Research into the consumer perception of assortment
and consumer purchase intention
M
ANAGEMENT SUMMARY
Due to the increasing number of hard discounters (HDs), competition increases the need for differen-tiation within the format (Deleersnyder and Koll, 2012). HDs are known to almost exclusively offer their own private label (PL) products (Ailawadi and Keller, 2004). However, HDs are now increasingly introducing national brands (NBs) into their assortment (Lourenço and Gijsbrechts, 2013). NBs can help generating consumer interest, patronage and loyalty (Ailawadi and Keller, 2004), but they can also increase the dependence on manufacturers and lead to product cannibalization (Deleersnyder and Koll, 2012). From a manufacturer perspective, the primary reason to sell a national brand at HDs is to expand the market. Moreover, manufacturers do not believe that selling at HDs will diminish their brand image. This present study aimed to indicate whether purchase intention increases if HDs introduce an extra brand, and if so, should they choose to introduce more NBs or a premium fancy label (PFL). The focus of this research is on NBs. In addition, to fully understand how introducing NBs at HDs may influence purchase intention of costumers, the present study took into account the consumer perspective. Assortment perceptions such as price, quality and variety can influence shopping habits in supermar- kets (Mantrala et al., 2009). This study therefore investigated whether assortment perceptions medi-ate the relationship between addition of an extra brand and purchase intention. Moreover, since introducing new brands also introduces differences in price level, this study additionally measured the effect of price level on the assortment perceptions and purchase intention. And lastly, to under-stand if the effect of price level and brand on assortment perception and purchase intention would depend on the hedonic level of a product, this variable was introduced as a moderator for each rela-tionship. The central question of this study is as follows: What is the impact of adding national
purchase intention. The next question was whether HDs should add PFLs or more NBs. The results of the MANCOVA analysis showed that the addition of an NB increased quality and variety perception compared to PL and PFL. The addition of an NB did not even damage the price perception, because an NB price perception was similar to the assortment with only PLs. Furthermore, the ANCOVA anal-ysis show that the addition of an NB increased purchase intention. The mediation model of Hayes showed that only quality perception and variety perception fully mediated the relationship between the group (PL, PFL, NB) and purchase intention. This study has shown that adding a PFL would not improve assortment perception or purchase intention at HDs. Surprisingly, this study did not find support for the notion that the price level of brands at HDs impacts the assortment perceptions and purchase intention. Additionally, this study did not find evidence for the moderator ‘hedonic level of a product’. A hedonic product did not influence the relationships between the group or price level and the assortment perceptions and purchase intentions. The study found the opposite, the utilitari-an product did positively influences the relationship between group and variety perception. The findings of this research provide insights for HDs who are considering to expand their assortment with NBs. This study strengthens the idea that HDs should add NBs to their assortment and comple-ment findings of earlier studies (Lourenço and Gijsbrechts (2013); Deleersnyder and Koll. 2012), by taking into account the consumers’ perspective on different price levels of brands and different brand types (PFL and NB).
Keywords: Hard discounters, national brands, private labels, price perception, quality perception, variety perception, purchase intention, hedonic level.
P
REFACE
After graduating at the Hanzehogeschool Groningen for my Bachelor of Commerce, I decided to con- tinue with studying. I started a pre-master in Marketing at the University of Groningen, which I fin-ished in one year. This year I hope to finish my master degree in Marketing Management. To be able to accomplish my master degree, I wrote a final master thesis. In front of you lies the result of my master thesis. First of all, I would like to thank my first supervisor Prof. Dr. Laurens Sloot for his feedback and ad-vice. Due to our conversations, I have learned a lot during my master thesis period. Secondly, my sincere thanks go to my family and friends for their support and their encouraging words. Thirdly, I would like to thank Dr. Wander Jager for his evaluation. Fourthly, I appreciated the helpful com- ments of my fellow students. Finally, I would like to thank all the respondents who filled in my sur-vey.
Miriam Jansen Groningen, June 2017
T
ABLE OF CONTENTS
MANAGEMENT SUMMARY 3
PREFACE 5
1.INTRODUCTION 8
1.1 BACKGROUND 8
1.2 NATIONAL BRANDS AT HARD DISCOUNTERS 8
1.3 RESEARCH QUESTION 9
1.4 ACADEMIC AND MANAGERIAL CONTRIBUTION 10
1.5 THE STRUCTURE 10
2. THEORETICAL BACKGROUND 11
2.1 HARD DISCOUNTERS 11
2.2 LISTING NATIONAL BRANDS AT HARD DISCOUNTERS 13
2.2.1 HARD DISCOUNTER PERSPECTIVE 13
2.2.2 MANUFACTURER PERSPECTIVE – EXPERT OPINIONS LISTING NB AT HD 15
1.I
NTRODUCTION
1.1 BACKGROUND
The retail industry has changed a lot over the last few years. In the Netherlands, traditional super-markets such as Albert Heijn and Jumbo still have a larger market share than hard discounters such as Aldi and Lidl (Berkhout, 2017). However, hard discounters (HDs) have significantly altered the market share of existing retail formats (Van Heerde, Gijsbrechts and Pauwels, 2008). Albert Heijn has the largest turnover and market share (35%) in the Netherlands. Jumbo comes second, with a 17% market share. Lidl and Aldi are in the third and fourth position (Berkhout, 2017). HDs distinguish themselves from traditional supermarkets by focusing on very competitive prices and strong reliance on their own private label (PL). Furthermore, HDs have an extremely efficient supply chain, mainly because of the limited numbers of SKU’s (Stock Keeping Units) per supermarket and their PL, which makes the operation simpler (Steenkamp & Kumar. 2009). To be able to offer lower prices, all HDs have a simplified, ‘no-frills’, store format. They have limited promotional activities, and only rarely introduce new products (Deleersnyder et al., 2007). HDs not only have a direct effect on the market share of traditional retailers, they also force other retail formats to increase operational efficiency and/or decrease prices (Van Heerde, Gijsbrechts and Pauwels, 2008). In response to the increasing number of HDs, traditional retailers were forced to develop defensive strategies and come up with an ‘economic private label’, which complements the standard line of PL products (Ailawadi, Pauwels, and Steenkamp, 2008).
1.2 NATIONAL BRANDS AT HARD DISCOUNTERS
Although HDs rely almost exclusively on their own store brands, HDs are currently increasingly intro- ducing national brands (NBs) into their assortment. Aldi and Lidl have now well-known NBs in differ-ent categories, e.g. Ola, Nutella, Coca Cola, Lay’s, and Calvé. Retailers use NBs to generate consumer interest, patronage and store loyalty (Ailawadi and Keller, 2004). Adding NBs to the assortment of-fers some benefits to discounters. Deleersnyder and Koll (2012) found that selling NBs at HDs in-creases the total performance of a discounter. This means that the benefits of selling NBs through a discount channel outweigh the potential cannibalization losses. In addition, listing NBs at HDs can provide ‘one-stop shopping’ service, which creates an obvious benefit for consumers (Semeijn and Vellenga, 1995), and enhance their assortment attractiveness (Deleersnyder and Koll, 2012), which will be the focus of the study presented here.question whether HDs should introduce more NBs to their assortment or premium fancy labels (PFLs). Another question concerns the type of products for which an HD can introduce new brands. For in-stance, a distinction can be made between hedonic and utilitarian products. Utilitarian products are characterized by practical benefits, whereas hedonic products have enjoyment-related benefits. Both aspects contribute to the overall goodness of the brand or product, but in different ways (Chitturi, Raghunathan & Mahajan, 2008). Sethuraman (2003) suggests that consumers pay a higher premium for NBs with more hedonic aspects, compared to NBs with more functional aspects. What does this mean for HDs? Does the purchase intention increase when introducing more brands in hedonic product rather than utilitarian product categories? Hence, this raises the following question: if HDs decide to introduce more brands in their assortment, should they choose products with more hedon-ic aspects or more functional aspects? Also, HDs must choose a price strategy for the brands. Should they be considerably cheaper than traditional retailers, or can they offer the brands at the same price as traditional retailers do?
