IMAGE OR QUALITY?

(1)

IMAGE OR QUALITY?

The effect of brand image and objective quality on the

willingness to pay a price premium for national brands over

private labels and the moderating effects of consumption

situation and hedonic product value

Master thesis, Marketing Intelligence

University of Groningen, Faculty of Economics and Business

January 2017

MARIEKE SCHULENBURG

Student number: 2034042

Van Brakelplein 41a

9726 HG Groningen

Tel: +31(0)651013528

(2)

MANAGEMENT SUMMARY

Nowadays, the threat of private labels for national brands is more apparent than ever. A major reason behind this threat is the fact that retailers increasingly develop and market their own private label brands. The increasing amount of private label sales impose important and challenging issues for national brands. Where private labels used to be low-cost alternatives to national brands, there is a clear trend towards differentiation and quality improvements. As a result, consumers’ perceptions of private label value and quality perceptions have been increased in recent years. Private labels have obviously developed throughout time and they are no longer low-cost alternatives. Therefore, competition between private labels and national brands remain an important subject for research, however there is a lack of research into whether and when consumers are willing to continue to pay a price premium for national brands over private labels. While some studies show that quality still remains an important factor in explaining the price premium consumers are willing to pay, others show that the strong emphasis on quality as a determinant of price premium should be questioned. A stronger emphasis on the brand itself and brand image elements, rather than only on quality and price, could enhance a better understanding of price premiums for grocery brands. The price premium for national brands may arise from different sources. First, consumers may think that there is a quality gap between national brands and private labels that justify a price premium for national brands. On the other hand, consumers may also believe that the brand image or positive associations with the national brand exceed quality perceptions which justify a price premium. Additionally, the growth of private label brands differs among different product categories, therefore this study also focuses on the influence on hedonic product value and their role in determining a price premium for national brands. Furthermore, this study provides deeper insight into the role of private labels and national brands in consumers’ consumption pattern. Thus, this study provides answers to the following research question; To what extent are consumers willing to pay a price premium for NB’s over PL’s and

what influences this?’. This research question is subdivided in the following sub questions; 1) Do

(3)

quality and image dependent on consumption situation. The purpose of this study is to enhance

our understanding of what drives consumers to pay a price premium for NBs over PLs in the consumer packaged food market.

A quantitative study with an online survey is used to gather the data. Participants in this research are exposed to four different product categories, three that dominate on hedonic product value (chips, cola and chocolate) and one that dominates on utilitarian product value (spaghetti). A model of the price premium consumers’ are willing to pay for NB’s over PL’s is developed based on the responses of 192 participants. First, a chow-test is executed in order to determine whether it was allowed to pool the data among the different products categories. This statistic showed that it was statistically not justified to pool the data, therefore further analyzes are done unit-by-unit. The program Latent Gold is used to analyze the data since it enables to run a regression analysis while at the same time specify different numbers of classes.

(4)

PREFACE

This Master Thesis has been written to accomplish my Msc. Marketing with a specialization in Marketing Intelligence at the University of Groningen. With this Master thesis I finish my life as a student in the wonderful city of Groningen and my ‘real’ working life is going to begin. I can’t wait….

Hereby I would like to take the opportunity to thank all the people who supported me and helped me during the writing of my thesis. First, and foremost, I would like to thank my thesis

supervisor Erjen van Nierop for his valuable, patient and friendly support. Furthermore, I would like to thank the respondents who participated in this study for providing me with relevant information. I also thank my friends and family for their unceasing encouragement and support.

I proudly present my Master Thesis about the willingness to pay a price premium for national brands over private labels. In my opinion this thesis has become an interesting, practical and also academic piece of reading. With writing this thesis I have developed my own knowledge about this topic but most importantly increased my statistical skills and I have learned to keep my patience and make realistic schedules.

Enjoy reading!

(5)

1. INTRODUCTION

Several studies outline the threat of private labels for national brands. A major reason behind this threat has been that retailers increasingly develop and market their own store brands, or private labels (Verhoef, Nijssen & Sloot, 2002). A recent study conducted by ACNielsen (2014) points out that the growth in the private label market is significantly increasing nowadays. Where private labels first emerged as low-cost alternatives to national brands, there is a clear trend towards differentiation for private labels instead of just a focus on simplicity and low prices (Anselmsson, Johansson & Persson, 2007; Martos-Partal, Gonzalez-Benito & Fustinoni-Venturini, 2015). As a result, the increasing amount of private label sales pose important challenges for national brands around the world, and private labels are becoming a major concern for brand manufacturers (Steenkamp, van Heerde & Geyskens, 2010; Verhoef et al., 2000). Nowadays, retailers offer a broad range of diversified and high-quality private label products (Martos-Partal et al., 2015). Some studies imply that on average the quality gap between NB’s and PL’s are very small (Kadirov, 2015), which influences consumers’ attitude towards private labels. The Nielsen Global PL survey shows that consumers perceptions of private label value and quality has significantly increased recent years (Nielsen, 2014). Especially on the European market, consumers have a strongly positive attitude towards private labels (Steenkamp et al., 2010). Private labels have clearly developed throughout time and they are no low-cost alternatives anymore. According to de Wulf, Odekerken-Schröder and Bakpayev (2015) private labels play different roles that fulfill a variety of consumer needs, which pose different implications for retailers and manufacturers.

(8)

actual price differential between the brands (Sethuraman, 1999). Therefore, insights into the factors that explain the price premium consumers are willing to pay for national brands would enhance our understanding of why some consumers buy national brands while others prefer to purchase store brands.

Previous research has primarily focused on two different dimensions of positioning, price and quality (Anselmsson et al., 2007; Bronnenberg and Wathieu, 1996; Hoch, 1996; Ghose and Lowengart, 2001). However, literature describes conflicting findings concerning the effect of the quality gap between private labels and national brands and the willingness to pay a price premium. While some researcher argue that the quality gap still remains an important incentive for consumers to pay a price premium (Steenkamp et al. 2010; Tikkanen & Vääriskoski, 2010), others suggest that consumers increasingly consider private labels to be as good as national brands (Rossi et al., 2015), and that the strong emphasis on quality as a determinant of price premiums should be questioned (Anselmsson et al., 2007). Even if there is still a perceived quality gap between NB’s and PL’s, this might not be sustainable in the long run (Kadirov, 2015). According to Anselmsson et al. (2007), a stronger emphasis on the brand itself and the brand image, rather than quality and price, could enhance a better understanding of price premiums for grocery brands. According to them there is little research done into which specific brand image elements influence price premiums.

Customer-based brand equity has often been mentioned as synonymous with price premium, customers’ willingness to pay for different brands. Brand equity often includes a number of dimensions and different frameworks include different dimensions. The most frequently used framework is from Aaker (1991) and includes awareness, loyalty, perceived quality and associations. More recently, Andersson et al. (2007) developed and refined a conceptual brand equity framework for grocery products based on a qualitative study which include perceived quality, brand associations, loyalty, uniqueness and awareness.

