National vs. Store Brands: The Effect of Crowding

(1)

National vs. Store Brands: The Effect of Crowding

The influence of perceived crowding in front of the shelf and the moderating role of the relative budget on brand choice

(2)

National vs. Store Brands: The Effects of Crowding

The influence of perceived crowding in front of the shelf and the moderating role of the relative budget on brand choice

University of Groningen Faculty of Economics and Business

Department of Marketing

Master thesis, MSc Marketing Management

January 12, 2015 Ilektra Mitsa Student number: S2628996 Damsterdiep 196 9713 EN Groningen Tel.: +31 (0)633337802 E-mail: elektra.mitsa@gmail.com

(3)

1 Abstract

Many prior researchers have focused on the effects of retail crowding on shopper behavior. While at the theoretical level a diverse number of reasons have been proposed in order to understand the reasons why consumers prefer store brands to national brands, none has focused on the effects of perceived crowding in front of the shelf on brand choice. The present study aims to fill this gap by testing the potential effect of different levels of perceived crowding in front of the shelf and level of relative budget as possible explanatory variables about brand choice in five product categories: cookies, coffee, milk, butter and sugar. The results indicate that perceived crowding has no effect on the overall likelihood of purchasing a national over a store brand. Analyzing each of the five product categories separately, perceived crowding had a direct significant effect only for the cookie category. A high perceived crowding in front of the shelf increases significantly the likelihood of purchasing a store cookie brand. Since perceived crowding had no significant direct effect on brand choice, no moderation effect could exist. In the cookie category, the interaction between the two variables was not significant. However, the results showed that the level of relative budget has a strong direct positive effect on the likelihood of purchasing a national over a store brand in the cookie, coffee and the overall category. To conclude this study the managerial implications and future research ideas are presented.

Key-words: perceived crowding, relative budget, store brands, national brands, shelf,

(4)

2 Preface

This master thesis is written in the field of marketing management, and performed at the University of Groningen, where it is the final assignment before graduating. The paper provides insights on how different levels of perceived crowding can affect brand choice. This topic has received increasingly more attention in the past years and was therefore proposed by the University of Groningen.

Special gratitude goes to Prof. dr. Lara Lobschat who guided me throughout the process and

provided me regularly with valuable feedback. My sincere thanks also go to Daniela

A. Naydenova for her evaluation, and my fellow students who provided me with useful feedback. Lastly, I would like to express my gratitude to all the respondents for their participation in this research as well as the people who helped distributing the online questionnaire.

Groningen, January 2014

(5)

3 Table of Contents

1. Introduction ... 4

2. Theoretical Background ... 8

2.1 Crowding ... 8

2.2 Retailers and Brand Allocation... 10

2.2.1 Shelf Placement ... 11

2.3 Relative Budget ... 13

2.4 Summary ... 16

3. Methodology ... 16

3.1 Research Design and Participants ... 16

3.2 Shelf presentation setting ... 17

3.3 Measuring the Variables ... 19

3.4 Analysis ... 20

4. Results ... 21

4.1 Pre-test ... 21

4.2 Sample Characteristics ... 21

4.2.1 Definitive Number of Participants ... 21

4.2.2 Description of the Sample ... 22

4.3 General Preferences Variables ... 23

4.4 Perceived Crowding Variable ... 24

4.5 Likelihood of Purchasing Variable ... 25

4.6 Delta Variables Before and After Manipulation ... 25

4.7 ANCOVA Analysis ... 26 4.7.1 Cookie Category ... 26 4.7.2 Coffee Category ... 27 4.7.3 Milk Category ... 28 4.7.4 Butter Category ... 28 4.7.4 Sugar Category ... 29 4.7.5 All Categories ... 29 4.8 Summary ... 30 5. Discussion ... 31 5.1 Managerial Implications ... 33

5.2 Limitations and Further Research ... 34

References ... 36

Appendix A: Questionnaire ... 42

(6)

4 1. Introduction

Nowadays, a significant amount of effort has been devoted to discovering and comprehending the processes which consumers arrive at some type of decision, especially a purchase. There is a substantial amount of research on how even a minor change of the in-store environmental cues can alter the consumer’s behavior. The accumulated literature has shown that changes in store’s atmosphere e.g., color, music, and crowding, can influence consumers’ judgments and purchasing behavior by evoking a wide range of behavioral responses as for instance sales, impulse buying and time spent in the store. These changes to environmental stimuli may or may not be perceived by the consumer (Turley and Chebat 2002; Turley and Milliman 2000). Moreover, there is prior literature focused on the extent that responsiveness to the physical design and condition can influence a shopper’s decisions on the location and time spent for shopping, as well as, the extent to which crowding can lead the shopper to alter purchases or even leave the premises (Ailawadi and Keller 2004; Eroglu and Machleit 1990; Milliman 1982; Turley and Milliman 2000). As a consequence, retailers are investing a huge amount of their budgets in order to gain insights about constructing a more effective retail environment that generates more sales, improves customer relationships, and boosts return on marketing investment, by altering physical conditions to make the shopping experience more pleasurable (Mulhern 1997).

(7)

5

crowding in front of the shelves can limit the visibility of some brands, especially the store brands as they are usually located on the lower shelves. Hence, many questions can be derived on this under-researched area such as: Does perceived crowding in front of the shelf affect brand choice, on planned purchases? Do consumers react to the feeling of confinement by choosing national brands or store brands? How does the arrangement of the products on the shelf affect their choice, when visibility is compromised? Does their level of relative budget affect their choice? All these questions, lead to the assumption that investigating these areas might prove to be highly insightful for retailers, manufacturers and shoppers.

Grocery shopping is considered a weekly routine for the shoppers, which is characterized by achieving multiple goals, through processing all in-store stimuli such as product and brand information. The variety of these situational factors can affect purchase intentions and outcomes, such as sales and brand switching (Park, Iyer and Smith 1989). Shankar (2011) mentions that in-store merchandising variables such as aisle placement, shelf place positioning, and displays, influence shopper perceptions and behavior. More specifically, Valenzuela and Raghubir (2009) suggested that the placement of the products on the shelf influences shopper perceptions as in their experiments they found out that a product in a central position was preferred over one at either end of the shelf, as consumers believed that the product in the center was the most popular. In addition to that according to Chandon et al. (2009) top and middle shelf positions gain the most attention. Furthermore, Turley and Milliman (2000) stress that even though there are numerous studies examining shelf-space decisions, there are still many research gaps such as the effect of crowding perceptions and that can be further examined. Moreover, one can question if brand category, national and store brands may have an effect on shelf-space decisions as well.