While other studies evaluated consumers’ reaction to adding private labels in an NB dominated as-sortment (e.g. Pauwels and Srinivasan, 2004), the present study investigates the opposite, namely how consumers perceive the assortment by adding NBs in a PL dominated HD assortment. If HDs choose to offer more NBs (at prices that are above the HD’s PL offer) the HD’s value image is likely to change, because it jeopardizes the HD’s reputation of offering value-for-money (Lourenço and Gijsbrechts, 2013).). Understanding how consumers perceive the assortment and how this influences purchase intention is thus of critical importance. The focus of the present study is the addition of an NB. However, the study will also investigate the effects of adding a PFL (instead of an NB).
1.3 RESEARCH QUESTION
Previous research mainly focused on the benefits and risks of listing NBs at an HD from an HD and manufacturer’s perspective (Deleersnyder and Koll, 2012). The study presented here expands the literature from a different perspective, by also considering the consumer perspective of adding NBs at HDs. The following main research question is formulated:
What is the impact of adding national brands to a hard discounter’s assortment on assortment per-ceptions and purchase intention, and besides that, what is the impact of different price levels of brands on the assortment perceptions and purchase intention, and lastly, to what extent are these effects moderated by the hedonic level of a product?
To be able to answer this research question, the following sub questions should be answered: 1) What is meant by assortment perception?
3) What is the influence of adding NBs to an HD assortment if product type (hedonic vs utilitarian) is considered?
1.4 ACADEMIC AND MANAGERIAL CONTRIBUTION
This research extends previous research of Lourenço and Gijsbrechts (2013). These authors found that there should be a sufficiently deep assortment in order to enhance the HD’s assortment percep-tion. The authors suggest that future research should focus on whether the HD’s shoppers are more likely to appreciate (or dislike) increased NBs listings. The current study will honour this suggestion by further elaborating on this issue. This study will investigate whether purchase intention will in- crease by adding NBs to the assortment of an HD. This research will also extend the study of Deleer-snyder and Koll (2012), that examined the effects of NBs at HDs from the HDs’ and manufacturers perspective, by investigating the consumers’ perspective. If discounters carry more and more NBs, it would change assortment perception. Consumer percep-tion has been shown to be the most important drive of store choice and spending (Srivastava and Lurie, 2004), even more so for HD shoppers (Lourenço and Gijsbrechts, 2010). Hence, results from the present research will be valuable for managers of HDs who consider adding (more) brands into their assortment. The present study will help understand which brand type (NB or premium fancy label) and price level for which product type (hedonic or utilitarian) will be more beneficial when added to the assortment of an HD. Although there is a lot of knowledge about listing NBs in assort- ment from a discounters’ perspective, a customer-perspective is needed to fully understand the im-pact of adding brands to the assortment of discounters.
1.5 THE STRUCTURE
The structure of this paper is as follows. In chapter 2 we will discuss the theoretical background. Chapter 3 will illustrate the hypotheses and the framework of this study, which will provide the fun-damental basis for this research. The methods used in this study are described in chapter 4. This chapter is followed by the results of the study in chapter 5. In chapter 6, the hypotheses are rejected or accepted. The paper finishes with conclusions, a discussion, indications of the limitations and sug-gestions for future research in chapter 7.
2.
T
HEORETICAL BACKGROUND
In this chapter, an overview of relevant academic literature will be given. The chapter will commence with general information about HDs and their business strategy. In section 2.2 relevant literature about NBs at HDs is discussed from an HD’s perspective and also from a manufacturer’s perspective.
2.1 HARD DISCOUNTERS
Grocery merchandise industries have changed over the last decade. Discounters have experienced growth and are one of the fastest growing segments in grocery retailing (Lamey, 2012). During the economic crisis, which started in 2008, consumers became more aware of prices. Hence, low-price oriented supermarkets became more popular. That is why HDs could grow rapidly in Europe (Lin, Deleersnyder, Dekimpe, Geyskens, 2013). Table 1 shows the market shares of Aldi and Lidl in differ-ent countries in Europe. It shows that all countries realized a substantial growth in the last six years, except for Aldi in the Netherlands and Denmark. Aldi and Lidl are the two largest HDs in the Dutch market (Berkhout, 2017). HDs are different from every-day-low-price retailers (e.g. Jumbo) and from traditional supermarkets (e.g. Albert Heijn). The vision of an HD is to offer products at lower prices than competing traditional formats (Cleeren, et al., 2009). Discounters offer only a limited selection of items, typically fewer than 1.400 SKUs (Cleeren, et al., 2009). HDs economize on store loyalty and customer service to be able to offer low prices (Gijsbrechts et al., 2008). The assortment is dominat-ed by private labels with quality comparable to NBs (Cleeren, et al., 2009). HDs drive large volumes through each product line (Weinswig, 2015). Due the growth of HDs, competition increased between retail formats. Discounters have put pressure on traditional retailers to increase their operational efficiency and/or decrease prices (Van Heerde et al., 2008).
Table 1: Market share grocery retailers Aldi and Lidl per country (Euromonitor, 2016a to o)1
Market
share Lidl Aldi
2011 2016 2011 2016 Austria 3.9 4.7 15.0 17.5 Belgium 4.6 5.2 8.1 8.3 Denmark 1.8 2.2 2.8 2.6 Germany 15.0 16.2 3.9 13.7 Ireland 4.3 10.9 5.2 6.9 Netherlands 5.4 7.7 6.4 6.0 Slovenia 5.0 7.9 6.1 10.6 UK 1.8 3.2 1.8 5.2
1
Category definition: Retailers selling predominantly food/beverages/tobacco and other everyday groceries. This is the aggregation of
tional formats (Cleeren, et al., 2009). According to Porter (1980) this implies that HDs pursue a cost leadership strategy. According to Treacy and Wiersema (1993) there are three value disciplines: operational excellence, product leadership and customer intimacy. These values are based on knowledge of customer val- ues. To be a market leader, a firm must be a leader in one of the disciplines and meet industry stand-ards in the other two. Firms that pursue an intimacy strategy, tailor and shape products to fit the needs of the customers. Firms that continuously improve their products and constantly search for new solution are following a product leadership strategy. The term operational excellence describes a strategic approach with focus on production and delivery of products (Treacy and Wiersema, 1993; P.85). A company should lead its industry in price and convenience. Firms that pursue this strategy focus on delivering their products to customers at competitive prices. Due the focus on low prices of HDs (Treacy and Wiersema, 1993), Aldi and Lidl both pursue an operational excellence strategy.
