MSc Business Administration
Master Thesis
The relationship between category
characteristics and both store loyalty and
brand loyalty
Study: MSc Business Administration – Marketing Track Institution: University of Amsterdam
Student: Wouter Raaijmakers Student number: 10881204 Thesis supervisor: dr. J.Y. Guyt Date: 23-06-2016, final draft
2 Acknowledgements
The Master thesis is my last experience before leaving the University of Amsterdam, as it is the final assignment for completing the master Business Administration. The thesis is written in the field of retail marketing and therefore a perfect fit with the chosen track Marketing. It examines the relationship between supermarket category characteristics and their influence on both store loyalty and brand loyalty.
I would like to dedicate this section to extend my gratitude to all people who were involved in my Master thesis. First and foremost, I would like to thank my supervisor, Jonne Guyt. I could not have accomplished this without his expertise, feedback and help throughout the last couple of months. And of course, my gratitude also extends to my friends and family and especially to my girlfriend Floor Geeven, who has been my biggest support and for her understanding that my Master thesis was top priority. Naturally, I would like to thank
everyone else who supported me throughout this process and who have contributed by filling in my survey.
Wouter Raaijmakers Breda, June 2016
Statement of Originality
This document is written by Wouter Raaijmakers who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3 Abstract
This thesis examines the direct effect of supermarket category characteristics and their influence on both store loyalty and brand loyalty. Additionally, it investigates the ratio between store loyalty and brand loyalty. The first part of the quantitative study examines the direct effects of five category characteristics and six out of ten hypotheses are supported. First, promotion frequency, purchase frequency, stage of product life cycle and mean category price are all characteristics that show a significant positive relationship with store loyalty. Second, purchase frequency and perceived quality variance both have a significant positive relationship with brand loyalty. In contrast with previous findings, the promotion frequency also has a positive effect on brand loyalty. In the second part, the ratios between store loyalty and brand loyalty are measured. After calculating the ratios, a multiple regression was
performed for each category to examine if there are significant variables explaining the outcome variable (ratio). Only two out of eight models are significant, but those categories both show a significant decrease for purchase frequency. This means that an increase in purchase frequency will lead to a significant higher increase of brand loyalty than store loyalty. The same effect occurs in another category for the category’s quality variance and education. An increase in education level will lead to a significantly higher increase in brand loyalty than store loyalty. Overall, purchase frequency and promotion frequency both have implications for store and brand loyalty. Although the effect of promotion frequency on store loyalty is stronger and purchase frequency plays a greater role for brand loyalty. As a final point, stage of product life cycle influences store loyalty positively and the same holds for quality variance on brand loyalty. These findings shed light for both the retailer and the manufacturer on which category characteristics they should focus and will give an indication of the source of sales displacement in the event of a promotion.
4 Contents Acknowledgements ... 2 Statement of Originality ... 2 Abstract ... 3 Contents ... 4 1 Introduction ... 5 1.1 Background ... 5 1.2 Research question ... 7 1.3 Scientific relevance ... 7 1.4 Managerial contribution ... 8 1.5 Structure ... 8 2 Theoretical framework ... 10
2.1 Attitudinal loyalty and behavioural loyalty ... 10
Sales promotions ... 11
2.2 Category characteristics ... 14
Purchase frequency ... 14
Stage of product life cycle ... 15
Perceived category quality variance ... 16
Mean category price ... 17
3 Method ... 18 3.1 Conceptual Framework ... 19 3.2 Sample ... 21 3.3 Measures ... 21 4 Results ... 23 4.1 Analytical strategy ... 23 4.2 Results I ... 24 4.3 Results II... 29 5 Discussion ... 33 6 Conclusion ... 41 6.1 General conclusion ... 41 References ... 44 Appendices ... 48
I Pre-test and questionnaire ... 48
II – Regression ratio for all categories ... 51
III Histogram for ratio and natural logarithm (ln). ... 52
5 1 Introduction
1.1 Background
Supermarkets and their consumer packaged goods (CPG) have been a well-reviewed topic over the past decades. Particularly, how retailers can increase profitability by enhancing consumers’ loyalty (e.g. Kumar and Shah 2004; Helgesen 2006). There are multiple factors that have an impact on how to achieve these more loyal customers and higher profits (Sirohi et al 1998). One of those factors is the growing number of sales promotions, despite the fact that previous research suggests that there is only a minimal effect on profitability and sales promotions can attract the wrong customers (Reichheld and Teal 2001). Both retailers and manufacturers continue to outperform competition by offering attractive discounts. Although they do not always have the same motives, the retailer influences the decisions of the
manufacturer and vice versa (Ailawadi et al. 2009). They both need each other to achieve their own goals. For example, one of the main reasons for the retailer to have sales
promotions is to attract current and new shoppers towards their store. And in contrast with the manufacturer, the retailer does not depend on a small amount of brands. However, the retailer still needs the manufacturer to shape highly attractive sales promotions for the consumers. These days, the relationship between retailers and manufacturers is complicated, which leads to tension between the two. To identify and understand how consumers shop across categories between stores and brands will positively increase the insights of both retailer and
manufacturer. This contributes to each of their objectives.
