order to find factors creating avoidance behaviour and possibly generating negative consumer feelings.

8_{Garlin & Owen (2006) mention the following facets of music: tempo, volume, complexity, genre,}

(21)

Elements S-O-R Discussed by:

Affective outcomes Mood, arousal,

emotion

O Lee et al, 2004; Tansik & Routheiaux, 1999; Hosea, 2004

Financial outcomes Gross margin, items

purchased R Milliman, 1982, 1986; Areni & Kim, 1993 Attitudinal/perceptional outcomes Product evaluation, brand loyalty, liking

R Spangenberg et al, 2005;

Yalch & Spangenberg, 2000; Dube & Morin, 2001

Temporal outcomes Duration perceived/

actual service time

R Milliman, 1982, 1986; Yalch

& Spangenberg, 1990; Areni & Kim, 1993; Spangenberg et al, 2005

Behavioural outcomes Patronage frequency, impulse behaviour

R Mattila & Wirtz, 2001; Milliman, 1982 Table 3: Music outcomes (Garlin & Owen, 2006)

2.3.2 Music presence

Customers often prefer stores that play music, over stores that do not play music (Burleson, 1979). The presence of music is likely to increase the level of pleasantness and arousal in a retail environment. Linsen (1975) also found that customers see the presence of music as a sign of customer service, and interest in customer wellbeing by store management. Music presence can also have a positive effect on patronage (Garlin & Owen, 2006).

Presence Low High

Affective outcomes Increase in arousal (Yalch &

Spangenberg, 1993)(Vanderark & Ely, 1993)

Attitudinal/perceptional outcomes

Increase in felt pleasure (Garlin & Owen, 2006)

Behavioural outcomes Positive effect on patronage

(Garlin & Owen, 2006) Table 4: Music presence9

As music presence only has positive influences on the behaviour of customers and their feelings, no hypotheses will be created based on music presence.

(22)

2.3.3 Music congruency

While the issue of congruency has been widely established in the scent literature, only few authors have mentioned the importance of congruency between music and the assortment or merchandise. Nevertheless, Areni and Kim (1993) state that the impact of congruency may hold for all atmospheric variables, including music. MacInnis and Park (1991) introduced the notion of fit or congruency in the music literature, and found that “music is more persuasive when it fits the persuasion context employed”. However, their research aimed on music in advertisements instead of music in a retail environment. The congruency concept in music literature has, until now, mostly focused on the congruency between music and another ambient factors (e.g. scent and music in the research of Spangenberg et al (2005)). Congruency between music and products, or music and store concept has not yet been discussed in literature.

When the music is not congruent to the items sold in the retail environment avoidance behaviour can also result. For example, clothing stores for elderly are not likely to play house, dance, or metal music as these musical genres are not congruent to the merchandise and the target group that is likely to visit the retail environment. As mentioned before, congruency has only been established in scent literature and is established as a variable that has a large impact on approach/ avoidance behaviour (Spangenberg et al, 2005). Therefore, decided is to research the effects of music congruency on avoidance behaviour and consumer feelings. Following from the discussion above, hypothesis 1 can be stated:

(23)

2.3.4 Music tempo

Music tempo, as a stimulus, has received more attention in music literature than music congruency. The observations made from the literature review on music tempo can be viewed in table 5.

Table 5: Music tempo10

The tempo of music is strongly related to the level of arousal in customers (Yalch & Spangenberg, 1993). High arousal typically results from fast tempo music, and the opposite is true for slow tempo music (Vanderark & Ely, 1993). Herrington and Capella (1996; p27) found that people “adjust their pace, either voluntarily or involuntarily, to match the tempo of music”. High arousal music or fast music can have two implications for the speed with which customers move through the store. On the one hand, “arousing music may result in customers quickly completing their shopping”, while on the other hand “customers might find the arousing music stimulating, and decide to explore more of the store” (Yalch & Spangenberg, 1993; p632).