2.2 LISTING NATIONAL BRANDS AT HARD DISCOUNTERS
This section will discuss the effect of adding national brands at discounters from a hard discounter’s and manufacturer’s perspective. 2.2.1 HARD DISCOUNTER PERSPECTIVEAs discussed earlier, HDs strongly rely on their own private label (PL) (Lourenco and Gijsbrechts, 2013). PL products are defined as products that are produced by, or on behalf of, retailers and sold under a trademark through an HD’s own outlet (Baltas, 1997). However, more and more HDs sell NBs (Lourenço, Gijsbrechts, 2013). NBs are sold under the brand of a manufacturer. These products are available at almost all retailers (Baltas, 1997). Retailers use manufacturer brand to generate consum-er interest, patronage, and loyalty to a store. Most of the time, NBs help to attract consumers, often more than the retailer brand does (Ailawadi and Keller, 2004). Furthermore, growing competition within the format increases the need for differentiation (Deleersnyder and Koll, 2012). Table 2 shows a small overview of NBs at HDs. A complete overview of NBs at HDs can be found in Appendix 1. The-se overviews show that NB assortments are dominated by a lot of brands of Unilever.
Table 2: List of national brands at hard discounters National brands Aldi Lidl Brauerei C & A Veltins X Coca cola X Dr. Oekter X Haribo X X Hela Thissen BV X Mars Inc. X Nestlé X Nutella X Perfetti van Melle -Mentos X Procter & Gamble: -Always -Dreft X X Unilever: -Blue Band -Calvé -Dove -Lipton -Unox X X X X X
Source: website Aldi & self-research in Lidl store (April 2017)
Deleersnyder and Koll (2012) show that listing NBs in the assortment offers some benefits. HDs im- prove their distribution coverage and enhance their assortment attractiveness. Pauwels and Sriniva-san (2004) found that when retailers have too much shelf space for PLs, consumers may feel that there is only a limited choice. Moreover, by listing NBs in the assortment, more profitable consum-ers, i.e., consumers who buy both NBs and PLs, will visit the stores (Deleersnyder and Koll, 2012). Ailawadi, Pauwels and Steenkamp (2008) found that PL buyers tend to be loyal to price savings and PLs in general. Shoppers are not loyal to any private label of a particular retailer. Therefore, private label shoppers are less loyal to a store. This implies that HDs can attract new consumers by adding NBs and increase store loyalty. Moreover, hard discounters are seldom the single store-of-choice for shoppers. Consumers may choose to buy particular products at traditional supermarkets and some other products at a discounter, which makes an HD complement the traditional supermarket rather than a substituting them (Vroegrijk, Gijsbrechts, and Campo, 2013). By listing NBs in their assort- ment, HDs will increase the probability that shoppers will buy all their products at HDs, and becom-ing a single store-of-choice for shoppers. However, there are also some risks. From the perspective of an HD, the addition of brands will blur the discount position, product cannibalization will arise, as-sortment will be more complex, and dependence on manufacturers will increase (Deleersnyder and Koll, 2012).
In a recent presentation about Aldi Sued on the 21st of March 2017, Sievers presented results about
and 2/3 of sales is due to their own PLs. This means that by listing NBs the effects of cannibalization are rather small. In 2016, the sales in listing categories decreased (in comparison with 2015), howev-er, sales were still 31% higher than in 2014. The sales ratio between NBs and PLs was about the same as in 2015. Furthermore, Sievers argues that discounters who are listing NBs in the assortment put pressure on traditional retailers to lower their prices, due to price erosion. For example, before Aldi started to list Redbull to its assortment, the usual shelf price for a Redbull was € 1.99. After Aldi de- cided to add Redbull in assortment, the price decreased to €1.29 at traditional retailers. This indi- cates that when Aldi introduces NBs in its assortment, the prices of NBs go down at traditional retail-ers. HDs offer unbeatable prices in comparison with traditional retailers. To understand the current price strategy of products of NBs at an HD, all prices of NBs were analysed at Aldi. Currently Aldi offers more NBs than Lidl, which makes it more interesting to analyse the price strategy of Aldi. In total, Aldi offers 54 products of NBs. These products are on average 11% cheaper than at Albert Heijn (market leader in The Netherlands). The analysis (see Appendix 2) shows that every NB is offered at a lower price at Aldi than at Albert Heijn. However, when the prices of NBs at Aldi are compared to Jumbo (second player), five products are more expensive at Aldi than at Jumbo, 9% of the NBs at Aldi are offered at a higher price than at Jumbo. Thus, it is concluded that not all NBs are cheaper at an HD, which means that Aldi uses varying pricing strategies for different NBs.
2.2.2 MANUFACTURER PERSPECTIVE – EXPERT OPINIONS LISTING NB AT HD
3.
C
ONCEPTUAL MODEL
&
HYPOTHESES
This chapter starts with explaining the conceptual model in more detail. In the next paragraphs, the variables of the model will be described. Each paragraph will start with a literature review about the variables, followed by a hypothesis.3.1 CONCEPTUAL MODEL
Figure 3 shows the conceptual model of this study. This study uses three different experimental vari-ables: The first experimental variable is the addition of a brand. To find out if HDs should add (more) brands in their assortment, the present study will make a distinction between NBs and PFLs. The PFL is an unknown brand which suggests high-quality. The second experimental variable is the price level of the NB/PFL. This present study uses different price levels to investigate if there are different responses to different price levels of NBs or premium fancy labels (PFLs). As discussed in section 2.2.1, the prices of NBs are compared between an HD (which already sell NBs in the assortment) and a traditional retailer. Therefore, two price levels will be used: 1) the average percentage that consumer pay less for NBs at an HD in comparison with the prices that consumers must pay at a traditional retailer (which is discussed in section 2.2.1) and 2) no difference between prices of NBs between an HD and a traditional retailer. The third experimental variable is a moderator, the hedonic level of products. This moderator is ex-pected to influence the effects of the first two experimental variables on the assortment perception and purchase intention. The assortment perceptions and purchase intention are the dependent variables. Because HDs add an NB or a PFL to their assortment, using different price levels, it is assumed that this will influence the assortment perception on three different dimensions: price perception, quality perception and variety perception. Subsequently, it is hypothesized that brand and price level will influence the se-cond dependent variable, purchase intention, due to the changes in assortment perceptions. It is also expected that the experimental variables have a direct effect on purchase intention due to heuristics cues. Moreover, it is expected that consumer characteristics can have an influence on the assort- ment perception and purchase intention. To account for these influences, the following characteris-tics will be included as control variables: gender, age, income level, education, family size, quality consciousness, price consciousness, brand consciousness, familiarity with store and buying behaviour brandstion of the actual price of the product. From this definition, the following definition of price percep-tion is given: Price perception of the actual price of the product. From this definition, the following definition of price percep-tion is a consumer’s perceptual representation of the actual price of the product. From this definition, the following definition of price percep-tion of the actual price of the
ers use well-known NBs to judge the quality of the assortment. This result is confirmed in other stud-ies (e.g. Richardson, Dick and Jain, 1994). NBs are perceived to be excellent in terms of appearance, taste and attractiveness whereas PLs are not. As consumers rely on cues like appearance and attrac-tive packaging when judging assortment quality, and NBs are therefore perceived as superior over PLs, it is likely that by listing NBs the perceived quality of the assortment will increase. • H2a: Compared to the situation where hard discounters only carry PLs, adding NBs to the as-sortment positively influences the perceived assortment quality Price can be an indicator of the level of quality (Dodds et al., 1991). This price-quality relationship means that ‘consumers use price as an informational cue to develop beliefs about the product’s qual-ity’ (Bauer, Kotouc & Rudolph, 2012 p.13). Higher prices lead to higher perceived quality (Dodds et al., 1991). Therefore, it is expected that by adding NBs to the assortment at lower prices than traditional retailers, the quality perception of the assortment will decrease. •