(8)

6

a consequence, store brands are usually placed on the bottom shelves, encompassing the leading national brands, offering good alternatives to the shoppers.

Considering the global economic situation and that more people seek products at lower prices as their budgets become tighter, the increasing market share of store brands can be partly explained. Nowadays, it is a common practice of shoppers to pre-assign a specific amount of their budget for a specific purchase before the actual time of the buying, in order to control their expenses. Following this line, a different stream of research has focused on mental accounting and its effects on shopper behavior. They found out that shoppers set mental accounts to control expenses by allocating budget limits in certain categories, such as, expenses for household purchases, entertainment or food. (Heath and Soll 1996). Further research on the field suggested that resource depletion and budget exhaustion can influence consumer behavior (Kamakura and Du 2012; Soster, Gershoff and Bearden 2014). For instance, consider a consumer who allocates €50 to his weekly grocery budget. After his first trip to the grocery store in which he will spend for example €10, he will still have access to the other €40. As he decides to make additional purchases during the week, his budget will reduce accordingly. For example later in the week, he may decide to purchase additional products worth €10. This €10 purchase will further reduce his budget balance, which may now approach complete exhaustion. As a consequence, he may alter his consumption patterns due to his financial situation (Soster, Gershoff and Bearden 2014).

Similarly in this study, we propose that assigning a specific amount of budget that will lead the shopper to its exhaustion, by purchasing planned common purchased products, such as milk, will affect his choice between national and store brand products, depending on the level of perceived crowding in front of the shelf. By assigning a relatively low budget the shopper has to take into consideration its exhaustion and thus balance his brand choices, by including some lower priced store brand products. This is in line with the tight budgeting that the majority of people are exercising due to the global economic situation. Moreover, we speculate that different levels of perceived crowding will vary the visibility of the majority of brands, making it more effortless or effortful to choose the store brands. Particularly in this study we examine the likeliness of purchasing a national over a store brand in 5 different categories: 1) cookies, 2) coffee, 3) milk, 4) butter and 5) sugar, as they are basic, low-cost, common purchased products that the majority of consumers buy regularly.

(9)

7

on brand choice when the shopper perceives high crowding in front of the shelf. It addresses the following research questions: How different levels of perceived crowding in front of the shelf, affects brand choice? Do shoppers choose the national brands that are usually placed on the upper and middle shelves or the store brands that are placed on the lower shelves? and Does shoppers’ level of relative budget has a moderating effect on the outcome of brand choice? In that way, we expect to contribute to prior literature as we will attempt to fill the knowledge gap on how perceived crowding can affect brand choice on a planned purchase. This research is relevant as it will connect two already researched areas, crowding and shopper brand decisions.

The managerial implications of identifying the effects of crowding on brand choice could facilitate brand switching to the retailer’s store brand products. It is possible to enhance the store’s image and shopper satisfaction by assisting the consumer in adapting to the perceived crowding (Harrell, Hutt and Anderson 1980). Moreover, store brand managers can identify the correct position of their product on the shelf during the busy days. This could enhance their brand image and increase sales, as they will become knowledgeable about the effect of high perceived crowding in front of the shelf which leads to the lower visibility of the brand. In the end, managers would be capable of anticipating the changes in shopper shopping behavior under different crowding conditions and adjust merchandising and promotion strategies accordingly (Harrell, Hutt and Anderson 1980). For example, they can adjust the merchandising on the shelves in a way that even during busy hours the store brands remain visible, by adjusting the aisle width or the structure of the shelves. It could be effective if on the busy days, when visibility of the store brands lowers and sales decline, to advertise promotions or special offers. Similarly, managers of the national brands could negotiate with the retailers about the relocation of their products in order to ensure that their sales would not transfer to the store brand products. As it was mentioned above, in-store environmental cues can affect shopper behavior therefore, the development of specific environmental designs could alleviate the feeling of being crowded, by re-adjusting the crowding density.

(10)

8 2. Theoretical Background

As it was mentioned before, there is no prior research that investigates the effect of perceived crowding in front of the shelf, on brand choice. This section reviews the prior literature on perceived crowding, shopper’s in-store budget considerations, brand choice in terms of national and store brands, proposes a conceptual framework for these variables, and derives some hypotheses for empirical research. Figure 1 shows a conceptual framework of how the variables are related to one another.

2.1 Crowding

The majority of research on environmental aspects of retail shopping behavior has focused on intangible elements such as shelf space, scent, lighting, noise and aisle designs. According to Kotler (1974), crowding due to the density of shoppers within the retail store may alter the retail atmosphere in certain ways that can allow predicting shopper behavior and thus purchasing intention. Grossbart et al. (1990) in their research, defined shopper atmospheric responsiveness, as ―the tendency to base patronage decisions on stores’ physical design and condition and alter shopping behavior because of crowding‖. They investigated the extent that responsiveness to the physical design and condition can influence a shopper’s decisions on the location and time spent for shopping, as well as, the extent to which crowding can lead the shopper to alter purchases or even leave the premises.

Harrell, Hutt and Anderson (1980) argued that crowding can be viewed as both a physical state of high density, which depends on the number of persons in a specific space, and as a psychological state of the individual (Stokols 1972, 1976). According to Eroglu, Machleit and Barr (2005), it is crucial to separate the meanings of density from perceived crowding. Density is an antecedent of perceived crowding as it is only when density interferes with shoppers’ goals and actions that the environment can be perceived as crowded. Due to different shopper characteristics or environmental constraints, two different shoppers in the same store may perceive different levels of crowding.

(11)

9

shoppers may choose to postpone unnecessary purchases and reduce their shopping time. The last adaptation strategy Milgram suggested is that social interactions with store employees and other shoppers are being limited. Moreover, Eroglu, Machleit and Barr (2005) argued that shoppers with higher levels of crowding perceptions receive lower levels of positive emotions while shopping. Similarly, Mackintosh, West and Saegert (1975) found that respondents performing an experimental task under high density conditions described their feelings as negative, as they felt ―tense‖ and ―confused‖. Evans (1979) supported that crowding acts as a stressor for humans as he found in his study that more frustration and hostility were voiced by crowded groups than by less crowded. Moreover, Saegert (1973) suggested that shoppers are incapable of recalling any details about merchandise and layout in crowded facilities. This is in line with the view that an overload of environmental stimulus can be distracting and can reduce the attentional capacity available for a specific task (Cohen 1978). Therefore, it can be assumed that environmental conditions, particularly crowding may affect not only the consumers' satisfaction with the shopping trip, but also their perceptions of a store’s image (Harrell, Hutt and Anderson 1980).