In a restaurant setting, Milliman (1986) concluded that fast music leads customers to quickly finishing their meal, while slow music reduced the level of arousal in customers and provided for a relaxing environment, in which more time and money is being spent. Also, customers have a more favourable view of the retail environment when shopping in slow tempo music conditions (Garlin & Owen, 2006). Furthermore, customers often have a strong preference for a certain music tempo. This can create avoidance behaviour and negative consumer feelings when the music tempo does not match with the customers’ preferences.

10_{High: high music tempo / Low: low music tempo}

Tempo Low High

Affective outcomes Low arousal (Vanderark & Ely, 1993)

High arousal (Vanderark & Ely, 1993)

Financial outcomes Increase in spending

(Milliman, 1986) Attitudinal/perceptional

outcomes

More favourable view of the retail environment (Garlin & Owen, 2006)

Temporal outcomes Marginally more actual time spent (Garlin & Owen, 2006)

(24)

As the music literature has mixed results with regards to tempo and the implications of time and money spent a retail environment, decided was to include music tempo in the upcoming experiment. Expected is the following:

H2 Customers listening to high tempo music in retail settings are likely to display more avoidance behaviour and experience more negative consumer feelings than customers who are listening to slow tempo music.

Next to discussing the influences of music congruency and music tempo on avoidance behaviour and consumer feelings separately, the interaction between these two variables is also examined. Following from the previous discussions, the next hypotheses can be stated:

H3a Customers hearing slow incongruent music display less avoidance behaviour and experience less negative consumer feelings than customers hearing incongruent music that is fast.

(25)

2.3.5 Music volume

Another facet of music is music volume. Like the tempo of music, volume influences levels of arousal in customers. Higher volumes of music are likely to increase arousal and lead to faster shopping behaviour (Yalch & Spangenberg, 1993). While the time customers spent in a store decreases for loud music conditions, sales and reported satisfaction did not change in the research of Smith and Curnow (1965).11 However, “lower volumes tend to make customers shop at a more leisurely pace and in certain instances spend more money” (Herrington & Capella, 1996). Also, customers reported more favourable views of the retail environment when shopping in low music volume conditions (Garlin & Owen, 2006).

Volume Low High

Affective outcomes High arousal (Yalch & Spangenberg,

1993)

Financial outcomes Increase spending money

(Herrington & Capella, 1996)

Attitudinal/perceptional outcomes

More favourable view of retail environment (Garlin& Owen, 2006)

Temporal outcomes Marginally more time

spend (Garlin & Owen, 2006)

Less actual time spend (Yalch & Spangenberg, 1993)

Table 6: Music volume12

Still, music volume will not be included in the following research and experiment. As the objective is to place a computer simulation with questionnaire online, measuring volume might be complicated as volume levels on computers might differ significantly, and the music volume thus cannot be kept constant.

11

The volume levels were within the levels allowed by store management. What store management perceives as loud music may still be soft or moderate volume in the eyes of the consumer. Store management is likely to be cautious, as they want to avoid loosing customers.

(26)

2.3.6 Music familiarity & music preference

Music familiarity and preference are largely interconnected. While these concepts might be easily distinguished in literature, making a distinction between their influences on avoidance behaviour and consumer feelings might be less straightforward.

2.3.6.1 Music familiarity

Research shows that individuals believe they had spent more time in a store when listening to familiar music (Yalch & Spangenberg, 2000). However, reported actual shopping times were shorter than those in unfamiliar music conditions. Cognitive theory suggests that customers believe they are shopping for a longer period than they actual do because of the greater attention to the stimulus, in this case a particular song. According to Block (1990), “the increased cognitive processing associated with liked music could lead to the perception that more happened during the hearing of liked music, and thus augment perceived time”. Research does not clearly state whether it is best to play familiar or unfamiliar music. For example, when playing familiar music in a retail environment actual shopping times by customers are found to be shorter than for unfamiliar music, stimulating a feel that customers have accomplished their goal in a shorter period of time. Unfamiliar music, on the other hand, decreases the level of arousal in customers, but customers’ actual shopping times might be longer.