(9)

difference in quality, but may still want to pay a price premium because of image or positive association with the national brand that exceed quality perceptions. In order to determine which elements of brand equity (quality vs non-quality) contributes most to the willingness to pay a price premium for national brands over private labels, this research will focus on separating quality and non-quality constructs.

As some other researchers revealed, the growth of store brands varies across different product categories (Batra & Sinha, 2000). It is possible that quality is more important in some categories while in other categories non-quality elements are more important. Sethuraman & Gielens (2014) argue that only a few studies focused on the influence of different product characteristics such as functional versus hedonic and durable versus non-durable and they suggest that future research should include more product characteristics. Moreover, they argue that future research should include more consumer characteristics, such as usage intention. Are national brands more likely to be purchased for specific consumption moments, whereas private labels are more common in other situations? Deeper insights into the role that private labels and national brand play in consumer’s consumption patterns might help manufacturers (Sethuraman & Gielens, 2014). It is possible that consumers perceive brand image to be more important when consuming a product in a social setting, which in turn influences the willingness to pay a price premium (WTPPP). Summarized this research attempts to answer the question:

To what extent are consumers willing to pay a price premium for NB’s over PL’s and what influences this?

The research question will be answered by the use of the following sub questions. 1) Do objective

quality and brand image differential impact the willingness to pay a price premium for NB’s over PL’s 2) Which brand image related aspects influence the willingness to pay a price premium for NB’s 3) To what extent is the relationship between objective quality and brand image dependent on the hedonic product value? 4) To what extent is the relationship between objective quality differential and brand image dependent on consumption situation. The purpose of this study is to

(10)

By answering these research question, this study makes several practical and academic contributions. First of all, this research attempts to close the gap in literature related to the influence of quality and brand image on the WTPPP. Furthermore, it contributes to the existing brand equity literature by showing which brand image elements influence consumers willingness to pay for national brands. Secondly, by understanding which factors influence the price premium customers are willing to pay for national brands, brand manufacturers can better develop effective marketing strategies. When it turns out that objective quality differential is the main reason that consumers pay a price premium for national brands over private labels, managers should focus on improving their product in order to increase the objective quality of their products. On the other hand, if non-quality elements outweigh non-quality elements, managers should focus on improving their brand equity through image-building strategies. Additionally, it is useful for managers to know how this differs between different product categories and consumption situations. Moreover, it will help managers by knowing which consumers care more about quality/image and for what products? This can help brand managers focus on the right consumer segments and design appropriate marketing strategies.

(11)

2. THEORETICAL BACKGROUND

2.1 Willingness to pay a price premium

The willingness to pay a price premium reflects the ability of a brand to set their prices higher than those of competitors (de Chernatony & McDonald, 2003). So, brands command price premiums when the amount that customers are willing to pay for a particular brand is higher than for products out of the same product category from other relevant brands (Aaker, 1996). According to Agarwal and Rao (1996) the WTPPP is the best measure that can explain choices of different brands at both an individual level as well as aggregated market shares. In order to develop optimal pricing strategies, marketing managers depend on this measure. The price of a product indicates the sacrifices that consumers make when purchasing a product related to the expected value of that product (Ismail, Melewar, Lim & Woodside, 2011). Since the aim of this study is to compare choices of products within grocery product categories (e.g. national brands versus private labels), the WTPPP seems a logical and useful measure (de Cernatony & McDonald, 2003). However, it should be noted that the willingness to pay a price premium is a relative measure and does not necessarily indicate real prices. In this study it is used as a comparative measure that compares the amount of a price premium that customers can be willing to pay more for one brand (NB) over another (PL).

2.2 Quality

The traditional view on national brand and private labels suggests that PL’s are generally lower in quality than NB’s (Ailawadi, Neslin & Gedenk, 2001). However, recently, retailers have given greater importance to the quality of private labels and are trying to close the quality gap. However, literature still describes conflicting findings concerning the effect of the quality gap between private labels and national brands and the WTPPP. Therefore, this research aims to give insight into this issue by investigating whether objective quality as rated influences the willingness to pay a price premium.

(12)

quality. Whereas objective quality is defined as the actual superiority or excellence of the products which determines the actual positioning of the product, subjective quality is a subjective impression of quality that exists solely in consumers’ minds, which is related to brand image. Since the aim of this study is to focus on the difference between quality and image, the focus will be on objective quality while perceived quality is seen as an outcome of brand image. In this research we use objective quality labels, as used by experts at the Dutch Customer Association (Consumentenbond) in order to communicate objective quality.

Apelbaum, Gerstner & Naik (2003) show in their research that national brands receive on average 28.7% price premium when their average objective quality is lower compared to an average private label brands. This price premium increases to 50.4% when the average objective quality of national brands is higher. In addition, a growing focus has been paid to determine which factors are used by consumers to deduce quality of food products and how they might influence purchase decisions (Aprile, Caputo & Nayga, 2012). Ingredients and nutrition information are often used as extrinsic attributes that serve as a quality sign of different products (Anselmsson et al. 2007, Oude Ophuis & Van Trijp, 1995). Ingredients are often linked to the quality of the ingredients and the naturalness (Anselmsson et al. 2007). Also taste is an important quality sign (Anselmsson et al. 2007). According to Anselmsson et al. (2007), quality means that products of a certain brand should be consistent, meet customers’ expectations and also be higher compared to other products and brands.

(13)

Steenkamp et al. (2010) argue that, notwithstanding the progress in product quality for PL’s, consumers are still more uncertain about the quality of PL’s compared to NB’s. Especially when products are difficult to produce, consumers believe that NB manufacturers have more expertise to produce high-quality products which increases the quality gap. Additionally, NB’s actively compete continuously on innovations and continuous improvements of their products. Since, retailers have to manage a broad range of different product categories, they often lack the financial resources and technical expertise to catch up (Steenkamp et al., 2010; Steenkamp & Gijskens, 2014). So, these continuous program of small improvements impacts the quality of products and puts PL’s constantly on a distance which results in the fact that the comparative quality of PLs is lower in some categories.

Moreover, closely related to perceived quality, is the perceived risk that consumers notice related to the consequences of purchasing a store brand (Mieres, Martín & Gutiérrez, 2005; Sethuraman & Cole, 1999). As argued above, the perceived inferiority in quality of private labels brings along a source of uncertainty for end user of the products which increases the risk related to the purchase (Mieres et al., 2005). So, the difference in quality decreases the perceived value of private labels through the increased risk in purchasing private labels (Richardson, Dick & Jain, 1994; Mieres et al., 2005). Taken this together it can be argued that increasing the objective quality differential between NB’s and PL’s and communicating this by using objective quality labels might influence customers to pay a price premium for NB’s. Based on the above, the following hypothesis can be formulated;

H1: Objective quality differential is positively related to the willingness to pay a price premium for NB’s.