In their research, Levav and Zhu (2009) discovered that high levels of perceived crowding, which were created by spatially constraining people, can evoke feelings of confinement and create the need for reactance. They proposed that in shopping contexts reactance can manifest in consumers’ product choices. In Western societies, choices, especially unique and different choices, are considered as acts to express one’s freedom, which is a form of exhibiting reactance (Kim and Drolet 2003). Therefore, they hypothesized that people in relatively small confining spaces will tend to display a greater variety seeking in their choices. By manipulating aisle width, they found a positive relationship between perceived crowding and variety seeking behavior.

(12)

10

Figure 1. Conceptual framework (*: Reference Category).

2.2 Retailers and Brand Allocation

As it was mentioned in the previous section, different elements of a retailer’s in-store environment, e.g., color, music, and crowding, can influence shoppers’ perceptions of a store’s atmosphere, which as a consequence will affect their decisions of whether or not they visit a particular store, as well as the amount of money and time they will spend there (Ailawadi and Keller 2004; Eroglu and Machleit 1990; Milliman 1982; Turley and Milliman 2000). According to Ailawadi and Keller (2004) store atmosphere can mediate shopper perceptions of multiple dimensions of store image as for instance the economic and psychological costs of shopping in a store (Baker et al. 2002). More specifically, when shoppers find a store aesthetically pleasing, they rate store brands’ quality higher than when they find it less pleasing. Particularly the ratings of the private label’s quality are higher when consumers perceive the store more aesthetically pleasing than when they perceive it as less attractive. However, these findings are not applicable to national brands (Richardson, Jain, and Dick 1996b). Therefore, it can be assumed that an appealing in-store atmosphere might offer the opportunity of improving store brand products’ image.

(13)

11

Shoppers’ demands encompass the presence of both national and store brands in stores, otherwise the profitability of the retailer will be jeopardized (Ailawadi and Keller 2004; Anselmsson and Johansson 2014). Retailers are considering their store brands as worthy competitors of the leading national brands, however, they should treat carefully manufacturers as they are both competitors and customers. The majority of national brand manufacturers, nowadays, consider store brands as a threat as they are guaranteed full distribution and good shelf placement. An experiment conducted by Iniesta and Agustín (2001) compared the shelf space allocated to the store brands according to their market share based on their value and volume. They found that only in four out of the 28 categories retailers granted a space on shelves that was proportionate to their market share. In fact, they usually occupy greater space than would normally correspond to their market share (Suárez 2005). Considering that shelf space is one of the retailer’s most important negotiating assets with manufactures (Amrouche and Zaccour 2007), the retailer is in a better position in the competition between store brands and national brands as he is the one that allocates shelf space and the level of advertising in his assortment. On the other hand, the manufacturer’s position is highly dependent on his brand equity and market share (Juhl et al. 2006).

There is extensive literature about the relationship between the shelf space assigned to a product and the results that can be achieved. The most significant result is that by assigning more shelf space to a product category or a particular brand, their visibility will increase, and as a consequence so will the likelihood of purchasing it. Manufacturers pay a significant amount of money on slotting allowances, in order to gain a good shelf position (Chandon et al. 2009; Feighery et al. 2003; Sigurdsson, Saevarsson and Foxall 2009). For example, Feighery et al. (2003) interviewed retailers in order to gain insights into how tobacco companies convey promotional allowances and special offers in order to shape the retail environment to their advantage. They found that the tobacco companies competed with each other, as they wanted to achieve prime placement for their products in order to attain high visibility. On the other hand, retailers’ store brands are located around the leading brands, occupying greater space than other national brands with the same market share. All the above contributes to the easier comparison between the brands, with the lower price of the store brand standing out (Suárez 2005).

2.2.1 Shelf Placement

(14)

12

display can influence not only sales but quality expectations and brand choices. They found that choosing among unfamiliar wines, consumers tend to select the brands located at the top and in the middle of the vertical displays and the brands located in the center of horizontal displays (Chandon and Hutchinson 2009; Valenzuela and Raghubir 2009).

According to eye-movement studies, not all shelf locations attract equal attention. Chandon and Hutchinson (2009) found that positioning the brand in different location on the shelf can influence the attention and the evaluation of the product. More specifically, they suggested that the best position is in the top and near the center of the shelf as this position can improve attention and evaluation. Moreover, position on the middle shelves improves only attention, whether positioning them on the left- or right-hand side of the shelf has no significant effect on either. Furthermore, they found that increasing the facings of low-market-share brands was more effective than increasing those of high-low-market-share brands, supporting the belief that the higher the number of facings, the more important the brand is perceived. In their previous research they had supported that shoppers tend to fixate their attention on the center shelves, where the majority of leading national brands are positioned. Other researchers posit that it is more profitable to position the brand near eye or hand level, which can be translated into positioning near the top shelves, than on the lower shelves (Drèze, Hoch, and Purk 1994). Moreover, Raghubir and Valenzuela (2008) in their experiment found that shoppers chose brands located at the top or in the middle of vertical displays, supporting the opinion that consumers perceive that retailers are positioning the high-quality, more expensive brands on the top and near the middle shelves (Chandon and Hutchinson 2009).

(15)

13

Taking into consideration all the previous literature on shelf positioning, one can assume that the effort associated with looking at middle and top shelves is lower than the effort associated with looking at lower shelves. Moreover, the majority of retailers place the leading national brands on the middle to top shelves not only due to slotting allowances but also for attracting more shoppers, thus increasing retailers’ profits. As a consequence, store brands are usually placed on the bottom shelves, encompassing the leading national brands. As it was mentioned in the previous section, crowding can affect consumer behavior by producing either unpleasant feelings, promote postponing of the shopping or reduce shopping time. In addition to that it can encourage the use of heuristics for a purchase decision. Taking into account all the above, it is logical to assume that crowding in front of the shelf can decrease the visibility of the products even further. As Chandon and Hutchinson (2009) suggest in their article, the traditional justification for in-store marketing studies is that ―unseen is unsold.‖

In this study we will examine the effect of perceived crowding in front of the shelf on brand choice. We expect that shoppers will choose a brand which is located on top and near the middle shelves, as it requires less effort. Moreover, we suggest that crowding will reduce the visibility of the brands on the lower shelves and also reduce the shopping time as shoppers will most probably base their decisions on heuristics. Taken together, these arguments lead to our first specific hypothesis:

H1 +

: We expect that high perceived crowding in front of the shelf will increase the likelihood of purchasing a national brand product over a store brand product.