Previous literature typically focused on either the retailer or the manufacturer and have conducted multiple studies on store loyalty (Ailawadi et al. 2008; González-Benito and Martos-Partal 2012) and brand loyalty (e.g. Chaudhuri and Holbrook 2001) independently. Little research combined brand loyalty and store loyalty or whether consumers choose the store or the brand first. Does this have an effect on each product category or do they differ?
6
For example, do consumers in the beer category choose the store or the brand first and how is this different with home care categories? There is some research (Briesch et al. 2009;
Gijsbrechts et al. 2008) about the differences between what characteristics make different categories create and increase store loyalty or brand loyalty. However, they did not test which characteristics of categories are more influential and which characteristics are important for retailers or manufacturers. Therefore the primary objective of this research is to identify why consumers choose the store first for specific categories and why the brand comes first in other categories. Although the retailer and the manufacturer have more or less shared goals, this research will shed light on how for example sales promotions per category can be optimized. Both the retailer and manufacturer can benefit from the outcomes of the research in their unending quest to attract and create loyal customers.
Sales promotions can influence both brand and store switching and therefore have an impact on consumer loyalty. However, most research has been done on either brand loyalty (e.g. Chaudhuri and Holbrook 2001) or store loyalty (Ailawadi et al. 2008; González-Benito and Martos-Partal 2012). The latter suggested that further research with attitudinal measures could provide new insights. Where most research about retailing is conducted primarily with
scanner data, this will only reveal the behavioural loyalty of consumers and not the reason why consumers (intent to) buy at different stores or different brands. Consequently, this research will only focus on the attitudinal loyalty of consumers in the consumer packaged goods (CPG) industry. In particular, CPG categories consist of several brands and those differ between stores. Does the consumer first decide which store to visit or does the consumer make a comprehensive consideration where to find the best bargains for their favourite brands? Are there specific characteristics per category that influence this decision? Does this for example depends on the category size of the supermarket, the average category price, the purchase frequency, the familiarity of the products, stage of the product life cycle, risk
7
reduction, time restrictions or the category sales promotions (in/out store)? This study analyses store and brand loyalty by looking at aforementioned category characteristics.
1.2 Research question
Based on the revealed literature and the identified research gap, the following research question is formulated:
What characteristics of supermarket categories influence consumers to be more brand or more store loyal?
In order to answer this question, store categories need to be narrowed down. Time restrictions make it impossible to conduct a thorough analysis about all categories for all stores. Therefore a selection will be made, but in order to find the most significant differences at least one category of home care, food and beverages needs to be represented.
1.3 Scientific relevance
The aforementioned articles (Chaudhuri and Holbrook 2001; Ailawadi et al. 2008; González-Benito and Martos-Partal 2012) are all part of the thoroughly studied CPG industry, however they only focused on a specific subject and to the best of the authors’ knowledge did not combined store and brand loyalty. What drives consumers to switch between stores and why are specific categories worth the effort of changing consumers’ shopping route? The
expectation is that certain categories have an overlap of characteristics predicting store or brand loyalty. Besides, previous research is mostly conducted with scanner data to examine behavioural loyalty, but little research is performed for attitudinal loyalty. This will lead to new insights in the literature and can be very valuable for further research. The final contribution would be the broader knowledge insight on attitudinal loyalty in combination with category characteristics.