2.3 Brand Image

(14)

non-product quality (brand image) related aspects in order to broaden the understanding about price premiums for branded products.

Brand image has been defined by Keller (1993) as “the information linked to a brand in customer

memory” (figure 1.). An underlying assumption of brand image is that a favorable, strong and

unique brand image creates a sustainable competitive advantage that will deliver economic returns (Aaker, 1991; Keller, 1998). From a theoretical and managerial perspective, brand image might be one of the most interesting aspects because it deals with all customer-based brand associations that companies can impact with a variety of efforts (Bondesson, 2013). In order to understand how brand image delivers value and add profits, it can be put in a broader brand equity framework (Persson, 2010; Anselmsson et al., 2007). Therefore, this research utilizes theories on customer-based brand equity. This research stream mainly focuses on non-product related customer perceptions that add value to products (Keller, 1993). Customer-based brand equity mainly put focus on removing the tangible economic value derived from brands, which makes it in particular relevant for the establishment of price premiums (Anselmsson et al. 2007). Lassar, Mittal and Sharma (1995), argue that brand equity can be broken down into that which;

• Indicates consumer perceptions rather than objective dimensions • A global evaluation related to the perceived value of the brand • Originated from the brand name rather than from physical elements

• It is a relative measure that should be evaluated in relation to other relevant competitors

(15)

premium is a distinct and relevant consequence of a favorable brand image (Aaker, 1996; Anselmsson et al., 2007; Persson, 2010). Brand strength, in contrast, is most often described as a global evaluation or an intention to behave, such as an intention to buy or pay for a brand (Netemeyer et al., 2004). Examples of general brand strengths and outcomes of brand image are price premiums, loyalty and satisfaction (Aaker, 1996; Netemeyer et al., 2004; Keller, 2001).

BRAND EQUITY

2.4 Brand Image Elements

The existing brand equity literature includes various different brand image models (Aaker, 1996; Keller, 2000). Some models are developed with the idea of being appropriate for various product types (Aaker, 1996; Netemeyer et al., 2004; Yoo and Donthu, 2001), while others are more focused on specific products (Anselmsson et al. 2007; Tikkanen & Vääriskoski, 2010). In the literature into consumer packaged foods, brand equity has been studied by Tikkanen and Vääriskoski (2010), Anselmsson et al. (2007), Netemeyer et al. (2004) and Sethuraman (1999). Aaker (1996) considers awareness, loyalty, perceived quality and associations as important elements. Yoo and Donthu (2001) suggested a multidimensional consumer-based brand equity model by assuming brand loyalty, perceived quality, brand awareness and brand associations to be important. Keller (1993) also points out brand awareness and brand associations as key underpinnings of consumer-based brand equity. More recently, Andersson et al. (2007) developed and refined a conceptual brand equity framework for grocery products based on a qualitative study which include perceived

BRAND IMAGE

BRAND STRENGTH

BRAND VALUE

Any information linked to the brand in customers’

memory.

The economic value of the brand to the

brand-owning firm. Customers’ global evaluation

of, and behavioral response towards the brand.

• Brand Loyalty • Price Premium

(16)

quality, brand associations, loyalty, uniqueness and awareness. In the study of Tikkanen and Vääriskoski (2010), quality, perceived brand value, uniqueness and brand associations are identified. The criteria under associations can be further categorized as origin, organisational associations, environment/animal friendliness and social image. Social image has also been shown to be a driver for a price premium for food brands in other exploratory studies (Lassar et al., 2007; Tikkanen & Vääriskoski, 2010). Considering the various suggestions, this research recognizes brand awareness, social image and brand uniqueness as the most important elements of brand image that are expected to influences brand strength in term of the WTPPP.

2.4.1. Brand Awareness

According to Keller (1993), brand awareness is reflected in the ability to recognize the brand under various circumstances and gives a reassuring message. Although it is often measured at the individual level, it is actually a collective phenomenon (Kapferer, 2008). When a brand is collectively known, each individual knows it is known, this often results in spontaneous inferences. Brand awareness is mostly correlated with trust, reliability, closeness to people, traditional styling and accessibility (Schuiling & Kapferer, 2004). According to Keller (1993), brand awareness is mainly important in low-involvement categories (Keller, 1993). Since, groceries can broadly be described as low-involvement products (Anselmsson, 2007), it is assumed that brand awareness is an important element in the willingness to pay a price premium for national brands. Consumers nowadays are often looking for ways to minimize time spent on doing groceries which reduces detailed considerations of alternatives (Silavoi & Speece, 2004), hereby brands convey certitude and trust (Kapferer, 2008). So, strong brand names serve as a time and risk reducer while doing groceries.

H2a: National Brand Awareness will have a positive impact on the willingness to pay a price premium for NB’s over PL’s

2.4.2. Brand Uniqueness

According to Netemeyer et al. (2004, p. 211), brand uniqueness refers to “to what degree

(17)

unique selling propositions and differentiation. Also Keller (1993) underlines the degree of uniqueness in brand’s associations as an important element. When consumers perceive a brand as unique, they see it as outstanding and lack of substitutability (Kapferer, 2008). According to Anselmsson et al. (2007) uniqueness is mainly important for retailers, since supermarkets often sell over hundred brands and thousands of articles. What motivates customers to choose one product over another? Brands help define what makes a product unique, different and more desirable than other products. When a brand is able to offer something ‘unique’ that other brands can not accomplish, consumers are motivated to pay more for that brand than for other competing brands. Often the name of national brands yields uniqueness of offer and an added value that competitors do not have (Kapferer, 2008). Based on these findings, the following hypothesis can be formulated;

H2b: Brand Uniqueness will have a positive impact on the willingness to pay a price premium for NB’s over PL’s

2.4.3. Social Brand Image

The social role and symbolic meaning of brands in general is often highlighted in customer brand equity literature (Biel, 1992). Social image is related to the way consumers are willing to express themselves and their ideal selves (Ball & Tasaki, 1992; Anselmsson, 2007). Brands help us to express who we are and who we want to be and consumers often choose to purchase products that maintain or enhance their self-image (Graeff, 1992). Often, people become emotionally attached to certain brands and see them as part of their self-image (Haskins, 2015). Additionally, the purchase of a certain brand is often used to reflect personalities and to remain a certain self-perception (Ball & Tasaki, 1992). Based on this it is assumed that national brands more often communicate a desirable social image that consumers desire in order to improve or maintain their self-image.