H2b: Compared to the situation where hard discounters use the same prices as traditional re-tailers, using prices that are considerably lower negatively influences the perceived assort-ment quality
3.2.3 PERCEPTION OF ASSORTMENT VARIETY
To meet the needs of changing circumstances or to comply with an internal drive to choose variety over time, consumers seek diversity in their choices of goods (Kahn, Kalwani, and Morrison 1986; Kahn, 1998). Kahn and Wansink (2004) found that consumers do not only evaluate the assortment variety by the number of SKUs offered in a category, but they also use other determinants to form a perception of the variety of the assortment. Kahn and Wansink (2004) stated that variety seeking consumers perceive greater utility when retailers increases their choice in brands, flavours, packag-ing and sizes of their assortments. Bauer, Kotouc & Rudolph (2012) found one more determinant that increases the perception of the variety of the assortment, namely quality ranges. By listing NBs in the assortment, the perceived variety of brands and quality ranges will increase, and therefore consumers have more options for choosing a product, which, in general, will increase the likelihood of consumers finding their desired items (Ailawadi and Keller, 2004). Furthermore, by adding new products and changing the assortment, consumers believe that they have more flexibility in their choices (Kahn and Lehmann, 1991). Therefore, the following hypothesis is formulated:
•
H3a: Compared to the situation where hard discounters only carry PLs, adding NBs to the as-sortment positively influences the perceived assortment variety
As discussed before, variety seeking consumers are looking for different brands, flavours, sizes and quality (Bauer, Kotouc & Rudolph, 2012), but not a variety in price levels. Therefore, it is expected that the price level of NBs has no influence on the perception of assortment variety. • H3b: Compared to the situation where hard discounters use the same prices as traditional re- tailers, using prices that are considerably lower do not influence the perceived assortment va-riety
3.3 PURCHASE INTENTION
Purchase intention can be defined as the consumers’ intention to purchase a product. Purchase in-tention is an effective tool to forecast sales. In the literature, consumer intention is often used as a predictor of subsequent purchase (Grewal et al., 1998). This is confirmed by Luo et al. (2011) who showed that purchase intentions predict actual purchase behaviour. Zeithaml (1988) shows that purchase intention might be altered by the price and quality perception. As stated above, by adding NBs to the assortment, the assortment quality perception is expected to change. As a result, NBs are likely to evoke superior perception (Richardson et al., 1994) and higher purchase intention than private labels (Das, 2014). As discussed, it is expected that adding NBs in the assortment negatively influences price perception. Hence, it is assumed that consumers will form a negative attitude towards prices in categories with NBs. In turn, this may have a negative impact on the purchase intention of existing customers. However, Deleersnyder and Koll (2012) showed that adding NBs to the assortment of HDs will attract more new customers, which will increase the pur-chase intention at HDs. Furthermore, it is assumed that adding NBs at HDs will have a positive impact on the variety perception of the assortment. Hoch et al. (1999) show that there is a positive relation-ship between assortment variety and fulfilment of consumers’ needs. This means that consumers are more likely to find the product they were looking for (Lancaster, 1990). This will likely influence the purchases intention in a positive way. However, researchers also found that there can be too much choice in an assortment; an assortment with too many choices can be overwhelming and render it more difficult to choose as it enlarges the chances of regret (Berger, Draganska & Simonson, 2007; Iyengar and Lepper 2000). It is unlikely however, due to the low SKUs at HDs compared to traditional supermarkets, that adding NBs to the small HD assortment will lead to a too wide assortment variety.Therefore, the following hypotheses are formulated. • H4a: Compared to the situation where hard discounters only carry PLs, adding NBs to the as-sortment positively influences purchase intention • H5a: Compared to the situation where hard discounters only carry PLs, adding NBs to the as- sortment positively influences purchase intention via quality perception and variety percep-tion, but also negatively influences purchase intention via price perception
In general, a price reduction increases demand (Cordell, Wongtada, Kieschnick, 1996), which subse-quently influences purchase intention. Therefore, it is expected that purchase intention will increase by using prices which are considerably cheaper than at traditional retailers. As stated above, it is assumed that the price reduction of NBs at HDs will have a positive effect on price perception. As a result, it is expected that the positive price perception will subsequently increase purchase intention. Additionally, it is expected that a price reduction has a negative impact on the quality perception of the assortment. As consumers use price as an indicator of quality, lowering prices will have a nega-tive effect on the quality perception of the assortment. This will subsequently negatively influence the purchase intention. Lastly, it is assumed that a price reduction will not influence the perception of assortment variety, which means that this will also not influence the purchase intention. There-fore, the following hypothesis are formulated. • H4b: Compared to the situation where hard discounters use the same prices as traditional re-tailers, using prices that are considerably cheaper positively influences purchase intention
•
H5b: Compared to the situation where hard discounters use the same prices as traditional re-tailers, using prices that are considerably lower positively influences purchase intention via price perception, but negatively influences purchase intention via quality perception3.4 HEDONIC & UTILITARIAN PRODUCTS
The distinction between utilitarian and hedonic products can be included when testing the effects of adding NBs at HDs. It is assumed that when consumers shop in a category with a high level of hedon-ic products, they are more interested in the emotional aspect of the NB. For utilitarian products, the brand name will be less important to consumers, since they mainly look at the instrumental value of the product (Babin, Darden & Griffin, 1994). This is confirmed by Fischer, Völckner & Sattler (2010), who show that in categories with higher brand relevance, consumers are looking for brand benefits and that the brand name plays a critical role in purchase decisions. Based on these results it is as-sumed that by listing NBs to the assortment, hedonic rather than utilitarian products will positively affect consumers’ assortment perception and purchase intention. • H6a: The effects of listing NBs at hard discounters on assortment perceptions and purchase intention are stronger for hedonic than for utilitarian products
Consumers may attribute high performance risk to PLs especially in hedonistic categories because consumers are worried that PLs are not able to deliver the preferred emotional benefits (Sethuraman and Cole, 1999). Consequently, it is expected that consumers are willing to pay a larger premium for NBs which have a hedonic dimension.
• H6b: The effects of using prices that are considerably lower than traditional retailers are
stronger for utilitarian than for hedonic products
3.5 CONTROL VARIABLES
4.
M
ETHODOLOGY
The objective of this study is to create insight in and understanding of the effect of adding NBs to the assortment of HDs on purchase intention and consumers’ assortment perceptions. An online ques- tionnaire was conducted. In this chapter, data collection and the design of the experiments are de-scribed. The chapter continues with a description of the presentation of product categories and the measurement of the constructs.