2.3 Relative Budget

(16)

14

self-control mechanisms that people can employ to prevent over-consumption and overspending, they often prove to be malleable and fallible. In their article they argue that consumers usually seek loopholes to circumvent their imposed self-control devices by classifying ambiguous expenses, by either assigning them to different mental accounts, or by constructing new mental accounts.

It is well known that consumers set budgets in advance of the actual consumption. Self-control literature is focused on separating the action of planning from doing, dividing consumers into rational "planners" and impulsive "doers" (Heath and Soll 1996). Categorization is crucial in the planning stage, as it determines whether or not a particular category’s mental budget should be depleted (Antonides, Groot and Raaij 2011). By dividing and labeling their available resources, consumers simplify cognitive calculations which allow them to monitor their consumption and expenses. However, Heath and Soll (1996) in their study they suggested that assigning money to particular category will lead to overspending or underspending as the consumer is likely to spend too much within that category, if the category budget is set too low. Moreover, they supported that previous expenditures within a category tend to decrease the likelihood of future expenses in that category.

Prior studies have further investigated the matter and suggested that consumers assign less resources on nonessential products during times of economic contractions (Kamakura and Du 2012), think differently about expenses depending their resource availability (Spiller 2011), or yield lower consumer satisfaction with a product as their budgets approach exhaustion (Soster, Gershoff and Bearden 2014). This phenomenon is known as the bottom dollar effect. Recent research supports that resource depletion can influence consumer behavior (Kamakura and Du 2012). Soster, Gershoff and Bearden (2014) suggested that two shoppers that have purchased the same product for the same price, might perceive different levels of product satisfaction, depending on their perceptions of payment pain, which is driven by their access to available budgetary resources. Stilley, Inman and Wakefield (2010), contributed to the mental account literature by proposing that mental budgets for grocery trips are composed of both an itemized portion and in-store slack. They defined the itemized portion as the budget that the consumer has allocated to spend on planned items and in-store slack as the portion of the mental budget that is assigned to be spent on unplanned in-store purchases.

(17)

15

weekly budget, for each shopping trip. It is a common practice that they calculate these expenses on their spending levels experienced in past shopping trips. Thus, it can be assumed that assigning a specific budget, for common repeat purchase products, will have an effect on shoppers’ brand choice. More specifically, by assigning a specific amount of money that will lead the shopper to almost the exhaustion of his budget, in order to purchase common products such as milk, will most likely affect his choice of purchasing a national over a store brand product. The shopper may have to consider that even though he may prefer the national brand product, his budget might not suffice for this purchase and therefore, he may opt for the store brand product. This effect will be weakened or strengthened by the different levels of perceived crowding in front of the shelf. It can be assumed that if the shopper encounters high crowding in front of the shelf, when the majority of the brands are not visible, he may base his decision on previous preferences or heuristics. Our study will therefore expand prior literature by incorporating brand choice, perceived crowding and relative budget theory.

We expect that when the shopper’s perception of crowding is high, and he is assigned to a low relative budget, the likelihood of purchasing a national brand product will decrease. We suggest that the shopper will most likely purchase a national brand product if he encounters high crowding in front of the shelf, due to the fact that he will most likely experience unpleasant feelings and therefore decide to reduce his shopping time by choosing a well-known national brand. In addition to that high crowding will lower the visibility of the majority of store brands which usually are located on the lower shelves. Moreover, as his budget may approach to exhaustion, he would have to consider purchasing store brands. On the other hand, we expect that in the other scenario, in which he is assigned to a higher relative budget, to increase the likelihood of choosing a national brand product as the budget will suffice for this purchase. We suggest that since choosing the national brand requires less cognitive effort due to mainly higher visibility, having sufficient budget will only strengthened the likelihood of purchasing a national over a store brand. Therefore, we expect relative budget to have a moderating effect on the likelihood of purchasing a national brand over a store brand, due to perceived crowding. These arguments suggest our second hypothesis:

H2

-: Lower relative budget in contrast to high relative budget, in high perceived crowding in

(18)

16 2.4 Summary

The prior sections show how it is expected that perceived crowding in front of the shelf and the level of relative budget can affect brand choice, on the basis of the likelihood of purchasing a leading national brand product over a store brand product. Next, we try to support our hypotheses by conducting several ANCOVA and factor analyses.

3. Methodology

In this chapter the setup of the research will be discussed together with the manipulation of the independent variables, followed by the operationalization of the concepts introduced so far. More specifically, the research design and the number of participants will be presented as well as, the plan for the data analysis and the measures of the variables.

3.1 Research Design and Participants

We designed a survey in which 138 participants (65 males and 73 females) were randomly assigned to two crowding scenarios. We conducted a 2 (low perceived crowding/high perceived crowding) × 2 (low relative budget/high relative budget) between-subjects experiment. The data were then analyzed using a full-factorial ANCOVA analysis.

(19)

17

again, the product category is specified as national cookie brands such as Oreo’s or store brands cookies such as Albert Heijn or Carrefour cookies.

In the first scenario, the level of budget was relatively high so that it would suffice for picking either brand product categories. In the other case scenario, the participants had to face the possibility of budget exhaustion and had to consider choosing store brand products instead of the more expensive national brand ones. More specifically, the participants were assigned to a budget of €12 which would suffice for choosing only three national brand products. The participants would then realize that they had to choose the store brand alternatives in the other two categories. We chose to introduce the concept of relative budget as the majority of our participants were students, between the ages of 18-34, who usually prefer store brands due to their restricted budget. All combinations taken together, leads to a total of 4 different experimental conditions. An overview is depicted in table 1.