(18)

2.5 The Moderating Effect of Hedonic Product Value

Insights into factors that influence price premiums for national brands can offer brand manufacturers guidance in developing appropriate marketing strategies. However, it is likely that this differs between product categories. For example, according to a study conducted by ACNielsen (2014) private label success is biggest in more commodity-driven and high-purchase categories, so mainly in categories where consumers perceive little differentiation. Market shares of private labels vary considerable across different product categories, however only a few studies have focused on different product characteristics that influence private label share (Sethuraman & Gielens, 2014). Therefore, this research attempts to obtain knowledge on the moderating effect of hedonic product value on the relationship between objective quality differential and brand image on the willingness to pay a price premium.

This study focuses on four different two product categories namely three which are seen as superior on hedonic dimensions and one that is seen as superior on utilitarian dimensions. Broadly speaking, hedonic products provide more experiential consumption, excitement and pleasure while utilitarian products are mainly functional and instrumental ((Dhar & Wertenbroch, 1999). Batra & Sinha (2000) found that consumers often switch from a PL to a national brand if they perceive that the degree of uncertainty about the quality of a PL is high. Especially in categories characterized with more hedonic and experience benefits such as coffee or chocolate compared to more functional products that serve a more functional benefit. This leads to greater felt purchase anxiety about the quality equivalence between private label products and national brands which reduces PL purchases. Therefore, it is expected that consumers are more likely to pay a price premium when the product they buy score higher on consumption pleasure (e.g. hedonic) instead of more functional and utilitarian products.

H3a: The level of hedonic product value positively moderates the relationship between quality and the willingness to pay a price premium for NB’s.

(19)

give the personal gratification, excitement and status desired by consumers, especially when they buy hedonic products. When an individual is highly involved with a certain product category (mostly hedonic categories) it is more likely that the choice is based on self-concept and the congruence between the brand image and their personal self-concept (Xue, 2008). Branded products are more often seen as products that serve as a status and image indicator that influence the buyer of the product (DelVechhio, 2001). Also Fischer et al. (2010) argue that when brands are more relevant to customers, customers should be willing to pay more for a brand name. According to Steenkamp and Geyskens (2014) national brands are more successful in product categories that score high on symbolic or sign value, since national brands communicate a higher brand imagery compared to PL’s, which are more often sold based on their functionality. Based on this it is expected that especially in hedonic product categories, consumers give greater importance to brand image aspects and are willing to pay more for branded products.

H3b: The level of hedonic product value positively moderates the relationship between brand awareness and the willingness to pay a price premium for NB’s.

H3c: The level of hedonic product value positively moderates the relationship between brand uniqueness and the willingness to pay a price premium for NB’s.

H3d: The level of hedonic product value positively moderates the relationship between social brand image and the willingness to pay a price premium for NB’s.

2.6 The Social Context of Consumption

(20)

that influences the social comparison of consumers is called public-self-consciousness (PSC). This is defined by Fenigstein, Scheier & Buss (1975, pp. 321) as “the consistent tendency of persons to

direct their attention toward themselves as social objects”. This means that consumers with a high

PSC are more aware of the opinions of others. Bearden and Rose (1990) found that individuals with a high PSC score have the tendency to be conscious about the fact that people around them form impressions and opinions based on the product they choose and their usage behavior. Additionally, the tendency to purchase the cheapest product is lower since consumers do not want to be seen as cheap or unable to afford higher prices product. People that know beforehand that they are going to consumer the products in a social setting are more likely to change their brand preferences (Gainer, 1995). Also Bearden and Etzel (1982) underline this finding by showing in their research that the social pressure on a brand decision is higher when the product is bought for hedonic reasons and intended to be consumed publicly. According to Müller et al. (2006) national brand products are more valuable when they are consumed in a social setting, for example with friends and other family members, which decreases the value of private labels. Based on this we can conclude that the willingness to pay a price premium is situation-dependent, therefore the following hypothesis can be formulated;

H4: The social context of consumption will positively influence the willingness to pay a price premium for NB’s over PL’s.

2.7 The Moderating Effect of the Social Context of Consumption

(21)

social recognition and status and the base for self-worth is the approval of other individuals (Ball & Tasaki, 1992). So consumers are more likely to change their brand preferences if they know that they are going to consume a certain product in a social context based on improving their public-self. Based on the above, it is expected that the different brand image elements are positively moderated by the social context of consumption. Therefore, the following hypotheses can be formulated;

H5b: The social context of consumption will positively moderate the relationship between

brand awareness and the WTP a price premium for NB’s over PL’s.

H5c: The social context of consumption will positively moderate the relationship between

brand uniqueness and the WTP a price premium for NB’s over PL’s.

H5d: The social context of consumption will positively moderate the relationship between

social image and the WTP a price premium for NB’s over PL’s.

According to Baltas (1997) national brands offer a more secure alternative that in social consumption situations are more accepted. Therefore, some consumers hesitate to buy PL brands because they are afraid for likely negative social consequences due to their product choice (DelVechhio, 2001). People want to show that they care by buying the best quality products and do not want to be perceived as cheap (Müller et al. 2006). Additionally, Semeijn (2004) found in his research that consumers are especially willing to pay more for branded products in a social setting when the quality variance between NB’s and PL’s in a certain product category is high. Therefore, it is expected that;

H5a: The social context of consumption will positively moderate the relationship between

quality and the WTP a price premium for NB’s over PL’s.

2.8 Socio-demographic Characteristics

(22)

(23)

2.9 Conceptual Model

An overview of the hypotheses and relationships addressed in this research can be found in the conceptual model depicted in Figure 2.

Figure 2: Conceptual Model

Objective Quality

Differential

Brand Image

- Brand Awareness - Brand Uniqueness - Social Image

Willingness to pay a

price premium

Hedonic product value

(24)

3. RESEARCH DESIGN

3.1 Design, participants and procedure

The aim of this study is to determine whether brand image and quality differential between NB’s and PL’s influence the willingness to pay a price premium with the moderating effect of product category and consumption situation. Data is collected using an online questionnaire designed by the survey platform Qualtrics (Appendix B.). The online questionnaire is distributed by the social media network Facebook, forums and face-to-face contact by asking consumers outside the supermarket Albert Heijn, Jumbo and PLUS. Participants are motivated to fill in the survey by a chance to win a gift card worth €25. Besides that, participants were free to fill in the survey in their own time and anonymity was ensured. Since this study is focused on the Netherlands, the core objective of the participants is to have a Dutch nationality or have resided for a longer time in the Netherlands. The measurement scales in this research are developed and validated in English but will be translated in Dutch. This will be needed since the targeted population is Dutch. However, it should be mentioned that the process of translating the concepts can be fraught with problems of semantics (Chang et al. 1999). When translating a questionnaire, it is important that the distortion from the source culture needs to be reduced (Chang et al. 1999). In order to reduce this distortion, translation of the constructs is done carefully by comparing different translation tools. Additionally, translations are thoroughly discussed with other peers by using back-translation.