4.1 DATA COLLECTION AND RESEARCH DESIGN
This study used an experimental design to investigate the effect of adding NBs at HDs. In an experi-mental design, the independent variables are systematically manipulated and their effect on one or more dependent variables is subsequently measured, while controlling for extraneous variables (Malhotra, 2009). The present research had two independent variables (addition of a brand and price level), four dependent variables (assortment perceptions [price perception, quality perception, varie-ty perception] and purchase intention), one moderator (hedonic level), and several control variables (gender, age, household size, income, education level, familiarity with Aldi, buying behaviour of brands, brand equity, and quality-, price-, and brand consciousness). The independent variable ‘addi- tion of a brand’ was manipulated by presenting participants with an assortment of only PLs, an as-sortment of PLs and an NB or an assortment of PLs and a PFL.was presented with one product type, either hedonic or utilitarian, which was evenly divided over the three groups. Lastly, the participant in experiment group 2 and 3 could either see a price of a NB/PFL which was considerably lower than traditional retailer, or the same price as traditional retail-ers. The experimental design is shown in Table 3. To collect the primary data, a survey was used that could be filled in online. The costs of this method are relatively low and the obtained data can be processed by using statistical software, such as SPSS (Malhotra, 2009). This study focused on one HD, namely Aldi. Both HDs (Lidl and Aldi) are currently listing NBs, so consumers are familiar with NBs in the assortment of HDs. As discussed in section 2.1, Aldi has a smaller market share than Lidl in the Netherlands, while Aldi used to be the stronger chain. For Aldi, it is thus valuable to get insight in whether more NBs or PFLs should be introduced into its assortment.
4.2 HEDONIC AND UTILITARIAN PRODUCT CATEGORIES
As discussed before, hedonic and utilitarian products for consumption need not be mutually exclu-sive (Batra and Ahtola, 1990). For this reason, it is necessary to establish whether certain product categories are truly perceived as hedonic or utilitarian. The study of Sloot, Verhoef and Franses (2005) was used to understand which products are perceived as more hedonic and which as more utilitarian. The two categories that were chosen are shown in Table 4. Table 4: Product groups Product type Product group Hedonic Chips Utilitarian Milk Currently, Aldi generally offers NBs that are comparable to their own PLs. For example, ‘Cornetto’ ice cream is comparable to Aldi’s product ‘Cornets’, ‘Coca Cola Zero’ is comparable to ‘River cola 0%’, and ‘Nutella’ is comparable to ‘chocolate paste’. It can be concluded that Aldi introduces mostly NBs Table 3: Experimental designAssortment type Product type Price level
Control group 1 4 Private labels
(4 facings per private label) Hedonic or utilitarian product cate-gories
of 1-3 cents, to have the same pricing strategy as Aldi (ending prices with 9 or 5 cents). These prices apply to both experimental groups 2 and 3. By not changing the prices of NBs between the experi- mental groups, it was possible to see if there were different results between PFLs and NBs. Moreo-ver, to reduce the risk of bias, it was important to take into account different ratios between and within product categories. The prices should have the same ratio among product categories to be able to compare the different groups. Table 5 presents the different ratios between product catego-ries. Table 5: Price ratio per product category with 11% discount (product names are in Dutch) Product type PL1 PL2 PL3 PL4 % 1 NB1 NB2 NB3 NB4 % 2 %3 Hedonic
Natur-el Papr-ika Bolog-nese Cheese & onion Lays Natur-el
Lays
Paprika Lays Bolognese Lays Cheese & onion €0.79 €0.79 €0.89 €0.89 13% €1.15 €1.15 €1.19 €1.19 4% 51% Utilitarian Half volle melk (HVM) Volle melk (VM) Houd-bare HVL
Houd-bare VM Cam-pina HVL
4.4.2. DEPENDENT VARIABLES
The construct perceived assortment variety was measured by the components of Chowdhury, Rear-don and Srivastava (1998) and Huddleston et al (2009). The following items were included in the questionnaire: ‘This assortment has a large variety of choice’, ‘In this assortment, Aldi offers the as- sortment from which I could always make the right decision’ and ‘This assortment has a good prod-uct selection’.
To measure the perceived product quality, the studies of Chowdhury, Reardon and Srivastava (1998) and Huddleston et al. (2009) were used again. Regarding the assortment quality perception, the fol-lowing questions were formulated: ‘The products in this category are of high quality’, ‘I trust the quality of these products’ and ‘The products in this category have consistent quality’.
Regarding the perceived price perception of the product group, the following statements were used (based on Bauer, Kotouc and Rudolph, 2012): ‘This assortment offers good value for the money’, ‘This assortment offers various price ranges to choose from (e.g. budget, value, premium)’ and ‘I am satisfied with the general price level of this product category’.
Lastly, purchase intention was measured. According to Wertenbroch and Skiera (2002) the method of measuring purchase intention can be criticized. They argue that willingness to purchase does not measure actual buying behaviour. An in-store experiment would have been more valuable, to estab-lish whether the purchase intention corresponds with the actual purchase behaviour. Due to the limited extent of this study, it was impossible to measure the actual purchase behaviour of NBs in the assortment at an HD. Bergvist and Rossiter (2007) argue that purchase intention should utilize a sin-gle-item measure. Therefore, the respondents were asked the following question about the category purchase intention: ‘Is it likely that I will purchase a product from this category?’. 4.4.3. MODERATOR
con-sciousness and brand consciousness. Lastly, several questions about the buying behaviour of re-spondents were included, such as the familiarity with Aldi, which brand they usually buy and brand equity. The psychological variables were measured through several items. A scale by Rao and Bergen (1992), was used to measure the construct ‘quality consciousness’. In this research, these scales are used to evaluate consumer’s quality consciousness towards store. A minor adjustment was made to measure quality consciousness towards products. The following questions were included to measure the con-struct: ‘High quality is very important to me when buying products’, ‘It would mean great monetary loss for me if the quality of products would turn out to be disappointing’ and ‘In general, I check to ensure that the quality of products is high before I purchase them.’ To measure the extent of price consciousness, a scale was used that has been developed and tested by Ailawadi, Pauwels and Steenkamp (2008). The construct was measured by the following question: ’For me, price is decisive when I am buying a product’, ‘Price is important to me when I choose a product’ and ‘I generally strive to buy products at the lowest price’. Finally, as proposed by Sproles and Kendall (1986) and Sharma and Chan (2011), brand consciousness was measured by the statements: ‘I tend to choose the more expensive brands’, ‘The well-known national brands are best for me’ and, ‘The most adver-tised brands are usually very good choices’. Additionally, familiarity with a store of Aldi was measured by asking respondents ‘Out of every ten supermarket visits, how often have you visited the following supermarkets?’. The question was fol-lowed by a series of listed supermarkets in the Netherlands. Familiarity with Aldi and their PLs may influence the results, because consumers tend to buy brands that are familiar to them and have met their expectations in the past (Nelson, 1970). To ensure that the results about the success of adding NBs in assortment are based on the enlarge-ment of the assortment and not on the fact that respondents’ favourite brands are included in the assortment, respondents were asked which brands they typically buy within the product category along with the brand equity of the presented brand. Brand equity was measured by three items de-veloped by Yoo (2000): ‘Even if another brand has same features as X, I would prefer to buy X’, ‘I have a strong preference for X’, ‘It makes sense that people pay a bit more for X than for other brands’. One more question was included: ‘X is of high quality’.
4.5 DATA ANALYSES
5.