Low Perceived Crowding High Perceived Crowding Low Relative Budget 1 person in-front of the shelf/

12€ budget

4 people in-front of the shelf/ 12€ budget

High Relative Budget 1 person in-front of the shelf/ 15€ budget

4 people in-front of the shelf/ 15€ budget

Table 1. Research Design

3.2 Shelf presentation setting

(20)

18

examples of existing brands to indicate the associations with the overall category of national and store brands. For example, we referred to Oreo cookies as a national brand in that category and Albert Heijn cookies as an example of a store brand in that category. In order to avoid the influence of established brand associations on the results we kindly incorporated a reminder that these are only examples and that we want to investigate their general preferences. The following text was used to briefly explain the situation to the respondent and to make sure that respondents did not indicate a preference for a specific brand but their overall preference in a specific product category:

“National brands are those product names that are promoted nationally, or even globally. For example, Oreo cookies, Nescafé coffee, Campina Milk, Becel butter, Van Gilse sugar are national brand products. Albert Heijn (AH) cookies or Carrefour cookies, AH coffee or Carrefour coffee, are store brand products. In this survey you are supposed to buy

five items, by choosing either category, in your hypothetical every day shopping list: 1) cookies, 2) coffee, 3) milk, 4) butter and 5) sugar. Please take into consideration that the

brands mentioned before are just examples of national brand names and store brands. We want to investigate if you choose national brands or store brands, not e.g. if you prefer Oreo cookies to AH cookies or Carrefour cookies”

In all experimental conditions respondents were asked to evaluate the level of crowding in front of the shelf, the likelihood of purchasing a national over a store brand in each product category, by using a structured questionnaire. The questions covered the dependent variable, the control variables and the moderator relative budget (see paragraph 3.3 for a deeper insight in the questions). Due to its simplicity and high feasibility the questionnaires have been distributed online, with the help of Qualtrics program, as it was regarded as an adequate research method. Each experimental condition contains at least 25 respondents, so that a total of 120 were assured. The online questionnaires were distributed using convenience sampling through Facebook, LinkedIn and via emails. The online questionnaire started with some demographic questions (age, gender, occupation, annual income and country of origin) followed by the two control variables (brand consciousness and brand preference on national or store brand for the five product categories). After completing these general questions, the participants were led to a screen with the introduction of the real research:

(21)

19 survey we are going to show you a picture of the crowding in front of the shelves at the moment. Please imagine that you have to purchase the items in your list, encountering the

situation that the picture shows you.”

The four different experimental conditions were randomly assigned to the participants. The questions below the picture were the same for all participants. The participants were able to see a picture of the shelf without crowding and then they were presented with either a picture of one person in front of the shelf (low perceived crowding condition) or a picture of four people in front of the shelf (high perceived crowding condition). After completing the questionnaire they were thanked for their participation. The complete questionnaire is included in the appendix.

3.3 Measuring the Variables

In order to measure each of our variables we adapted already existing validated and tested scales. As it was mentioned before, in the beginning of the questionnaire we included demographic questions such as age, gender, country of origin, occupation and annual income. Moreover, before any manipulation we included control questions about their existing brand preferences that would serve as covariates in the subsequent analysis. We used Shim and Gehrt’s (1996) brand consciousness scale, which contains three five-point Likert-type statements that measure the degree to which a person focuses on buying well-known brands, with reported alpha of 0.72. The respondents had to state their preferences by agreeing or disagreeing to statements such as: “The well-known national brands are best for me”, “The

more expensive brands are usually my choices”, “The higher the price of the product, the better its quality”. In addition to that we used two five-point Likert-type statements

measuring the degree to which a person views a focal brand as preferable to a reference brand. This is Sirgy et al.’s (1997) brand preference scale with reported alphas ranging from .72 to .98, for eight different products. We adapted the wording in order to correspond to our selected product categories. Therefore, respondents indicated how much they agree with statements like: “I like national brand cookies, e.g. Oreo cookies better than store brand

cookies, e.g. AH (Carrefour) cookies”.

(22)

20

is the current crowding in front of the shelves of all the items in their list. Respondents after that had to state for each of the five pairs of adjectives the extent to which they believed the adjectives described the picture. The adjectives included were: Not stuffy/Stuffy,

Uncrowded/Crowded, Uncramped/Cramped, Restricted/Free to move, Confined/Spacious.

Furthermore, we included a manipulation check by incorporating a question about the visibility of the majority of the brands in the picture.

In the next step, we introduced the relative budget scenario in which the respondents were randomly assigned to either the low or high scenario. They were assigned to either €12 or €15 budget and they were presented with the prices of each product in their list, national and store brand, in all five product categories. In each scenario they were informed if their budget would suffice for picking either brand categories or else that they had to take into consideration its exhaustion and thus, divide it wisely. The respondents then had to decide which of the two brand categories will most likely purchase in the five product categories, considering always not exceeding their budget, the crowding in front of the shelf, and their existing preferences. They had to state if the agree with statements they encountered in the beginning of the survey such as “I would be most likely to purchase Oreo cookies (€2,

Budget:€13) over AH cookies (€1, Budget €14)”.

Moreover, in order to measure that our relative budget manipulation was successful we employed Urbany, Dickson, and Kalapurakal’s (1996) budget constraints scale. It is a five-point Likert-type statements to measure the degree to which the consumer perceives having insufficient funds to cover his/hers needs with reported alpha 0.87. In the end of the survey, we asked the participants to indicate if the budget in the survey was tight for the specific purchases (manipulation control variable) and whether their budgeting in real life is always tight as we expected it will affect brand choice (covariate variable).

3.4 Analysis

(23)

21

an effect on the dependent, the likelihood of purchasing a national over a store brand in all product categories. Moreover, the test showed whether there is an interaction between the perceived crowding and the relative budget and thus, whether moderation effects exist.

The ANCOVA (analysis of covariance) is considered an extension of the one-way ANOVA to incorporate "covariate" variables. Like the one-way ANOVA, the ANCOVA is used to determine whether there are any significant differences between the means of two or more independent groups. However, the ANCOVA has the additional benefit of allowing us to "statistically control" for more variables, which are also known as a "confounding variables", which may be negatively affecting our results.

4. Results

In this chapter the results of this study will be presented and explained. First of all, a description of the pre-test will be presented. Secondly, the description of the selected sample characteristics will be provided, followed by the discussion of the main analyses. Furthermore, the effects among the dependent, independent variable and the moderator will be explained. Finally, the summary and the major findings are presented.

4.1 Pre-test

As mentioned above, a pre-test was conducted in order to be certain that the manipulations of the crowding and the budget were correct. After the first analyzing of the completed questionnaires and the valuable feedback of fellow classmates a new corrected questionnaire was conducted. Pre-testing the new version of the questionnaire with the help of 15 participants showed that the estimated means were relative low and similar for the two low crowding scenarios, indicating that the perceptions of the crowding in front of the shelf of the participants were the anticipated. The estimated means for the two high crowding scenarios were relative high and similar as well. That indicates that our manipulation was successful.