2.3.6.2 Music preference

(27)

Familiarity/ Preference Low High

Temporal outcomes Actual time spent

decreased (Yalch & Spangenberg, 2000)

Perceived time spent increased (Yalch & Spangenberg, 2000)

Behavioural outcomes Positive effect on patronage (Garlin &

Owen, 2006) Table 7: Music familiarity & music preference

As the two concepts of familiarity and preference are heavily intertwined, the following research will not focus on these two concepts. Correctly separating the effects of familiarity and preference is likely to be a large challenge, and results on these variables may be biased to a large extent.

2.4 Conceptual model

As discussed above, music, as a variable, has several facets. From these elements, music tempo and music congruency were selected as variables for the upcoming research, as these variables are likely to impact consumer feelings and approach/avoidance behaviour significantly. The resulting conceptual model can be seen in figure 3.

(28)

(29)

Chapter 3 Methodology

Chapter 3 introduces the research methodology used for this research. This section examines the research design, computer simulation, target population, and sampling design. Next, an overview of all operational variables is discussed and a pre-test conducted. This section ends with conclusions related to the pre-test. These conclusions will be used as input for the main experiment done to confirm/reject the hypotheses in chapter 4.

3.1 Research design

This research is based on a 2 x 2 (within participants) experimental research design. For this research, two variables are manipulated: 2 levels of music congruency x 2 levels of music tempo (see table 8). The 2 x 2 research design results in 4 experimental conditions. Per experimental condition between 20-30 respondents are included.13

Experimental groups Tempo

Music Slow music Fast music

Congruent A B

Congruency

Not congruent C D

Table 8: Research design

Before determining the type of research that was used for this research, the research designs of the articles examined in chapter 2 were reviewed. A small majority of the studies used field experiments, where data is collected using questionnaires, interviews or observations (see Appendix A). In true experiments, several options were used to recreate the retail environment. These included product displays, written descriptions, photos, videotapes and computer simulations. With true experiments, data was mostly gathered using questionnaires. In 2 studies, only questionnaires were used, without an experiment.

(30)

However, behaviour often does not lend itself to be measured solely by questionnaires (Malhotra, 2007).

The experiment used in this research can be characterized as a true experiment, as a simulated retail environment was used. Using a simulated retail environment will benefit from less intervening variables (e.g. promotions, other customers) and a higher level of control for the researcher. However, we have to take into account that results can be biased due to an increased evaluation apprehension and less natural experimental settings. Still, a true experiment will provide us with more possibilities to research the dark side of marketing. If the research was done in a real store environment (i.e. field experiment), restrictions to the range of music and the music tempo are likely to be put in place by store management in order to minimize the possibility of losing customers (Smith and Curnow, 1965).

To create more natural experimental settings, a computer simulation was used in this research. A computer simulation, as according to Massara and Pelloso (2006), could benefit a true experiment, as the setting will look more realistic than a single product display in a laboratory. Also, a computer simulation allows for interactions between the respondent and the environmental stimuli, which will actively involve the respondent in the experiment (Massara and Pelloso, 2006). The computer simulation will be discussed further in section 3.3. First, the target population and sampling procedure are examined.

3.2 Target population & sampling design

(31)

3.2.1 Target Population

A target population is the “collection of elements or objects that possess the information sought by the researcher and about which inferences are to be made (Malhotra, 2007; p.336). Malhotra (2007; p336) determines the target population by looking at the elements, sampling unit, extent and time frame. In this research, the target population can be defined as male or female (international) students able to perceive ambient factors (with an interest in mp3 players), located in the Netherlands (see table 9).

Target Population Definition

Elements Male or female students able to perceive ambient factors (with an interest in mp3 players)

Sampling unit Students

Extent Netherlands

Time 2010

Table 9: The target population

Social media and (student) email were used as media to contact the target population, as students are frequent customers of these particular media.