D
ESCRIPTIVE RESULTS
In this chapter, the descriptive results are described. It starts with data clarification. To gain insight into the representativeness of the group, the demographics of the respondents are described in the second section. Subsequently, the reliability and consistency of all the variables are tested, to check whether the variables are measured in the correct manner. Next, the assumptions of the models are discussed. After that, the results of the scores on the hedonic level are measured.5.1 DATA CLARIFICATION
The data for this study were collected within two weeks. Respondents could fill in the survey online via a link via social media. 351 respondents filled out the questionnaire. Within this group, 30 re-spondents only answered approximately the first 10 questions. Two rere-spondents had clearly not taken it seriously, they filled in the same answers at every question, indicating unreliable results. Hence, after excluding these respondents, the sample consisted of 319 respondents.5.2 REPRESENTATIVENESS
Furthermore, regarding the income variable, the Dutch Central Bureau of Statistics (CBS) shows that the average income of a person is € 32,300 per year before taxes (CBS, 2016). Due the network of students and young employees it was not surprising that the largest group had an income below €15,000 per year (30.1%). The second largest group had an income between €25,000 and €35,000 which is comparable to the average income in The Netherlands. In general, the distribution in income was evenly distributed. Moreover, most participants were highly educated and had finished HBO or higher education (68%). This could be explained by the fact that students were targeted for participa-tion. The results of nationality, education level and income of the sample are summarized in Table 7. Table 7: Socio-Demographic characteristics Sample Nationality Dutch 93.4% Other 6.6% Education level Elementary school 0.3% High school 13.2% MBO 18.5% HBO 36.7% University bachelor 16.6% University master 14.1% PHD 0.6% Income (per year, before taxes) Less than 15,000 30.1% Between €15,000 and €25,000 14.7% Between €25,000 and €35,000 21.3% Between €35,000 and €45,000 14.7% More than €45,000 19.1%
Because of the difference between the sample and the regular Dutch shopper, it should be taken into account that some groups had a larger impact on this research than other groups. Therefore, age, education level and income level were included as control variables for further analyses.
5.3 RELIABILITY AND CONSISTENCY
To explore whether the variables from the survey establish the underlying dimensions, a factor anal-ysis was performed. Subsequently, the Cronbach’s alpha’s were calculated to check whether the variables were also internally reliable and consistent. 5.3.1 FACTOR ANALYSES(KMO) and Barlett’s test of sphericity measure were used to test if the factor analysis was an appro- priate technique for analysing the data (Malhotra, 2009). Malholtra (2009) considers that KMO val-ues less than 0.5 should not be acceptable. A KMO value above 0.5 indicates that the correlations between pairs of variables can be explained by other variables. The principle component extraction method was used to derive the factor score coefficients. This analysis was used to explain the maxi-mum amount of variance with the fewest number of principal components (Malhotra, 2009). Fur-thermore, in case of multiple components, the component matrices were rotated using Varimax to simplify the interpretability of the factors (Malhotra, 2009).
Table 9: Factor analysis price-, quality-, and variety perception Rotated component matrix 1 2 3 P1. 0.769 P2. 0.708 0.288 P3 0.871 Q1 0.744 Q2 0.808 Q3 0.853 V1 0.876 V2 0.683 V3 0.797 Before making any decision, it was important to look at the meaning of the item P2. For this item, respondents answered the following question: This assortment offers various price ranges to choose from (e.g. budget, value, premium)? It was not surprising that this item loads heavily on the compo-nent associated with variety perceptions, since it referred to a variety of assortment prices. Before deciding whether or not to exclude this item (P2) from further analysis, internal consistency was evaluated. 5.3.2 INTERNAL CONSISTENCY To measure the reliability of the variables, the Cronbach’s alpha’s were calculated for each subscale. This coefficient varies between 0 and 1. According to Malholtra (2009), a value of 0.6 or more indi- cates satisfactory internal consistency reliability. This analysis was performed for the variables: quali-ty-, price- and variety perception, hedonic level, price-, brand- and quality consciousness, and brand equity.
and all included items. After conducting a factor analysis and reliability analysis, the items of each construct were merged into one scale by calculating the mean.
5.4 ASSUMPTIONS
6.
H
YPOTHESES TESTING
In this chapter, the conducted analyses are discussed. The first part shows an overview of the means of each dependent variable. The second part discusses the effect on the assortment perceptions and includes H1a/b, H2a/b, H3a/b and H4a/b. The third part focuses on the mediation analysis and in-cludes H5a/b. The fourth part analyses the moderation effect which include hypotheses 6a/b.
6.1 COMPARING MEANS
Table 10 shows the means of each four dependent variables. Based on Table 10, it is concluded that the assortment with NB and 11% discount has the highest score on price perception, quality percep-tion and variety perception. The experimental group with an NB and 0% discount has the highest purchase intention. However, there is only a small difference with the experimental group NB and 11% discount. Both PFL groups (0% and 11% discount) have the lowest score on price perception, quality perception and purchase intention. The control group which included only a PL has lowest score on variety perception in comparison.Table 10: Comparing M(SD) of the dependent variables Conditions
Price perception Quality perception Variety perception Purchase intention Private label (N=64) 5.25 (0.83) 4.67 (0.99) 3.81 (1.28) 4.20 (1.61) Premium fancy label & 0% discount (N=62) 4.65 (1.15) 4.15 (1.10) 3.86 (1.42) 3.85 (1.74) Premium fancy label & 11% discount (N=66) 4.97 (1.13) 4.30 (1.11) 3.92 (1.29) 3.91 (1.83) National brand & 0% discount (N=67) 5.25 (1.05) 5.00 (1.00) 4.40 (1.41) 4.85 (1.62) National brand & 11% discount (N=60) 5.32 (0.96) 5.13 (1.01) 4.45 (1.55) 4.82 (1.74) Total (N=319) 5.09 (1.06) 4.65 (1.10) 4.09 (1.41) 4.33 (1.75)
6.2 (M)ANCOVA ANALYSES
A (M)ANOVA test can be used to determine whether there are significant differences between the means of three or more independent variables (Field, 2009). The control variables are included in each model.
6.2.1 THE EFFECT OF GROUP ON THE ASSORTMENT PERCEPTIONS
The differences between group (PL/PFL/NB) on the assortment perceptions, with the moderator hedonic level (hedonic/utilitarian) were analysed. The following hypotheses are discussed:
•
H1a: Compared to the situation where hard discounters only carry PLs, adding NBs to the as-sortment negatively influences the perceived assortment price
•
• H3a: Compared to the situation where hard discounters only carry PLs, adding NBs to the as-sortment positively influences the perceived assortment variety First, we investigated if the means of all assortment perceptions together differed significantly be-tween the control group (PL) and the two experimental groups together (PFL and NB), to examine whether HDs should add an extra brand to their assortment. The overall model showed that this was the case: V=.03, F(3,306)=3.63, p=.013. However, the two groups only significantly differed on variety perception, F(1, 308)= 4.85, p=.028. The next step was to analyse which of the groups scored differ-ently on assortment perceptions. Table 11 shows that the experimental groups scored higher on variety perception. Table 11: Means of the control group and the experimental groups on variety perception Group Mean Control group 3.74 Experimental groups 4.17 Second, we investigated the effects of group (PL, PFL and NB) and hedonic level (hedonic/utilitarian) on assortment perceptions (price, quality and variety). The overall MANCOVA model showed that, using Pillai’s trace, there was a significant effect of group, V=.15, F(6,606)=8.45, p<.001, no effect of hedonic level, V=.02, F(3,302)= 1.15, p=.210, and a significant interaction between group and hedonic level on the combination of the three assortment perceptions, V=.081, F(6,606)=4.25, p<.001. Fol-low-up analyses to interpret the interaction are described in section 6.4.