3.2 Measures

(25)

Variant 1 Variant 2 Variant 3 Variant 4

Chips Quality label: Yes Situation: Social

Quality label: No Situation: Social

Quality label: Yes Situation: Non-social

Chocolate Quality label: No Situation: Non-social

Spaghetti Quality label: No Situation: Social

Quality label: Yes Situation: Social

Quality label: No Situation: Non-social

Cola Quality label: Yes Situation: Non-social

Quality label: No Situation: Non-social

Quality label: Yes Situation: Social

Figure 3: Survey variations

Brand image will be measured using three categories, brand awareness, brand uniqueness and social image. Brand awareness is defined as “the ability for a buyer to recognize or recall that a

brand is a member of a certain product category” (Aaker, 1991, p. 61). This implies that brand

awareness consists out of brand recall and brand recognition (Yoo and Donthu, 2001). Brand awareness will be measured with 2 items based on research by Yoo and Donthu (2001). In their research they made a distinction between brand awareness and brand associations. Brand awareness is measured using 2 items. An example of a statement is: “I can recognize brand X

among other competing brands”. Brand uniqueness will be measured using the scale developed by

Netemeyer et al. (2004). Uniqueness is defined as “the degree to which respondents feel that the brand is different from other competing brands” (Nettelmeyer et al. 2004: pp. 211). Respondents are asked to indicate their agreement on 3 statements on a 7-point likert scale. An example of a question is; (Brand name) really “stands out” from other brands of (product category). Lastly, social image will be measured using the scale validated by Sweeney and Soutar (2001). They define social value as “the utility derived from the product’s ability to enhance social self-concept” (Sweeney & Soutar, 2001, p. 211). Participants indicate their agreement on three statements. One example is “Buying products of this brand would help me to feel accepted”. The quality gap and image gap are measured by subtracting the quality and image scores of NB’s by those of PL’s.

(26)

like toilet paper, laundry detergent and milk (Sloot et al. 2002). Since, this study focus on hedonic product value and the relationship toward the willingness to pay a price premium, three products that dominate on hedonic features are included and one products that dominates on utilitarian features. The products that are chosen for this study are chocolate, chips and cola for products that dominate on hedonic features and spaghetti that dominate on utilitarian features. However, even though some products are on average more hedonic or utilitarian, this may differ between consumers (Sloot et al. 2002; Okada, 2005). Therefore, hedonic product value will be measured using the scale adapted from Okada (2005). In their research they compared a single-item scale (Dhar & Wertenbroch, 2000; O’Curry & Strahilevitz, 2001) with a multiple-item scale (Crowley, Spangenberg & Hughes, 1992; Voss, Spangenberg & Grohmann, 2003). He found significant and positive correlations between the single-item scale and the sum of the multiple-item scale for both hedonism and utilitarianism (Okada, 2005). So both scales capture comparable concepts and attitudes. Therefore, this research will adapt the single-item measure for the purpose of parsimony (Okada, 2015). Participants are asked to rate each of the four products on two scales; a hedonic scale of 0 to 6, where 0 = “not al all hedonic” and 6 = “extremely hedonic” and likewise a utilitarian scale of 0 to 6. Since it is likely that participants do not know the meaning of these words, they will be explained in the survey. Hedonic product value is defined as “products that can be

described with affective characteristics that provide a pleasant, nice and comfortable feeling”,

whereas utilitarian product value is defined as “products that can be described as functional and

useful”.

(27)

3.3 Questionnaire Design

The online questionnaire is divided into three different sections (Appendix B.). The first section gives a small introduction about the study and the questionnaire along with a confidentiality statement. In this section participant are also asked about their gender, age, family size, income and education. It is expected that age, income, family size and education influence buying behavior (Ailawadi et al., 2010; Urbany et al., 1996). The answers related to income are based on the Dutch CPB (Centraal Plan Bureau), who indicated that the modal gross margin income in 2015 was around 34.000. Therefore, the category option 30.000 – 40.000 is the middle out of 5 categories. As a result, we can divide participants as considerably beneath modal, beneath modal, modal, above modal and considerably above modal. In the second part participants are asked to indicate their agreement on different statements related to the awareness, uniqueness and social image of the products on a 7 point-likert scale, with one indicating fully disagree and seven indicate fully agree. In the third section respondents are asked how much more they are willing to pay for a certain national brand over private label in a given category, pictures of both products are shown. National brands in a certain category are based on the leading national brands based on market share in the Netherlands. Respondents rated both products (PL and NB) at the same time, which made it easier for them to compare the products. This is repeated for 4 different products in different categories. The questionnaire ended with the possibility to fill-in their email address in order to make chance to win a gift card worth €25.

3.4 Analytical procedure

(28)

4. RESULTS

4.1 Descriptives

(29)

Gross monthly income -‐ < €1.000 -‐ €2.001 - €3.000 -‐ €3.001 - €4.000 -‐ €4.001 – €5.000 -‐ > €5.000 14,6% 14,1% 23,4% 20,3% 15,6% 12,0% 9,9% 9,1% 23,3% 29,1% 15,2% 13,4% Education -‐ Primary school -‐ LBO/VMBO -‐ HAVO -‐ VWO -‐ MBO -‐ HBO -‐ WO 0,5% 3,6% 5,7% 2,1% 24,5% 40,1% 23,4% 1,7% 8,0% 8,1% 1,5% 32,0% 30,8% 17,9% Supermarket Preference -‐ Albert Heijn -‐ Jumbo -‐ PLUS 61,5% 34,4% 4,2% 55,6% 34,8% 15,6% Table 1: Descriptives 4.2 Reliability

(30)

Construct Cronbach’s Alpha (number of items) Brand Awareness Private Label National Brand 0,837 (2) 0,899 (2) Brand Uniqueness Private Label National Brand 0,862 (3) 0,895 (3) Social Image Private Label National Brand 0,904 (3) 0,937 (3)

Table 2. Reliability statistics

4.3 Willingness to pay a price premium

The WTPPP for NB’s over PL’s is scored using real amounts in eurocents. In order to use them in further analysis, these amounts are rescaled into percentages using the average private label price of the corresponding products. The overall average WTPPP and the scores for the different product categories can be found in table 3. Additionally, a one-way ANOVA test is used to test whether the mean score for WTPPP significantly differ between the four product categories. This test is significant (p<0.01) and therefore there is a statistically significant difference in the WTPPP for NB’s between the four different product categories. In order to see which of the specific product categories differed, the results of the Tukey post hoc test are taken into account. The Tukey post hoc test shows that all different product categories significantly differ from each other (p<0.05) except for the chips and chocolate category (p=0.573).