3.2.2 Sampling design

(32)

3.3 Computer simulation & website

To ensure that the retail environment in the computer simulation looks realistic, photographs from an existing retail environment were used as input for the computer simulation. Several store chains were approached. The Media Markt (store chain selling consumer electronics) in Groningen allowed us to take photos from the retail environment that could be integrated into the computer simulation.

3.3.1 Website

A special website was created to host the computer simulation and questionnaires.1415 Next to the actual computer simulation and questionnaires, information was provided regarding the goal of this research, confidentiality, and instructions on how to browse through the simulated retail environment. Information on the website was provided in English. The questionnaire, however, was available in both Dutch and English.16 The website containing the questionnaires was linked to the website that hosted the computer simulation, so that respondent did not have to open another Internet browser.

3.3.2 Product choice

The products on display in the computer simulation were mp3 players.17 In order to avoid biased results, the product that was displayed in the computer simulation should be gender-neutral; meaning that both males and females should be willing to buy or have an interest in the product (Fiore et al, 2000). This is crucial when the focus of a research is on something other than the product. For example, if the product were male oriented, females might react differently because of the product, and to a lesser extent due to the

14_{The questionnaire was pre-tested among 10 students; no major alterations were done after this review.} 15_{The website could be accessed using the following link: www.bramliemburg.com/linda. Current status}

website: offline.

16_{The questionnaire was translated using a translate-translate back method. The translation was done by}

second- year English students from the NHL.

(33)

ambient factor or the fabricated retail environment. Today, both males and females own an mp3 player. The product is thus relevant for both genders. Also, Fiore et al (2000) mention that products used in a research should not be too expensive and have a hedonic nature. Again, mp3 players comply with this description. Therefore, the mp3 player was chosen as the product to be displayed in the computer simulation.

3.3.3 Procedures

Potential respondents received an email with an invitation to participate in this research. Upon clicking the link that was provided in the email, respondents entered the website. After reading the information on the website, respondents accessed the computer simulation, where music starts playing (depending on the experimental condition). In the computer simulation, respondents are able to browse through the display of mp3 players, using their computer mouse. Clicking a selection of products (marked with a purple star), enlarged the product image and the accompanying product information. When respondents finished browsing through the selection of mp3 players, they were asked to proceed to the questionnaire. A selection of pictures of the website and computer simulation can also be found in figure 4 to 6. These, and other pictures can also be found in appendix F.

(34)

3.4 Questionnaire design & operational variables

In order to gather data and capture the opinion about the independent variables and their impact on the dependent variables, a questionnaire was attached to the computer simulation. This section discusses the design of the questionnaire and the measurement of the operational variables.

3.4.1 Questionnaire design

The questionnaire is based on a structured- indirect survey design with mostly fixed- alternative questions.18 The questionnaire itself is structured into several parts. The questionnaire can be described as having an indirect survey design, as the true purpose of the questionnaire is somewhat disguised.

3.4.1.1 Cover story

At the beginning of the questionnaire, the true purpose of the research (i.e. researching the influence of music on the behaviour of respondents) is not disclosed. Rather, the purpose will be described in a more general manner, stating that this research aims to get more insight into how customers purchase small electronic devices (mp3 players) and how the retail environment influences customers in general. When disclosing the true purpose of this research, the attention of the respondents might be solely on music rather than the retail environment or product in the computer simulation. This could lead to biased results.

3.4.1.2 Structure questionnaire

The questions range from general to specific. Questions regarding demographic information were asked first, after which questions regarding the dependent and independent variables were asked. The questionnaire ends with questions regarding music (i.e. questions on music congruency and music tempo). The complete questionnaire as used for the research can be found in appendix C. A short debriefing is placed at the end of the questionnaire to inform the respondents about the true purpose of this research and to thank the respondents for their participation.

(35)

The questions related to the dependent variables (consumer feelings and avoidance behaviour) and independent variables (music congruency and music tempo), are mostly measured using pre-validated scales from scientific literature. Please refer to appendix B for an overview of the scales used to measure the dependent and independent variables.