The separate ANCOVA’s explained 11.6% of the quality perception, 4.5% of the price perception, and 9.1% of variety perception. These were the adjusted R squared. Furthermore, the separate AN-COVA’s showed that group (PL/PFL/NB) differed significantly on quality perception, F(2,304)=19.61,
p<0.001, price perception, F(2,304)=7.54, p=.001, and variety perception, F(2,304)=7.35, p=.001.
Table 12: comparing means per group
Dependent variable Group Mean Std. Error Price perception PL 5.23 .130 PFL 4.80 .092 NB 5.28 .092 Quality perception PL 4.64 .132 PFL 4.23 .093 NB 5.06 .093 Variety perception PL 3.75 .170 PFL 3.91 .120 NB 4.44 .120 6.2.2 THE EFFECT OF GROUP ON PURCHASE INTENTION To measure the direct effect of group on purchase intention and the moderation of hedonic level, an ANCOVA was conducted. The next hypothesis was tested: • H4a: Compared the situation where hard discounters only carry PLs, adding NBs to the as-sortment positively influences purchase intention First the means of the control group (PL) and experimental groups together (PFL and NB) were ana-lysed. The ANCOVA showed that purchase intention did not differ significantly between the groups, F(1,309)=.98, p=.324. As can be seen in section 6.1, PFL scored lowest, whereas PL and NB scored higher on purchase intention. Therefore, the difference between PL and NB on purchase intention were further examined. Next the means of the three groups were analysed separately. Levene’s test was violated and there- fore it was decided to use a p-value of .01 instead of .05 for the effects of group on purchase inten-tion. The ANCOVA model explained 7.4% variance of the total variance of purchase intention. The ANCOVA showed that group significantly differ from purchase intention, F(2,305)=11.35, P<.001. There was a non-significant interaction effect of the moderator hedonic level, F(2,305)=.49, P=.612. Figure 7 shows the mean scores per group on purchase intention. Based on figure 7, it is concluded that the purchase intention is highest by introducing an NB in assortment. However, it was not yet clear whether this effect was significant. 4.15 3.87 4.87 1 2 3 4 5
Control group Fancy label National brand
The addition of NB increased purchase intention in comparison with PFL (p<.001) and PL (p=.019). However, the means between PL and PFL were not significantly different from each other (p=.878). The addition of an NB positively influenced purchase intention. Hence, hypothesis 4a was supported. 6.2.3 THE EFFECT PRICE LEVEL ON ASSORTMENT PERCEPTIONS In this section, the following hypothesis are discussed: • H1b: Compared to the situation where hard discounters use the same prices as traditional re-tailers, using prices that are considerably lower positively influences the perceived assortment price •
H2b: Compared to the situation where hard discounters use the same prices as traditional re-tailers, using prices that are considerably lower negatively influences the perceived assort-ment quality • H3b: Compared to the situation where hard discounters use the same prices as traditional re-tailers, using prices that are considerably lower do not influences the perceived assortment variety A MANCOVA was used to investigate if differences in price level (0% vs 11% discount) influenced the assortment perceptions and if this effect was moderated by hedonic level (hedonic/utilitarian). This analysis also included the PFL and NB groups to be able to measure the price level effects across dif-ferent groups. The group who saw only a PL was excluded, as this group did now view different price levels.
Pillai’s Trace showed that a significant effect of group on the assortment perceptions, V=.14,
F(3,238)=13.39, p<.001. However, price level did not influence assortment perceptions, V=.012, F(3,238)=.98, p=.404. Hedonic level had a significant effect on the perceptions, V=.07, F(3,238)=6.17, p<.001. The interaction between group and hedonic level was significant again, V=.04, F(3,238)=3.28, p=.022, but there was no interaction effect between price level and hedonic level, V=.01,
F(3,238)=1.12, p=.341.
The adjusted R² of the follow-up univariate ANCOVA for quality perception was .141 and therefore the model explained 14.1% of the total variance of quality perception. The model explained 4.2% of the total variance of price perception. Finally, the model explained 8.7% of the total variance of vari- ety perception. To continue, the ANCOVA showed that group differed significantly on price percep-tion, F(1,240)=12.39, p=.001, quality percepety perception. To continue, the ANCOVA showed that group differed significantly on price percep-tion, F(1,240)=39.13, p<.001, and variety percepety perception. To continue, the ANCOVA showed that group differed significantly on price percep-tion,
F(1,240)= 9.28, p=.003. More importantly, the price level did not significantly influence any of the
and hedonic level only reached significance for variety perception, F(1,240)=4.24, p=.041. The inter-actions between price level and hedonic level were non-significant for all perceptions, F<3.07 and
not hold for price level as independent variable. It is concluded that H5b was not supported. There- fore, the focus is on H5a and mediation analysis was conducted with group (PL/PFL/NB) as independ-ent variable, assortmfore, the focus is on H5a and mediation analysis was conducted with group (PL/PFL/NB) as independ-ent perceptions as mediators and purchase intention as dependfore, the focus is on H5a and mediation analysis was conducted with group (PL/PFL/NB) as independ-ent variable. The interaction effect of hedonic level was not included. Due to the multi-categorical independent variable, it was impossible to include a moderator in the model of Hayes. Each step is discussed be-low.
• Condition 1: effect of X on Y has to be significant
The analysis started by measuring the effect between group and purchase intention. The analysis showed that NB positively influenced purchase intention in comparison with PL, 𝛽=.71, P=.008. PFL and PL have approximately the same score on purchase intention, these groups did not differ from each other, 𝛽=-.26, p=.334. The first condition was met. • Condition 2: effect of X on M has to be significant Table 13 visualizes the effect of group on the different assortment perceptions. NB had the highest positive influence on quality perception and variety perception in comparison with the PL and the PFL. There were no differences between PL and NB on price perception. PFL negatively influenced price perception compared to PL (and thus also to NB). Hence, the groups significantly influenced assortment perceptions.
Table 13: regression results for assortment perceptions; standard errors between brackets Price perception Quality perception Variety perception
𝜷 Sig 𝜷 Sig 𝜷 Sig
Dependent variables PFL -.42 (.16) .009*** -.40 (.16) .015** .18 (.22) .405 NB .05 (.16) .758 .42 (.16) .009*** .68 (.21) .002*** P-value .005 .000 .014 F-value 2.62 5.15 2.27 R-square .079 .144 .069 *significant at 10% **significant at 5% ***significant at 1% • Condition 3: Effect of M on Y has to be significant
Quality perception and variety perception positively influenced purchase intention, resp. 𝛽=.47, p<.001 and 𝛽=.58, p<.001. Moreover, price perception did not significantly influence purchase per-ception, 𝛽=.03, p=.668. This condition was met for quality perception and variety perception.