4.2 Sample Characteristics

4.2.1 Definitive Number of Participants

(24)

22

Relative Budget) scenario, 39 were assigned to the second scenario (High Perceived crowding, Low Relative Budget), 30 were assigned to the third scenario (Low Perceived crowding, High Relative Budget) and 35 were assigned to the fourth scenario (High Perceived crowding, High Relative Budget). An overview of the groups can be seen in the Table 2 below.

Scenarios-Groups Participants Percent

Low Crowding, Low Budget ₃₄ _24.6

High Crowding, Low Budget ₃₉ _28.3

Low Crowding, High Budget ₃₀ _21.7

High Crowding, High Budget ₃₅ _25.4

Total ₁₃₈ _100.0

Table 2. Number of participants per scenario.

4.2.2 Description of the Sample

A description of the whole sample (N=138) is displayed in Table 3. As it can be derived from the table, 65 males and 73 females (52.9%) were divided almost equally over the scenarios. Moreover, the majority of the participants were between 20-34 years old (94.2%), as only 5 of them were below 20 years old and 3 of them were between 35-49. This is logical as the majority of the participants indicated that they were students (84.1%), with only 14 being employed, 6 self-employed and 2 unemployed. Therefore, 50 of the respondents stated that their annual income was below €10000 (37%) and 58 that it was between €10000-€29999 (43%). Moreover, the majority of participants stated that their country of origin was The Netherlands (30.4%) and Greece (35.5%).

Age: Gender: Occupation:

<20 20-34

35-49 Male Female Student Employed

Self-employed UnSelf-employed Group Low Crowding,

Low Budget 1 33 0 18 16 29 2 2 1 High Crowding, Low Budget 3 35 1 17 22 30 6 2 1 Low Crowding, High Budget 1 28 1 14 16 27 3 0 0 High Crowding, High Budget 0 34 1 16 19 30 3 2 0 Total 5 130 3 65 73 116 14 6 2

(25)

23

In order to check how the sample is distributed over the four scenarios, the frequencies of the variables age, gender, annual income and occupation were calculated. Moreover, an additional Chi-square test was used in order to examine whether there were significant statistical differences of these demographic variables among the groups, in each scenario. A Chi-square test was used in order to investigate whether distributions of the demographic categorical variables differ from one another. The results showed that age, gender, country of origin, occupation and annual household income have no statistical significant differences among the four scenarios (see Table 4). Thus, the four scenarios did not differ with respect to the demographic variables.

Age Gender Occupation Country of Origin Annual Income

Pearson Chi-Square 4.274a .688a 5.622a 5.651a 16.636a

Degrees of Freedom 6 3 9 9 15

p-value 0.640 0.876 0.777 0.774 0.341

Table 4. Chi-square test, p-value and df per variable per group.

4.3 General Preferences Variables

As it was mentioned in the previous chapter, before any manipulation we included questions about the already existing brand preferences of the participants as further in the analysis we aim to control for these variables. Even though the scale was already validated and tested, in order to create our own brand consciousness variable we first checked if we can factorize efficiently the original variables by conducting a factor analysis. The results showed that Kaiser-Meyer-Olkin measure of sampling adequacy =0.702(>0.5), Bartlett’s test of sphericity results rejected the null hypothesis that the variables are uncorrelated and each variable’s component in the analysis was above 0.826(>0.5). Therefore, we run a reliability analysis of the factors found and the reported Cronbach’s alpha was 0.770(>0.7).

(26)

24

Cookies Coffee Milk Butter Sugar

Cronbach's Alpha 0.925 0.930 0.950 0.934 0.934

Table 5. Cronbach’s Alpha for each product category.

As we wanted to test the overall preferences for national and store brands we conducted a factor and reliability analyses with our new five variables for each category, in order to create one factor that will serve as the pre-assessment of their overall preferences. The results of the factor analysis showed a KMO=0.762 and Sig.=0.000, which means that the variables are correlated, significant, and can be factorized. However, investigating the eigenvalues and the

scree plot we encountered the dilemma that SPSS recommended to create two factors (1st

factor: cookies and coffee, 2nd factor: milk, butter and sugar, more details can be seen in Table

1 in the appendix). As our main research question is on the overall likelihood of choosing national over store brands, we chose to create one factor and name it pre-assessment. The results of the reliability test reported that Cronbach’s alpha=0.709, which indicates that our new variable is acceptable and reliable.

4.4 Perceived Crowding Variable

Conducting a descriptive analysis on the perceptions about each adjective included in the scale of the perceived crowding, allowed us to compare the means in each scenario group. The estimated means were approximately 2.5 for the two low crowding scenarios and 5.5 for the two high crowding scenarios. That indicates that our manipulation was successful. Carefully analyzing the other available data it was noted that the range in each group was rather high (see Table 3 in appendix). These results are logical as perceptions can vary among individuals, however, it should be taken into account in the further chapters. Moreover, an investigation to our control variable about the visibility of the brands, supported the statement that our manipulation was successful, as almost half of the respondents indicated that the majority of products were visible and the other half not.

Running the same analyses as before, the results showed that the five variables can be factorized and are reliable and significant to form one variable, which we named perceived crowding. The results are presented in the following table.

KMO df Chi-Square Sig. Alpha Mean N of Items

Perceived Crowding 0.791 10 640.125 0.000 0.927 21.020 5

(27)

25 4.5 Likelihood of Purchasing Variable

We constructed five different variables that measure the likelihood of purchasing a national brand over a store brand for each of the five product categories, after our manipulation. These variables are needed for the analysis on the basis of product category. Moreover, in order to measure the differences in the overall likelihood of purchasing a specific brand category, a factor and a reliability analysis were required. As with the pre-assessment variable, we encountered the dilemma of choosing between two factors that SPSS recommended and one factor that will measure the likelihood of purchasing a national over a store brand in all product categories. According to the eigenvalues, the scree plot and the loadings of the five components, cookies and coffee should become one factor and milk, butter and sugar another (for a more detailed overview see Table 2 in appendix). However, following the same logic as before we chose to create one new variable which we named likelihood of purchasing, though we kept into consideration the limitations for the further chapters. The reported KMO score was 0.664 allowing us to proceed with factorizing the original variables. The reported alpha was 0.506 which is below the recommended value of 0.7. Moreover, if the variable of the likelihood of purchasing national over store brand coffee would be deleted, then the alpha would become 0.634. If we chose to delete the variable about the cookie category then the alpha would increase even more. The cause of these differences may be the strong preferences of the participants for these two product categories. The limitation that our chosen alpha is not ideal will be taken into consideration in the interpretation in the subsequent discussion chapter.