3.4.2 Operational variables - Dependent variables 3.4.2.1 Approach/ avoidance behaviour

Approach and avoidance behaviour are the two main responses to an environment according to the S-O-R framework of Mehrabian and Russel (1974).19 Approach behaviours are positive behaviours towards the retail environment and the merchandise within that environment, while avoidance behaviours represent the opposite (or negative behaviour towards the merchandise and retail environment). The concept of approach/avoidance behaviour has served as dependent variable in several studies (e.g. Spangenberg et al, 2006; Matilla and Wirtz, 2001; Donovan and Rossiter, 1982; Fiore et al, 2000; Chebat et al, 2009). In these studies, approach/avoidance behaviour is typically measured by 2 to 5 different variables. An overview of the variables used to measure approach/avoidance behaviour can be found in table 10.

Research Variables

Garlin & Owen (2006) Temporal outcomes (Duration, perceived vs. actual time) Behavioural outcomes (Patronage frequency)

Attitudinal/perceptional outcomes (Product evaluation) Spangenberg et al (2006) Perceived shopping time

Intent to visit store

Fiore et al (2000) Global attitude/product – global attitude/environment Purchase intention

Price willing to pay ($)

Chebat et al (2009) Intent to revisit

Table 10: Variables within the approach/avoidance behaviour concept

(36)

The variable constructs perceived/ actual time, product/ retail environment evaluation, purchase intention, intention to revisit, and price willingness to pay were addressed in the questionnaire. The variables were mostly measured on 7-point Likert scales. ‘Actual shopping time’, and ‘perceived shopping time’ are exceptions and will be measured by an open question.

3.4.2.2 Consumer feelings

In d’Astous (2000), where the ‘dark side of marketing’ was introduced, irritation was used as a measure for consumer feelings. He found that some environmental factors were more responsible for irritation than others (ambient factors being the factors that led to most irritation among customers). d’Astous (2000), however, recommended that it was also useful to use other emotions than irritation. For this research the CES, or Consumption Emotion Descriptors (Richins, 1997), will be used. The CES scale includes a diverse range of emotions. An overview can be found in table 11. A selection of these consumer feelings will be used in the experiment, to test the reaction of customers on the musical stimuli. In total 6 emotions from the Consumer Emotions Set will be used; 3 overall positive emotions (joy; excitement; surprise), and 3 overall negative emotions (anger; discontent; worry). These emotions are subdivided into a range of emotions. Respondents rated these 17 different emotions on a 7-point Likert scale.

Table 11: Consumer feelings

Positive consumer feelings Negative consumer feelings

Romantic love Anger

(37)

3.4.3 Operational variables - Independent variables 3.4.3.1 Music congruency

Congruency is measured in terms of perceived fit or appropriateness (Spangenberg et al, 2006; Bone & Ellen, 1999). What kind of music do respondents see as fitting with the image of the Media Markt? What do they expect to hear in the Media Markt? What customers see as (in)congruent will be determined by means of a pre-test, which will be further discussed in section 3.6. Statements related to music congruency were: ‘I expect to hear this song in the Media Markt’, ‘This music is suitable for the Media Markt’, and ‘It is likely that this music will be played in the Media Markt’. Responses were measured on a 7-point Likert scale.

3.4.3.2 Music tempo

Herrington and Capella distinguish between slow and fast paced music on the basis of research done by Milliman (1982, 1986). Music that has fewer than 72 BPM (beats per minute) is selected as slow music, while music that has 94 BPM or more is selected as fast music. Tempo was measured using Beatunes 2 (a computer program measuring the tempo of songs in your music library), ERGmusic.com and the BPM reference guide by Donny Brusca (2005). Statements related to music tempo were: ‘The tempo of music was too fast’, ‘The tempo of music was too slow’, and ‘The tempo of music was about right for the Media Markt environment’. Again, responses were measured on a 7-point Likert scale.