8 1.4 Managerial contribution
First of all, the outcomes of this study are relevant for all retailers. By defining the right characteristics of categories, they are able to create an uplift in sales and eventually attract more loyal customers. This would give them more insight in how to optimize their categories and sales promotions. It will also provide them with useful information in which categories their store brands have a higher chance of succeeding. Secondly, these results are also important for brand managers, who deliver their products to the specific retail chains. In the end, managers can benefit from the knowledge which characteristics make their category more brand or store loyal. They can adjust their products/services and promo-mechanisms accordingly to help increase the sales and create more loyal customers. Finally, knowing more about brand and store loyalty will also give an indication of the source of sales displacement in the event of a promotion. For example, brands want consumers to switch to their brand and stores want store-switching.
1.5 Structure
This section will show the structure of this research. Initially, the introduction is presented to introduce the subject and to illustrate the necessity of this study. The next chapter discusses the previous research and the theoretical framework. This chapter consists of a critical review on the key concepts in the field. Finally, the chapter provides this study’s hypotheses. Chapter three introduces the method and gives a detailed description of measurements, data collection and the sample that is used. Chapter four is the follow-up and shows how the data is analysed with the aforementioned methods of chapter three. Furthermore, this chapter will present all the empirical findings of the tested hypotheses. The fifth chapter discusses the main findings of the study and refers them to the theoretical framework. This chapter will also provide an answer to the research question and stress the contributions of this study. The aim of this
9
chapter is to show the theoretical relevance and new insights for both retailers and manufactures. Finally, the last chapter provides the limitations of this study and gives suggestions for future research.
10 2 Theoretical framework
This chapter provides a critical review of the most relevant findings from previous literature and their core concepts. Besides, it will introduce the hypotheses of this study. The chapter starts with explaining the difference between attitudinal loyalty and behavioural loyalty. Second, it explains how sales promotions can influence the consumers’ decision making for both store choice and brand choice. Next, different category characteristics and their influence on store and brand loyalty are discussed. Finally, the chapter visualizes the hypotheses that have been tested with a theoretical research model.
2.1 Attitudinal loyalty and behavioural loyalty
This chapter will explain the different forms of loyalty to fully understand what is being examined in this study. Previous research makes a distinction between real purchase behavioural loyalty and attitudinal loyalty. Chaudhuri and Holbrook (2001) describe the difference between purchase and attitudinal loyalty: “Purchase loyalty is defined as the
willingness of the average consumer to repurchase the brand. Attitudinal loyalty is the level of commitment of the average consumer toward the brand” (p. 82). Behavioural loyalty occurs when consumers repetitively visit the same store or purchase a specific brand (Quester & Lim 2003). Particularly, when they do not go anywhere else or purchase a different brand in the same category. The first one specifies behavioural loyalty towards a store and the latter behavioural loyalty towards the brand. Moreover, Lin and Gijsbrechts (2014) argue that consumers’ store choice and loyalty also depends on the outlet instead of the chain.
Consumers will keep purchasing at this particular store, even after a takeover. Nonetheless, this effect depends on if the new outlet is a more upscale tier, instead of having the same or even a lower positioning. In the end, high-end retailers benefit more from a takeover, because they have a higher chance of retaining consumers.
11
Additionally, according to Thompkins and Tam (2013) behavioural loyalty can be divided into two mechanisms: attitudinal loyalty and habit. Both mechanisms will lead to dissimilar purchase patterns. Attitudinal loyalty is at first a strong intention to buy the brand and in time leads to repurchase behaviour. Habit on the other hand is defined as automatically associated behaviour, when one or more stable contextual cues occur. These cues can be time, location, social setting and preceding and ensuing events (Wood and Neal 2009). Their results show that there is only a weak correlation between the two, which suggests that it can be seen as two individual drivers.
Sales promotions
A broad range of research has been conducted about sales promotions. This has been an interesting research topic in the consumer packaged goods industry, since retail chains came to the conclusion that sales promotions can trigger consumers to visit their stores. Although they have more than just one intended outcome. Sales promotions are frequently used by marketers to increase product trial. The products are only discounted in a certain period and can have limited availability (Kotler 1988; Webster 1971). Sales promotions influence where consumers shop and consumers are willing to adjust their shopping routes for better prices (Ailawadi et al. 2001). This inevitably has an effect on the marketing budget for brands. Brands have focused on how to improve their promotions, so they can increase their sales. The confidence of marketers that promotions are essential tool to create an uplift in sales is mainly because they want to keep up with the competition (Ken Peattie, Sue Peattie and Emafo 1997). Over the years, this belief in promotions ensured the tremendous increase in marketing budget allocated to promotions (Ailawadi et al. 2001).