Variant Average willingness to pay price premium (%)

(31)

Furthermore, the brand image gap scores are measured by subtracting the scores of the private labels from the national brand scores (Erdem, Swait & Valenzuela, 2006). As a result, the perceived brand awareness, brand uniqueness and social image gaps ranges from -6 (private labels are rated higher) to +6 (national brands received higher scores).

An independent sample t-test is used to test whether the scores for NB’s significantly differ from those of PL’s. So do the scores of brand awareness (BA), brand uniqueness (BU) and social image (SI) significantly differ for NB’s and PL’s? For brand awareness, brand uniqueness and social image the Levene’s test is significant at the 0.05 level (p<0.05). So, equality of variances is not assumed. These findings show that they are all significant (p<0.00). Therefore, it can be concluded that most consumers perceive a difference in the brand image from NB’s and PL’s. The statistic shows that NB’s score on average higher on brand image compared to PL’s (table 4.)

Mean t-value Sign. Mean difference

NB PL Brand Awareness 5.46 3.76 19.185 0.00 1.70 Brand Uniqueness 4.49 2.34 26.782 0.00 2.15 Social Image 2.95 1.77 15.082 0.00 1.17

Table 4. T-test statistics

(32)

Product category

Mean t-value Sign. Mean

difference NB PL Brand Awareness Chips 4.90 3.95 5.042 0.00 0.95 Chocolate 5.07 3.99 6.092 0.00 1.08 Spaghetti 5.99 3.61 14.769 0.00 2.38 Cola 5.88 3.52 14.700 0.00 2.36 Brand Uniqueness Chips 4.42 2.56 12.200 0.00 1.86 Chocolate 5.15 2.36 19.936 0.00 2.79 Spaghetti 3.06 2.23 5.739 0.00 0.83 Cola 5.33 2.23 20.819 0.00 3.10 Social Image Chips 3.17 1.89 8.275 0.00 1.28 Chocolate 2.98 1.82 7.531 0.00 1.16 Spaghetti 2.14 1.66 3.870 0.00 0.48 Cola 3.50 1.71 10.433 0.00 1.79

Table 5. T-test statistics per product category

4.4 Chow test

In order to test whether all data from the four different product categories can be pooled or not, a Chow-test is used (Leeflang et al., 2015). When the hypothesis of parameter homogeneity is not rejected, pooling the data across observations is statistically justified (Leeflang et al., 2015). The null hypothesis is defined as parameter homogeneity. The Chow-test is an F-test with degrees of freedom v1 and v2 and the following formula is used (Leeflang et al., 2015).

𝐹𝑣1, 𝑣2 ~(𝑅𝑆𝑆1 − 𝑅𝑆𝑆2)/𝑣1 𝑅𝑆𝑆2/𝑣2 where

RSS1 = Residual Sum of Squares of the pooled regression RSS2 = Residual Sum of Squares of the unpooled regression

(33)

The table with the residual sum of squares for the unit-by-unit, pooled and partially-pooled models can be found in table 6. The degrees of freedom for the pooled regression can be calculated by extracting the total number of parameters from the total number of observations (Leeflang et al., 2015). In this study there are 4 cross sections (n) (product categories), each with 192 observations (T), which makes the total number of observations n×T. For the fully pooled model with thirty-five explanatory variables (K, including dummy variables) and one constant the total amount of parameters is K+1. Therefore, the degrees of freedom for the fully pooled model is nT – K – 1. The total degrees of freedom for the unit-by-unit regression can be calculated by 𝑣2= n(T-K-1). Consequently, the difference in the degrees of freedom of the pooled regression and the unit-by-unit regression can be calculated by 𝑣1= (n-1)(K+1).

Model Residual Sum of Squares

Pooled Model 704490.44 Unit-by-Unit total 488304.55 Chips 148017.47 Chocolate 105720.37 Spaghetti 61288.73 Cola 173278.22 OLDSV-Model 669525.531

Table 6. Residual Sum of Squares

𝐹108, 624 ~(704490.44 − 488304.55)/108

488304.55/624 = 2.56

(34)

partially pooled regression and the unit-by-unit regression can be calculated by 𝑣1= (n-1)(K). This leads to the following formula;

𝐹117, 624 ~(669625.531 − 488304.55)/117

488304.55/624 = 1.98

F(117,624) = 1.98 which is significant a the α<0,01 level, so the null hypothesis of parameter homogeneity is rejected. The parameters of the OLSDV-model and unit-by-unit model significantly differ, so it is not allowed to partially pool the model. Therefore, further analysis of the data will be done using a unit-by-unit model in order to estimate a separate model for each product category. This results in a separate set of parameters for each of the four product categories which provides maximum flexibility in accommodating all potential differences between the different product categories (Leeflang et al., 2015).

4.5 Latent Class Modelling

Since the current data shows a lot of heterogeneity this section will present the best fitting latent class model based on estimation and comparison among different models. Due to time constraints the naïve model will be used, all explanatory variables as identified in the literature and control variables as covariates are included. Effect coding is used for the covariates; this implies that the parameters will sum to zero over the different categories of the nominal covariate. Therefore, the category-specific effects should be interpreted in terms of deviation from the average (Vermunt & Magidson, 2013). Eventually, the amount of appropriate latent classes will be chosen based on fit information criteria BIC, CAIC, AWE, the classification error and the R2. It is important to determine the appropriate number of different classes because specifying too few classes ignores differences between them while specifying too many might lead to an unstable model (Vermunt & Magidson, 2007). Due to time and space constraints in the current study, the product category of chips will be described thoroughly while the others are discussed briefly. This section will end with a comparison between the four product categories.

(35)

The AIC and AIC3 are often too small, they tend to favor more complex models and are better suitable for small sample sizes. Therefore, the focus will be on the BIC and CAIC. Additionally, the AWE is taken into account, this criterion is similar to the BIC but this also takes the classification performance into account (Vermunt & Magidson, 2007). Besides that, it is important to look at the classification error, this information criteria value indicates the proportion of cases that are estimated to be misclassified, as a result the closer this value is to zero the better (Vermunt & Magidson, 2007). However, besides the information criteria’s, the ideal number of segments should also be based on the interpretability of segments. This implies that segments should be homogeneous within the segment and heterogeneous between the different segments. As a result, segments have an appropriate size so marketing actions targeted on the particular segment will be profitable. Besides that, segments should be measurable and reachable in order to be able to communicate with the segment.

Model LL BIC(LL) CAIC(LL) Npar Class.E

rr. AWE _R2 1-Class 2-Class 3-Class 4-Class 5-Class 6-Class 7-Class -926,7350 -805,6154 -735,1876 -629,1474 -597,1982 -548,6833 -523,5458 1963,7368 1863,2694 1864,1857 1793,8771 1871,7504 1916,4923 2007,9891 1984,7368 1911,2694 1939,1857 1895,8771 2000,7504 2072,4923 2190,9891 20 46 72 98 124 150 176 0,0000 0,0336 0,0481 0,0286 0,0564 0,0526 0,0516 2137,0038 2291,4568 2531,1425 2660,7942 2983,1382 3252,2869 3564,4490 0,2229 0,5396 0,7442 0,8175 0,9822 0,9619 0,9948 Table 7. Information Criteria Chips

(36)

segments show homogeneous respondent groups whereas in the 2 and 3 segment too much variance is lost. Based on this, the four-class segment seems the most appropriate segment solution.