• Condition 4: Including M must significantly reduce the effect of X on Y
Including the assortment perceptions in the model resulted in a non-significant relation between group and purchase intention (p>.05). Hence a full mediation was present. It was checked if the mediation of each assortment perception was significant. After bootstrapping, the confidence intervals of the reduction in the regression coefficient for group on purchase inten- tion should not contain zero. The indirect effect of group on purchase intention through quality per-ception was significant, the interval did not contain zero. As expected, the mediation effect of price perception was not significant, the interval contained zero. Finally, the indirect effect through variety perception was significant. However, for the confidence intervals of the different groups on variety perception, the significant effect applied only to NB compared to PL. PFL compared to PL revealed an insignificant result. It was concluded that hypothesis 5a was supported. However, the mediation effect only applied for quality perception and variety perception (NB compared to PL). 6.3.3 REGRESSION OF THE WHOLE MODEL EXCEPT CONTROL GROUP All variables were included in two regression models. As the price level was not evenly distributed across the groups, i.e., the control group was not presented with different price, there was high mul-ticollinearity between group and price level. For this reason, it was decided to only include PFL and NB in the regression model (eliminating the multicollinearity problem). The two models in Appendix 8 were compared to each other. In the first model, two independent variables were included (group and price level) and also the moderation effects. In the second model, the three assortment perceptions were included in the model. The model showed the same effects of the mediation model as discussed in the previous section: adding NB did not significantly influence purchase intention anymore when the assortment perceptions were included in the model. Further-more, hedonic level did not moderate the effects in either one of the models.
6.4 THE EFFECT OF HEDONIC PRODUCTS
In the previous sections, it was discussed which interaction effects were significant and which were not. Hence, this section discusses the following hypotheses: • H6a: The effects of listing NBs at hard discounters on assortment perceptions and purchase intention are stronger for hedonic than for utilitarian products• H6b: The effects of using prices that are considerably lower than traditional retailers are
The MANCOVA analysis showed a significant interaction between group and hedonic level on variety perception. Figure 9 shows that the scores for milk did differ per group on variety perception where-as the scores on chips were about the same. However, it is necessary to understand what the reason was for the significant interaction effect on variety perception. To establish if this was a significant result, an ANCOVA analysis was conducted. By splitting the file on hedonic level, the overall model showed a significant result for group of the product type milk,
F(2,149)=13.32, p<.001, and a non-significant result for chips, F(2,147)=.30, p=.739. Figure 9 shows
7.
C
ONCLUSION
&
RECOMMENDATIONS
After measuring and analysing the data, the results of these analyses are discussed in this chapter. The chapter will continue by discussing the managerial and the academic contribution. Finally, the limitations and ideas for further research are described.
7.1 CONCLUSION & DISCUSSION
7.1.5 HEDONIC LEVEL OF THE PRODUCT
Hedonic level of the product only influenced the relation between brand and variety perception. Results showed that for milk (utilitarian product), adding an NB increased variety perception, where-as for chips (hedonic product) this was not the case. An explanation could be that consumers have different expectations of utilitarian and hedonic products when it comes to variety. Consumers may not expect a lot of variety in a product category as milk, whereas for chips they do. For the relation- ships between group and price perception, quality perception and purchase intention, it did not mat-ter which product type was used. This also applied to the relationship between price level and the outcome variables (price-, quality-, variety perception and purchase intention). According to the study of Lee and Hyman (2008), consumers perceive supermarkets as functional stores. Functional stores are more successful with selling utilitarian brands, instead of hedonic brands, which are more successful in department stores. Hence, it could be that the different effects between hedonic and utilitarian products are smaller for functional stores like supermarkets than for department stores.
7.2 MANAGERIAL CONTRIBUTION
7.3 ACADEMIC CONTRIBUTION
R
EFERENCES
• Ahold (2012). Reshaping retail: Annual report 2012. Retrieved from https://www.aholddelhaize.com/media/1417/ahold-full-ar-2012.pdf. Viewed 12 March 2017. • Ailawadi, L.K., Lehmann D.R. & Neslin, S.A. (2003). Revenue Premium as an Outcome Meas-ure of Brand Equity. Journal Of Marketing, 67(4), 1-17. • Ailawadi, L.K. & Keller, K.L. (2004). Understanding retail branding: conceptual insights and re-search priorities. Journal of Retailing, 80(4), 331-342. • Ailawadi, L.K., Pauwels, K. & Steenkamp, J-B. (2008). Private-label use and store loyalty. Jour-nal of Marketing, 72, 19–30. • Alford, B.L. & Biswas, A. (2002). The effects of discount level, price consciousness and sale proneness on consumers' price perception and behavioural intention. Journal of Business re-search, 55(9), 775-783. • Argentesi, E., Buccirossi, P., Cervone, R., Duso, T. & Marrazzo, A. (2016). The effect of retail mergers on prices and variety: An ex-post evaluation. DICE Discussion Paper. • Babin, B.J., Darden W.R & Griffin M. (1994). Work and/or fun: measuring hedonic and utili-tarian shopping value. Journal of Consumer Research, 20, 644–656.
• Baltas, G. (1997). Determinants of store brand choice: a behavioural analysis. Journal of
product and brand management, 6(5), 315-324. • Batra, R. & Ahtola, O. (1990). Measuring the Hedonic and Utilitarian Sources of Consumer At-titudes. Marketing Letters, 2(2), 159-170. • Bauer, J.C., Kotouc, A.J. & Rudolph, T. (2012). What constitutes a ‘‘good assortment’’? A scale for measuring consumers’ perceptions of an assortment offered in a grocery category. Jour-nal of Retailing and Consumer Services, 19, 11–26. • Bell, D. R. & Bucklin, R. E. (1999). The role of internal reference points in the category Pur-chase decision. Journal of consumer research, 2, 128-143. • Bellizzi, J.A., Kruckeberg, H.F., Hamilton, J.R. & Martin, W.S. (1981). Consumer perceptions of national, private, and generic brands. Journal of Retailing, 57(4), 56–70. • Berger, J., Draganska, M. & Simonson, I. (2007). The influence of product variety on brand perception and choice. Marketing Science, 26(4), 460–72. • Bergvist, L. & Rossiter, J.R. (2007). The predictive validity of multiple-item versus single-item measures of the same constructs. Journal of Marketing Research, 44, 175-84. • Berkhout, P. (2017). Food Economic Report 2016 of the Netherlands; Summary. Wageningen Economic Research, Booklet. 16. • Burt, S. (2000). The strategic role of retail brands in British grocery retailing. European Jour-nal of Marketing, 34(8), 875 – 890. • CBS (2016). Gemiddeld inkomen van personen naar kenmerken en naar regio. Retrieved from http://statline.cbs.nl/StatWeb/publication/?VW=T&DM=SLNL&PA=80500NED&D1=2-3&D2=l&D3=0-38&D4=0-4&D5=l&HD=100903-1406&HDR=G4,T,G1,G3&STB=G2. Viewed 16 May 2017. • Chang, T.Z. & Wildt, A.R. (1994). Price, product information, and purchase intention: An em-pirical study. Journal of the Academy of Marketing Science, 22, 16–27.
• Chitturi, R., Raghunathan, R. & Mahajan, V. (2008). Delight by design: The role of hedonic versus utilitarian benefits. Journal of Marketing, 72(3), 48-63. • Chowdhury, J., Reardon, J. & Srivastava, R. (1998). Alternative Modes of Measuring Store Im- age: An Empirical Assessment of Structured versus Unstructured Measures. Journal Of Mar-keting Theory And Practice, 6(2), 72-86. • Cleeren, K., Verboven, F., Dekimpe, M. G. & Gielens, K. (2009). Intra- and interformat compe-tition among discounters and supermarkets. Marketing Science, 29(3), 456–474.