Promotions in numerous consumer packaged goods categories were mostly single items or brand discounts and much literature focused mainly on this aspect. This would lead
12
come up with new ideas to create more store loyalty. That is one of the reasons why these days not only single item discounts but across category discounts are introduced.
According to DelVecchio, Henard and Freling (2006) sales promotions do not have a
significant effect on brand preferences after the sales promotion. Although this depends on the characteristics of the product. These findings suggest that sales promotions do not have an effect on brand loyalty, but only increase sales for a short period of time. If sales promotions do not have a significant positive effect on brand loyalty (DelVecchio et al. 2006), why do marketers invest between 58% and 60% of their marketing spending on it (Low and Mohr 2000; Ailawadi et al. 2009)? Moreover, Allender and Richards (2012) examined the effect of sales promotions in two different categories: ice cream and carbonated soft drinks. One important conclusion that has been drawn from their research, is that sales promotions have more impact on brands with weak loyalty (weaker brands). Despite the fact that they are less often discounted. In contrast with stronger brands, consumers are easily influenced by sales promotions to switch between weaker brands. However these findings are not necessarily generalizable to other product categories, the authors believe to find the same results in different categories. Further research from Ataman et al. (2010) examined both the short-term effect of sales promotions and the long-term effect. They investigated scanner data for 25 product categories and 70 brands in the four largest retailers in France. One of their findings is the short-term boost in sales after promotions. Their second finding is the contrary effect of promotions on long-term sales. This effect can be explained by consumers who only purchase discounted products and never pay the base price or consumers who shop across different stores.
Consumers switching from one brand to another is generally beneficial for one of the two manufacturers and not for the retailer. The single exception is the increase in margin, which can be a trigger for a retailer to promote brands with higher margins. According to van Heerde
13
et al. (2004) higher margins are not the only reason for retailers to promote particular brands over others. The increase in product sales for the discounted brands will lead to a decrease for other brands in the same category, however sales promotions can also lead to category
expansion. In contrast with the aforementioned article of Ailawadi et al. (2001), consumers do not necessarily switch stores for the highest discounts, but are willing to purchase more across categories. This would indicate that consumers are more loyal to the stores than to the brands and therefore sales promotions are extremely beneficial for the retailers to attract more consumers. According to the article of Gijsbrechts et al. (2008), empirical evidence that sales promotions do increase store loyalty is limited (Rhee & Bell 2002; Srinivasan, Pauwels, Hanssens, & Dekimpe 2004). This indicates that there are several reasons for consumers to switch between stores. One of their findings is that consumers prefer at least one product category at each store. They argue that consumers are loyal towards the stores’ specific category. Even though previous research is not consistent if sales promotions have a positive effect on store loyalty or brand loyalty, there is less research indicating that sales promotions do not have an effect. Besides, retailers continue to persuade consumers with sales promotions to purchase in their store, so the frequency of promotions does matter to them. On the other hand, sales promotions induce choice (DelVecchio et al. 2006) and this is a tool to encourage product trial (Kotler 1988). However, after the sales promotion this benefit no longer occurs. Both aforementioned effects will be examined in this study. Therefore the first hypotheses regarding sales promotions are formulated as follows:
H1a. The promotion frequency in a category has a direct positive relationship with store loyalty.
H1b. The promotion frequency in a category has a direct negative relationship with brand loyalty.
14 2.2 Category characteristics
Although little is known about characteristics of supermarket categories and their influence on consumer behaviour, there are some articles indicating the importance and differences of these characteristics. For example, the category purchase frequency influence which retailer to visit. Consumers are more likely to shop at one specific retailer for high frequency purchase products (Gijsbrechts et al. 2008).