4.6 Delta Variables Before and After Manipulation

(28)

26 4.7 ANCOVA Analysis

As it was mentioned in the previous chapter, we chose to conduct multiple ANCOVA analyses as they allow us to statistically control for confounding variables, which may be negatively affecting our results. Taking all our variables into consideration we concluded that we should not only control for the overall brand consciousness of the participants, which indicates if they prefer national or store brands in general, but also for their specific brand preferences in each product category and their overall brand preference in all product categories (pre-assessment variable). Moreover, we included the variable that measure their real life budget as a control variable as we believe it is a very important determinant on brand choice and will affect our final results. By controlling for all this variables we can clearly investigate the effects of perceived crowding and relative budget on our dependent variable, likelihood of purchasing a national over a store brand. We proceeded to conduct six different ANCOVA analyses as we wanted to draw conclusions not only on the general aspect, but also for each product category. The results on every product category and the overall likelihood of purchasing a national over a store brand are presented in the following chapters.

4.7.1 Cookie Category

The results of the ANCOVA on the likelihood of purchasing a national brand cookie over a store brand, after the manipulation of crowding and relative budget, controlling for brand consciousness, already existing cookie brand preference and real life budget are presented in the table below.

Cookie Category df F-value Sig.

Brand Consciousness 1 1.325 0.252

Existing Cookie Brand Preference 1 107.423 0.000

Real Life Budget 1 5.745 0.018

Crowding 1 5.075 0.026

Relative Budget 1 6.940 0.009

Crowding*Relative Budget 1 0.829 0.364

Table 8. ANCOVA results, cookie category.

The Levene’s test of equality of error variances result rejected the hypothesis that there is homogeneity of variance among the groups (Sig.=0.744). Moreover, as it can be derived from the table the already existing cookie brand preference and real life budget (control variables) of the participants, were significant in the cookie category. Furthermore, the effects of crowding and relative budget on the dependent variable are significant, in contrast to their interaction. Therefore, hypothesis 2 cannot be supported as there is not an interaction effect

Mean

Low Crowding 3.039

High Crowding 2.695

Low Relative Budget 2.664

High Relative Budget 3.070

(29)

27

between the independent variables, in the cookie category. Table 7 shows the means in each crowding and budget scenario. Even though the means are not crucially different, they depict that participants are more likely to purchase store brands cookies in all scenarios. More importantly, it is in the low crowding condition and high relative budget condition that the participants would be more inclined to purchase a national brand cookie thus, hypothesis 1 is partly supported. The direction of the effect of crowding in the category of cookies is the opposite of what it was expected. In contrast the direction of the effect of the relative budget is the same as it was assumed. The people assigned to the high condition were more prone to buy a more expensive brand than the people assigned to the low condition.

4.7.2 Coffee Category

In Table 9 the results of the ANCOVA on the likelihood of purchasing a national brand coffee over a store brand, after the manipulation of crowding and relative budget, controlling for the already mentioned variables are presented.

Coffee Category df F-value Sig.

Existing Coffee Brand Preference 1 178.396 0.000

Crowding 1 0.150 0.902

Table 9. ANCOVA results, coffee category.

(30)

28 4.7.3 Milk Category

A similar ANCOVA was conducted on the likelihood of purchasing a national milk brand over a store brand.

Milk Category df F-value Sig.

Existing Milk Brand Preference 1 175.260 0.000

Crowding 1 1.472 0.227

Table 10. ANCOVA results, milk category.

The Levene’s test of equality of error variances result supported the hypothesis that there is homogeneity of variance among the groups (Sig.=0.007). In Table 10 it can be seen that only the control variable for the existing preference on milk brand has a significant effect hence, both hypotheses were not supported.

4.7.4 Butter Category

The SPSS output from conducting an ANCOVA on the likelihood of purchasing a national butter brand over a store brand showed that the underlying assumption of homogeneity of variance in all groups is rejected (p-value=0.349). As it can be derived from the table below, in the butter category the existing brand preference and real life budget are significant. Moreover, it is the only product category in which brand consciousness is marginally significant (for the 10% of the participants). Similar to the milk category, both hypotheses are not supported.

Butter Category df F-value Sig.

Existing Butter Brand Preference 1 72.378 0.000

Crowding 1 0.057 0.811

(31)

29 4.7.4 Sugar Category

Similarly to the previous analyses, the results (see Table 12) on sugar category showed that only the two control variables existing sugar brand preference and real life budget are significant. Hence, both hypotheses are not supported in this product category.

Sugar Category df F-value Sig.

Existing Sugar Brand Preference 1 61.480 0.000

Crowding 1 0.031 0.861

Table 12. ANCOVA results, sugar category.

4.7.5 All Categories

Conducting the same analysis on the overall likelihood of purchasing a national over a store brand showed that the control variable of already existing preferences (pre-assessment) is highly significant. The results are presented in the next table.

All Categories df F-value Sig.

Existing Overall Brand Preference 1 124.391 0.000

Crowding 1 1.170 0.281

Table 13. ANCOVA results, all categories.

(32)

30 4.8 Summary

We conducted several ANCOVAs, including three control variables in order to observe the changes of the likelihood of purchasing a national over a store brand, due to different levels of crowding in front of the shelf. After analyzing all the results, it is clear that different levels of crowding do not have an effect on brand choice in the majority of the product categories. The only category that showed a different result is the cookie category. That was expected after reviewing the skewness of the answers for the dependent variable for each product category. According to the histograms (see Figures 1-6 in the appendix), the participants in every product category except cookie category, answered almost similarly. Expressing almost unanimously their strong brand preference in each product category might have compromised the significance of crowding on brand choice.

We expected that high perceived crowding in front of the shelf will increase the likelihood of purchasing a national brand product over a store brand product, due to the lower visibility. Moreover, we made the assumption that a lower relative budget in contrast to a higher relative budget, in high perceived crowding in front of the shelf, will weaken this effect. Looking more closely to the means for each category in every condition, it is clear that the direction of the effects are as we expected, however, the results are not significant (see Table 4 in the appendix). A summary of our results is presented on the table below.