Figure 4: Class size

4.6 Main Effects

In order to test the main effects a regression is run using the program Latent Gold, this program enables to run a regression analysis while at the same time specify different numbers of classes. The results of the regression are shown in table 8. The beta parameter for each predictor variable is a measure of the influence of that predictor on the willingness to pay a price premium for national brands. However, it is important to note that the main effects as described here are conditional effects because an interaction model is used. The beta-coefficients that are shown are the simple effects when the interacting variable is zero (Pickery, 2008). The Wald-statistic test the restriction that every estimate in the sample of beta parameters is associated with that indicator equals zero (Vermunt & Magidson, 2005). So, the Wald statistic tests the null hypothesis of equality between segments, so it determines whether an indicator discriminates between the different segments (Vermunt & Magidson, 2005). Wald statistics that can not reject the null-hypothesis of equality are restricted by imposing zero restrictions on the different betas associated with that variable. As a result, they are assumed to be equal across segments in order to obtain a more parsimonious model (Vermunt & Magidson, 2005).

It was hypothesized that brand image will positively influence the WTPPP for NB’s. When looking at the different brand image element, the following can be concluded. First of all, it was expected

29%

26% 24%

21%

CLASS SIZE (N=192)

(37)

(38)

Mean Standardized Coefficients Beta (β) Wald(=)

Constant 15.858 45.231 -2.278 0.403 16.367 N/A

Main effects Segment1 Segment2 Segment3 Segment4 Brand Awareness Gap 4.033 5.296** 7.475** 3.827** -1.805** 77.412** Brand Uniqueness Gap 5.524 4.824** 12.217** 2.521** 1.625** 215.224** Social Image Gap -2.983 -3.636** -9.168** 1.319* 0.613** 92.010** Objective Quality 10.637 7.028** 3.133** 2.550** 33.665** 92.200** Social Context of Consumption -2.821 -15.441** 1.291** -13.532** 73.321** 59.610** Hedonic product value 0.616 -4.945*** 1.168*** 8.676*** -1.940*** 90.601*** Moderation effects Quality x Hedonic product value -4.774 -8.781 -6.74 0.451 -3.750 11.876 Quality x Social setting 11.967 9.904*** -13.400*** 28.256*** 27.455*** 30.831*** Brand Awareness x Hedonic product value 0.250 0.036** 0.086** -0.572** 1.729** 3.529 Brand Awareness x Social setting 2.163 5.333*** 7.160*** -9.859*** 5.735*** 6.779* Brand Uniqueness x Hedonic product value 0.152 -0.036 0.169 0.913 -0.512 1.877 Brand Uniqueness x Social setting -1.630 3.993*** -11.342*** 18.865*** -21.472*** 128.546*** Social image x Hedonic product value -0.474 -3.839*** -1.056*** 0.535*** 3.717*** 123.222*** Social image x Social Setting 4.609 -1.315*** 21.097*** -5.586*** 4.125*** 85.337*** Covariates

(39)

1= 18-29 2= 30-54 3= 55 and older Gender 0= Men 1= Women N/A -2.060* 0.705* 1.296* 0.060* N/A Income 1= < €1.000 2= €1.001-€2.000 3= €2.001-€3.000 4= €3.001-€4.000 5= €4.001-€5.000 6= >€5.000 N/A 2.085** 0.393** 0.932** -3.411** N/A Household size 1= 1 person 2= 2 persons 3= 3 or more N/A -0.821** 1.121** -1.420** 1.102** N/A Education 1= Primary school 2= LBO/VMBO 3= HAVO 4= VWO 5= MBO 6= HBO 7= WO N/A 0.356* 0.991* -1.288* -0.059* N/A

Table 8. Overall model estimation chips

(40)

4.7 Moderation effects

There is an interaction effect if the effect of one variable differs depending on the level of another variable (figure 5). So the effect of a dependent variable (X1) on WTPPP is dependent on the level of another variable (X2).

Figure 5: Moderation effect

In this study it is hypothesized that hedonic product value and social setting of consumption play a moderating role. Therefore, eight interaction terms are included in the model (table 8.).

First of all, it was expected that hedonic product value positively moderates the relationship between objective quality and brand image with the WTPPP for NB’s. When looking at the average scores of hedonic and utilitarian product value (figure 6.), it becomes clear that chocolate receives the highest hedonic value while spaghetti receives the lowest hedonic value.

(41)

When looking at the interaction effect of hedonic product value at each of the three brand image components and objective quality, the following can be concluded. The moderating effect of hedonic product value on the relationship on brand awareness and WTPPP is significant (p<0.05) and indicates a positive interaction effect in segment 1 (β=0.036), 2 (β =0.086) and 4 (β=1.729), whereas it shows a negative interaction in segment 3 (β= -0.572). However, the z-value in segment 3 indicates an insignificant relationship in this segment. This implies for segment 1 that the effect of brand awareness on WTPPP is 0.036 (β) procent-point higher for each 1-point increase in hedonic product value. Therefore, hypothesis 3b can be partly accepted for the product category of chips. The moderating effect of brand uniqueness and hedonic product value does not indicate a significant relationship (p<0.05), therefore hypothesis 3c can be rejected for the chips product category. Furthermore, the interaction effect between social image and hedonic product value is significant (p<0.001) and shows a positive interaction effect in segment 3 (β=0.535) and 4 (β=3.717). Therefore, hypothesis 3d can be partly accepted. The interaction effect between hedonic product value and objective quality is not significant (p>0.05), therefore hypothesis 3a can be rejected.

(42)

and indicates a positive interaction in segment 1 (β=9.904), 3 (β=28.256) and 4 (β=27.455). Hypothesis 5a can be partly accepted. In table (9) and overview of the hypothesis is provided.

Hypothesis Segment

1 2 3 4

H1: Objective quality is positively related to the WTPPP for NB’s + + + - H2a: The perceived brand awareness gap between PL’s and NB’s

will positively influence the WTPPP for NB’s over PL’s

+ + + -

H2b: The perceived brand uniqueness gap will have a positive impact on the WTPPP for NB’s over PL’s.

+ + + +

H2c: The perceived social brand image gap will have positive impact on the WTPPP for NB’s over PL’s.

- - + +

H3a: The level of hedonic product value positively moderates the relationship between objective quality and the WTPPP for NB’s.