3.4.4 Other variables measured in the questionnaire

(38)

Furthermore, individual differences will be measured using an Optimum Stimulation Level (OSL) scale. The OSL literature argues that individuals differ in their optimum stimulation level. It depends per individual how loud or fast the music should be in order to achieve optimal sales or increase/decrease the time spend in the retail environment. The OSL literature features various scales, but the Change Seeker Index (CSI) has found support in numerous studies (Orth & Bourrain, 2005). The CSI scale of Steenkamp and Baumgartner (1995) will be used. This scale, consisting of 7 statements (measured on a 5-point Likert scale), may statistically not be the strongest one, but using the original scales (consisting of 32 to 95 statements) will not be beneficial to results as respondents may feel reluctant to fill out all the answers (Steenkamp and Baumgartner, 1992).

Furthermore, questions related to familiarity with the Media Markt, song preference, and music liking (i.e. ugly vs. beautiful, bad vs. good) were asked in the questionnaire. Other variables related to music (i.e. song preference or music liking) were included to check whether the selected songs only differ significantly on music congruency and music preference, and not on other facets as well. Also, brand preference and shopping enjoyment will be measured. Brand preference measures whether customers like to shop at the Media Markt (“When shopping, I always enjoy visiting the Media Markt”). Shopping enjoyment, on the other hand, measures whether customers enjoyed shopping in this particular retail environment (“Did you enjoy shopping in this retail environment?”).

3.5 Pre-test

It is imperative to know what customers see as congruent or incongruent music in a retail setting, as this will allow for more precise manipulation of the variable music congruency in the main experiment. Therefore, a pre-test was conducted to determine what people see as congruent music.20

20_{While pre-testing usually refers to “the testing of the questionnaire on a small sample of respondents to}

(39)

3.5.1 Procedures

A small sample of respondents from the target population (see section 3.4) was selected to participate in a pre-test. The questions in the pre-test questionnaire are related to music congruency. Goal was to research (a) which music is most typical for the Media Markt concept, (b) which music is seen as most likely to be played in the Media Markt, and (c) which music is seen as most typical for the mp3 player product category. The pre-test was held under 21 people, 11 females and 10 males, in the age of 18 to 28. The pre-test questionnaire can be found in Appendix D, whereas the results can be found in Appendix E.

Participants were given a list of music songs, both slow (<72 BPM) and fast paced (>94 BPM) music, and a set of questions to determine which songs have to be played during the actual experiment, in the four different experimental conditions (see Appendix C). The list of music was composed out of songs that were released between 1999 and 2010. Furthermore, all songs were listed in the Dutch top 40 and were released on a Hitzones CD (i.e. a CD that features the most popular pop songs released in a certain period of time). Also, it was made sure that various genres of music were included in the list to appeal to a wide variety of people. In total 11 slow songs, and 12 fast songs were selected, getting a total of 23 songs from which the respondent could choose (see Appendix D for an overview).

3.5.2 Music congruency with Media Markt concept21

The first two questions of the pre-test questionnaire were related to congruency with the Media Markt concept. Respondents were asked to rank their top 5 of songs that they found to be typical for Media Markt concept (congruent music) and their top 5 of songs they found to be atypical for the Media Markt (incongruent music). They were asked to select songs from high and low tempos. Overall, the results regarding these questions were quite conclusive.

(40)

Respondents highly agreed on what music is most typical for the Media Markt concept. When looking at the results (Appendix E), 67 % of respondents selected “Cold Play- La Viva La Vida” as highly congruent with the Media Markt concept from the list with high tempo music. From the list with low tempo music, 52% of respondents selected “Plain White T’s -Hey There Delilah”.

When looking at the results indicating what music respondents found to be incongruent with the Media Markt concept, the results were less conclusive. 52% of respondents agree that “Usher Ft. Will.I.am – OMG” is not congruent to the Media Markt concept, when selecting from the list with low tempo music. From the list with high tempo music, only 43% of respondents agree that “Katy Perry – I Kissed A Girl” was atypical for the Media Markt concept.