Hypothesized Relationship Results

Cookie Category

H1: High Crowding + Partly Supported

H2: Low Budget - Not Supported

Coffee Category

H1: High Crowding + Not Supported

Milk Category

Butter Category

Sugar Category

Overall Category

(33)

31 5. Discussion

Sales of store brands have been growing rapidly in recent years as store brands are appearing in an ever-increasing number of categories and their acceptance by consumers is unquestionable (Miquel, Caplliure and Aldas-Manzano 2002). Retailers are promoting their store brands in order to increase store loyalty, chain profitability, more control over shelf space, which can be translated into more bargaining power over manufacturers (Ailawadi and Keller 2004; Amrouche and Zaccour 2007; Morton and Zettelmeyer 2004; Richardson, Jai, and Dick 1996b; Semeijn, Riel and Ambrosini 2004). The main reason for their popularity and growth is their price advantage over the national brands (Batra and Sinha 2000). Prior research has focused on the effects of retail crowding on shopper behavior. While at the theoretical level a diverse number of reasons have been proposed to try and configure the reasons why consumers prefer national brands to store brands, none has focused on the effects of perceived crowding in front of the shelf on brand choice. As it was mentioned in previous chapters, the effort associated with looking at middle and top shelves is lower than the effort associated with looking at lower shelves. Moreover, it is a common practice of retailers to place the leading national brands on the middle to top shelves and store brands on the bottom shelves. Crowding in front of the shelf can decrease the visibility of the majority of the brands on the shelf, especially the ones located on the bottom. This study has hypothesized, and tested the potential effect of perceived crowding and level of relative budget as possible explanatory variables about brand choice. We proposed that shoppers would most likely choose a brand which is located on top and near the middle shelves, as it requires less effort, when the crowding in front of the shelf was perceived as high. Moreover, we expected that a low level of relative budget would weaken this outcome as the shopper would have to consider that his budget would not suffice for purchasing a national brand product.

(34)

32

between the alternatives, make them a low perceived risk purchase decision (Miquel, Caplliure and Aldas-Manzano 2002). That can be also one of the reasons why crowding was not significant in the other product categories. Since these products are some of the most frequently purchased products, people have already established their preferences and would probably not even comprehend the crowding in front of the shelf, choosing their usually purchased store brand. Hoyer (1984) in his study suggested that choice is the outcome of multiple trials and purchases. At the end of these trials, consumers tend to develop a set of simple choice tactics which allow them to decide in a quick and satisfactory manner. We suggest that cookies have many different flavors and varieties that people tend to consider the alternatives store brands that can fulfill the same need.

Moreover, we assumed that people would be more likely to purchase a national cookie brand over a store in the high crowding condition. The results showed that people would be more likely to purchase a store cookie brand in both conditions. However, it should be noted that in the high crowding condition people were more inclined to purchase a store cookie brand than in the low crowding condition. That can be explained on the basis that even though cookies are a low involvement product category, with low perceived risk, people tend to display a variety seeking behavior, altering types and flavors. Since it is a highly hedonic product people who will encounter high crowding in front of the shelf will not ponder and proceed to buy the brand they usually purchase. However, if they do not encounter crowding in front of the shelf they might consider their alternatives and opt for a different type of cookie, depending on their current mood.

(35)

33 5.1 Managerial Implications

All these results can prove to be crucial for brand managers as they can be used to alter consumer behavior in order to evoke favorable outcomes. Levav and Zhu (2009) in their research they suggested that by spatially confining people, either by crowding them with others or by manipulating architectural cues, leads them to feel restricted and consequently to exhibit reactance in the form of making more unique choices and testing new brands and products. In their experiments they altered the aisle width to manipulate spatial confinement. Following the same logic, we suggest that in the cookie category retailers could alter the width of the aisles in order to evoke more crowding in front of the shelves. According to our results this will increase the likelihood of purchasing a store cookie brand product and as a consequence will increase the retailers’ sales in that category. Another strategy would be to relocate their store cookie brands in order to gain higher visibility. According to the results shoppers are more inclined to purchase store cookie brands regardless the level of crowding in front of the shelf. By relocating these brands to higher shelves, preferably near the middle and close to the top, where visibility is higher, the effort of searching these brands will decrease. Therefore, the likelihood of purchasing a store cookie brand will increase which can lead to the increase of their sales.

(36)

34

its expiration that will urge the consumer to go to the cookie aisle, increasing the crowding in front of that shelves and thus, increasing store cookie brand sales.

Our results also indicated that consumers’ perception of their available relative budget can affect their likelihood of purchasing a national over a store brand. Retailers can employ various marketing in-store tools in order to influence consumers’ perceptions and consequently increase their sales. More precisely, they can provide consumers with scanners and devices that can scan and calculate the expenses of the products placed in their shopping cart. Consumers would be able to see instantly the reduction of their available budget resulting in influencing their budget perceptions. According to our results a lower relative budget perception could lead to a stronger inclination of purchasing the store brand alternatives. Retailers can also use discounts as a tool to influence shoppers’ budget perceptions. The shopper can place a product in his cart and be presented with a discount on his scanner. By providing him with an unexpected discount, his relative budget perception will be altered, as the consumer would consider that now he has more available budget than he originally calculated in his mental account. Moreover, he can be presented with an additional promotion to another store brand. This can lead to the increase of the store brand sales in another product category.

5.2 Limitations and Further Research

(37)

35

proceeded to analyze with reported Cronbach’s alphas below the recommended value, which could have confounded our results.

Moreover, all the product categories were low involvement, therefore it would be recommendable investigating whether crowding have an effect on brand choice in high involvement product categories, in which consumers are more inclined to engage into cognitive thinking about brand choice. As Richardson, Jain and Dick (1996a, b) suggested the decision about purchasing a store over a national brand, is based on certain consumer perceptions of the particular product category. They argued that the level of perceived quality, perceived risk, and perceived value in terms of money, as well as the degree of consumer involvement about the product category can affect their purchase decision. According to the literature on perceived risk, shoppers will choose a national over a store brand, if the level of perceived risk is perceived as high (Narasimhan and Wilcox 1998). Batra and Sinha (2000) argued that purchases of certain product categories as for instance baby foods are perceived as more risky than the purchases of others, as for instance sugar or cookies. Therefore, it would be rather insightful to investigate the effects of crowding in front of the shelf in the high involvement product categories, where consumers are more knowledgeable and more conscious about the brands. Another area of further investigation concerns the role of store type e.g. discount stores. Consumers visiting discount stores are usually more deal-prone and price-conscious and therefore, more prone to purchase the cheaper store brands alternatives. Investigating the effect of crowding on brand choice in different store types can fill many gaps in the existing literature.