- - + -

H3b: The level of hedonic product value positively moderates the relationship between the perceived brand awareness gap and the WTPPP for NB’s.

+ + - +

H3c: The level of hedonic product value positively moderates the relationship between the perceived brand uniqueness gap and the WTPPP for NB’s.

- + + -

H3d: The level of hedonic product value positively moderate the relationship between the perceived social image gap and the WTPPP for NB’s.

- - + +

H4: The social context of consumption will positively influence the WTPPP for NB’s.

- + - +

H5a: The social context of consumption will positively moderate the relationship between objective quality and the WTPPP for NB’s.

+ - + +

H5b: The social context of consumption will positively moderate the relationship between the perceived brand awareness gap and the WTPPP for NB’s.

+ + - +

H5c: The social context of consumption will positively moderate the relationship between the perceived brand uniqueness gap and the WTPPP for NB’s.

+ - + -

H5d: The social context of consumption will positively moderate the relationship between the perceived social image gap and the WTPPP for NB’s.

- + - +

(43)

4.8 Segments description

4.8.1. Segment descriptives

After segmentation and information in difference of WTPPP for national brands, it is interesting to see what are the main differences and what could explain these differences. In order to do so, the covariates and predictors are taken into account. The segments sizes can be found in figure 7. As can be seen, the segment sizes are equally distributed.

Figure 7. Segment sizes

First of all, it becomes clear that the differences in amount of the WTPPP differs clearly between the different segments (figure 8.). Segment 3 stands out with an average WTPPP of 59%, so respondents in this segment are willing to pay more than half of the price of a private label for a national brand. Figure 8. Average WTP 0% 5% 10% 15% 20% 25% 30% 35%

Segment1 Segment2 Segment3 Segment4

Class size (n=192)

0 10 20 30 40 50 60 70

(44)

When looking at gender (figure 9.), it is clear that segment 2, 3 and 4 show similar results with clearly more women than men. Only in segment 1, there is an equal distribution between men and women.

Figure 9. Gender distribution

When looking at age (figure 10.), it is observable that segment 3 is the segment with the lowest average and the youngest people (18-34). In contrast, segment 4 contains the older people (≥55). Segments 1 and 2 show homogeneous results with both most middle-aged people (35-54). It is interesting that younger customers (18-24) are on average willing to pay the highest price premium for national brands.

Figure 10. Age Distribution

When looking at income (figure 11.), it becomes clear that segment 1 has the highest average income (>€4000) while segment 4 has the lowest average income (€0-€2000) (figure 11.). It is

0% 20% 40% 60% 80% 100% 1 2 3 4

Gender

Men Women 0% 20% 40% 60% 80% 100%

Class 1 Class 2 Class 3 Class 4

Age

(45)

interesting that people with the lowest average income are also willing to pay the lowest average price premium for NB’s.

4.8.2. Segment specific description

What truly stands out is that each segment has their own variable that is considerably higher compared to those of the other three segments. In table (10) the differences are shown.

Segment 1 – A little bit of everything guys

This group contains the most men of the four segments, with the highest income (>€4000) and average age (35-54). This group is mostly motivated by brand awareness, brand uniqueness and objective quality. Interestingly, also the moderation effects these variables with social image are all positively related to the WTPPP. This implies that they are motivated to pay a price premium by brand awareness, brand uniqueness and objective quality when in a private setting, however when consumed in a social setting this effect becomes even stronger. So people in this group are willing to pay a higher price premium when the product is consumed in a social setting.

Segment 2 – The unique middle

This is the group with most people with an average income and average age and contains the most women. This group allocates a great importance to the uniqueness of the product and they are willing to pay 12.21% more based on this. Interestingly, this does not show in the moderation effects, since the interaction with social setting is negative while the interaction with hedonic

0% 20% 40% 60% 80%

Class 1 Class 2 Class 3 Class 4

Income

Low (€0-‐€2000) Average (€2001-‐€4000) High (>€4000)

(46)

product value is really small. So they are motivated by uniqueness of the products irrespectively of the social setting or the hedonic product value. Quality is the second most important factor that influence them to pay more for a national brand. When looking at the moderation effects, it becomes clear that the interaction between social image and social setting motivates them to pay 21.09 procent-point more for national brands. So these people care about the social image that the product communicates to others while in a social setting.

Segment 3 – Social image youngsters

This is the youngest group with the highest average WTPPP for NB’s. All brand image elements show a significant positive relationship with WTPPP, although these effects are relatively low. The interaction effect between quality and social setting is very high (β=28.526), while the main effects of social setting and quality are either negative or really low. This implies that people in this segment are mainly willing to pay more for national brands based on the quality in combination with a social setting of consumption. Also the interaction between uniqueness and social setting of consumption is high (β=18.865). This implies that customers in this segment are mainly willing to pay a price premium based on uniqueness when they know that they are going to consume the product in a social setting.

Segment 4 – The quality oriented older generation

(47)

Segment Information Segment 1 Segment 2 Segment 3 Segment 4 Average WTPPP 32% 40% 59% 24% Descriptives Age 35-54 35-54 18-34 ≥55 Income Average-High

Average Low-Average Low

Education

Average-High

Average High Low

Household size 3 3 1 2-3

Gender Men/Women Women Women Women

Main effects Quality ++ + + ++ Brand Awareness ++ ++ + - Brand Uniqueness + ++ + + Social Image - -- + -/+ Social setting -- - -- ++ Moderation effects Quality x Hedonic product value -- -- +/- -

Quality x Social setting ++ -- ++ ++

Brand Awareness x Hedonic product value

+ +/- +/- -

Brand Awareness x Social setting

+ ++ -- +

Brand Uniqueness x Hedonic product value

+/- +/- + +/-

Brand Uniqueness x Social setting

+ -- ++ --

Social image x Hedonic product value

- - +/- +

Social image x Social Setting

- ++ -- +

Table 10. Segments description

IMAGE OR QUALITY?

IMAGE OR QUALITY?

The effect of brand image and objective quality on the

willingness to pay a price premium for national brands over

private labels and the moderating effects of consumption

situation and hedonic product value

Master thesis, Marketing Intelligence

University of Groningen, Faculty of Economics and Business

January 2017

MARIEKE SCHULENBURG

Student number: 2034042

Van Brakelplein 41a

9726 HG Groningen

Tel: +31(0)651013528

MANAGEMENT SUMMARY

PREFACE

TABLE OF CONTENTS

1. INTRODUCTION

2. THEORETICAL BACKGROUND

BRAND EQUITY

BRAND IMAGE

BRAND STRENGTH

BRAND VALUE

Objective Quality

Differential

Brand Image

Willingness to pay a

price premium

Hedonic product value

3. RESEARCH DESIGN

4. RESULTS

CLASS SIZE (N=192)

Class size (n=192)

Gender

Age

Income