Looking only at the number one choices of respondents for both congruent and incongruent perceived music, results confirm initial outcomes (see Appendix D). Resulting from the pre-test and the accompanying results, the four songs displayed in table 12 will be used as background music in the computer simulation.

As results are not always distinctive, the questionnaire of the experiment will again contain questions about the perceived fit of the music with the Media Markt. Although the respondents chose both slow and faster paced music, music from the fast- paced music list (60%) were selected more often than music from the slow- paced music list (40%). These results were consistent for both males (61%) and females (57%). Remarkable is that while respondents marked fast-paced music as more typical for the Media Markt (61%), respondents also selected more songs from the fast- paced music when selecting music that they perceived as incongruent with the Media Markt concept (57%).

Table 12: Music per experimental condition

Tempo

Slow music Fast music

Congruent Plain White T’s – Hey There Delilah Cold Play- Viva La Vida

Congruency

(41)

Chapter 4 Results

This section describes the experimental conditions, respondents, pre- analysis steps and the actual analysis. The outcome of the analyses can be found in section 4.4.

4.1 Experimental conditions

There are 4 different experimental conditions in this research (see table 13). Each of the four experimental conditions yielded a high enough response, namely 30 respondents or more per condition (165 respondents in total). Data was collected within a time frame of approximately 2 months (31-7-2010 till 08-10-2010). This research used a between participants design, meaning that each respondent could only participate in one experimental condition, once. Below, a description of the respondents that participated in this research is given.

Experimental conditions Tempo

Music Slow music Fast music Congruent

Plain White T’s – Hey There Delilah

Cold Play – Viva La Vida Congruency Not Incongruent UstherUsher – Usher - OMG Katy Perry – I Kissed A Girl

Table 13: Experimental conditions

4.2 Respondents

(42)

Music Hey There Delilah – Plain White T’s Manipulation Slow/Congruent Time frame 31.07.2010 – 07.08.2010 Number of respondents N=30 60% female (N=18) 40% male (N=12)

Average age respondents 27 (range between 18 and 64)

Familiarity with Media Markt 87% (N=26)

Average time spent in computer simulation

6 minutes (range between 1 min. – 25 min.) Table 14: Respondents experimental condition A

Table 15: Respondents experimental condition B

Table 16: Respondents experimental condition C

Music I kissed a girl – Katy Perry

Manipulation Fast/Incongruent

Time frame 13.08.2010 – 29.08.2010

Number of respondents N=33

36% female (N=12) 64% male (N=21)

6 minutes (range between 1 min. – 20 min.) Table 17: Respondents experimental condition D

Music Viva la Vida – Coldplay

Manipulation Fast/Congruent

Time frame 07.08.2010 – 13.08.2010

62,5% female (N=20) 37,5% male (N=12)

11 minutes (range between 1 min. – 90 min.)

Music OMG – Usher ft Will.I.AM

Manipulation Slow/Incongruent

Time frame 29.08.2010 – 08.10.2010

45% female (N=15) 55% male (N=18)

(43)

4.3 Pre-analysis steps

This section discusses all necessary steps that should be performed before SPSS analysis can commence. This includes preparing the data generated by the questionnaires and creating constructs of all the questions related to a particular variable.

4.3.1 Before starting the analysis

After gathering the data, a few steps have to be undertaken before data analysis can start. Thesistools.com, the website used for creating the questionnaires, distributes the output in a standard excel format. While this format is easily exported to SPSS, it is the data itself that needs to be processed, and/or modified. Respondents have to be taken out that did not hear music, the format of data has to be synchronized, and variables recoded (i.e. consumer feelings, Intention to Revisit, and Optimal Stimulation Levels).

4.3.2 Checking the assumptions

In order to analyse data, variable constructs have to be formed using factor analysis. Doing a factor analysis requires the data to conform to several assumptions. The data has to be normally distributed, have a high internal consistency and be appropriate for performing a factor analysis (Hair et al, 2006). The constructs General Attitude Environment, General Attitude Product Merchandise, General Attitude Product Purchased, and Intention to Revisit, will be examined using factor analyses (see appendix G) (for an overview of all variable constructs please refer to appendix B). Results related to normal distribution, internal consistency and appropriateness of factor analysis can be found in appendix G (1 t/m 4). The results indicated that a factor analysis can be performed for all the constructs mentioned above.

(44)

feelings, on the other hand only had support for three factors, to which no good label could be attached. Decided was to take the average of all consumer feelings.

4.3.3 Manipulation check

A manipulation check was done to ensure that the manipulation of the variables was perceived as intended for this research. Two ANOVAs were done for this purpose. These ANOVAs show that there is a significant relationship between the perceived music tempo22 and the manipulated music tempo23 (F(1, 23) = 18.79, p<0.01). No such significant relationship existed between perceived music tempo and the manipulated music congruency (F(1, 2) = 1.577, p>0.05).There is a similar relationship for the perceived music congruency. A highly significant relationship exists between the manipulation variable music congruency24 and the perceived music congruency25 (F(1,26) = 18.541, p<0.01). However, there is also a significant relationship between the manipulation variable music tempo and perceived music congruency (F(1, 5.8) = 4.101, p<0,05). While this result is surprising, it is not expected to negatively influence the reliability and significance of results in later analyses. Please refer to appendix I for the ANOVA output related to the manipulation check.

4.4 Testing the hypotheses

Section 4.3 prepared the data for importation in SPSS and a factor analysis. These steps are necessary in order to start testing the hypotheses. Since the research includes several dependent variables, as well as multiple independent variables, a Multivariate Analyses Of Variance (MANOVA) was done next to an ANOVA in order to discover differences between the experimental conditions. Coupled with a MANOVA and ANOVA test, a planned contrast analysis26 was done to determine which underlying relationships are significant. This section describes how the hypotheses were tested and whether the hypotheses can be confirmed/rejected.

22_{Perceived Music Tempo – Music Tempo as indicated by respondents on the questionnaire.} 23_{Manipulation Music Tempo – Actual manipulation of tempo, as intended by the researcher.}

24_{Manipulation Music Congruency – Actual manipulation of congruency, as intended by the researcher.} 25_{Perceived Music Congruency – Music Congruency as indicated by respondents on the questionnaire.} 26_{A planned contrast analysis was used in this research, as post-hoc analyses could not be performed. The}

(45)

4.4.1 ANOVA consumer feelings27

First, an ANOVA was done related to consumer feelings. All variables that could have an influence on consumer feelings were included in the ANOVA test. Several relations appeared to be significant (see table 18/ appendix J).

4.4.1.1 Music congruency

As can be viewed in table 18, music congruency (p < 0.05) significantly influences consumer feelings, confirming hypothesis 1. Incongruent music can decrease happiness and enthusiasm, or increase irritation and frustration.

Table 18: ANOVA consumer feelings in- and output

27_{These results were obtained using a filter. Only respondents that rated their involvement with mp3} players as equal to or higher than 3 (on a 7-point Likert scale) were selected in the database (N=106). When including all respondents, relations to the construct of consumer feelings and approach/avoidance

behaviour were insignificant.

Input and results

Dependent variable _{Independent variable} F-value Significance level

Consumer feelings _{Manipulation Congruency} _4,361 _,040

Manipulation Tempo ,006 ,936 Music Liking 3,558 ,033 Shopping enjoyment ,821 ,444 Gender ,533 ,467 OSL ,112 ,894 Familiarity 2,689 ,105 Involvement 1,957 ,148 Brand preference ,084 ,920 Remembering stimulus 3,456 ,036 Song preference ,226 ,798 Genre preference ,354 ,703

Music and the Dark Side of Marketing -Music in the Media Markt-