Purchase frequency
Liu (2007) examined in her study about loyalty programs the influence of heavy buyers versus light buyers. Heavy buyers are consumers who purchase many products in the same store and light buyers usually shop their products at different stores. Even though she
expected heavy buyers to claim their loyalty awards more often than light buyers, it were the less heavy buyers that became more loyal due to loyalty programs. The loyalty program strengthened the relationship between the firm and light and moderate buyers. These groups became more loyal towards this store. Her research argues the importance of purchase frequency, which is in line with other articles that also suggest that purchase frequency is an important predictor of a consumers’ relationship with a retailer (Schmittlein, Morrison, and Colombo 1987; Venkatesan and Kumar 2003). The more consumers tend to purchase from a supermarket category the more they probably will be acquainted with a store. On the other hand, purchasing frequently from a category also leads to less loyalty towards a specific brand (Dawes et al. 2015). Overall, based on the previous literature, purchase frequency has a stronger effect on store loyalty than on brand loyalty. This leads to the following formulated hypotheses:
H2a. The average purchase frequency of a supermarket category has a direct positive relationship with store loyalty.
15
H2b. The average purchase frequency of a supermarket category has a direct relationship with brand loyalty.
Stage of product life cycle
Bijmolt et al. (2005) state that also the product life cycle of a product matters and that product differentiation can lead to more loyal customers in the later stages of the product life cycle. This suggests that categories that are relatively new are easier to differentiate from others. And in the later stages of the product life cycle products differentiation is key to distinguish products and categories. This contradicts with the findings of Johnson et al. (2006). Their research shows stronger ties between consumers and firms as the category matures. The consumers’ experience grows, which lead to stronger relationships and attitudes towards brands. One of their limitations is that these findings are only examined in the service industry and should be replicated into different contexts to discover if the results are generalizable. For retail products it can be completely different. In fact, relatively new categories with many innovations may perhaps lead to higher loyalty for both brands and stores. If a retailer is one of the first to introduce a new category consumers are willing to try this store if they want to purchase these new innovations. If consumers positively perceive these innovations they will automatically prefer these brands, as there are not many competitors. On the other hand, new and young categories carry more risk. To avoid this effect, brand affect and brand trust are essential. Risk aversion will eventually lead to brand loyalty (Matzler et al. 2008).
The following hypotheses are related to the stage of product life cycle:
H3a. Products in early stages of the product life time cycle have a direct positive relationship with store loyalty.
H3b. Products in early stages of the product life time cycle have a direct positive relationship with brand loyalty.
16
Perceived category quality variance
The aforementioned articles of van Heerde et al. (2004) and Ailawadi et al. (2001) give contradicting arguments for sales promotions being valuable to increase store or brand loyalty. Nonetheless, there are more characteristics of categories that influence the shopping route of consumers. For example, Briesch et al. (2009) show the importance of category characteristics. According to them, retail assortments are after convenient locations and low prices the most significant reason for the consumers’ store choice. Not only larger assortments would attract more consumers, but also the number of items inside the category influence the store choice. Furthermore, the presence of a consumer’s favourite brand also affects their store choice. If there are more favourite brands available, this will increase the likelihood of visiting that store. Finally, the broader the category assortment and the higher the quality variance will increase the number of satisfied consumers (DelVecchio, 2001). By offering different products with different quality the possibility that consumers find their favourite products will grow. On the other hand, if the quality variance is high, consumers are more able to differentiate products and this may lead to higher brand loyalty as well. According to Szymanowski, and Gijsbrechts (2013) brand quality and the need for variation differs per category. For high-involved products and categories with high performance risk, consumers tend to purchase different products to determine the quality and become more loyal to one brand after evaluating different brands. This would assume that quality variance is more important for the brand than for the retailer, which leads to the following hypotheses about the perceived quality variance:
H4a. The category quality variance has a direct positive relationship with store loyalty. H4b. The category quality variance has a direct positive relationship with brand loyalty.
17
Mean category price
The two factors that are essential to describe loyalty are favourable attitude and repeated purchases (Dick and Basu, 1994; Bandyopadhyay and Martell, 2007). The latter research discusses the difference between behavioural loyalty and attitudinal loyalty. To really
understand these differences, they were discussed briefly in the previous chapter. In line with the research of Chaudhuri and Holbrook (2001), their research indicates that behavioural loyalty leads to an increase in market share and attitudinal loyalty to relatively higher brands’ prices. This research will shed light on if the category mean price also plays a role in both brand loyalty and store loyalty. It is expected that consumers will switch their shopping routes more easily for categories with higher mean prices than for categories with low mean prices. Furthermore, consumers will probably be more critical for categories with high mean prices. The trial percentage will be lower for those products, because consumers do not want to waste money. As a result, also the brand loyalty will be higher. So higher category mean prices are estimated to have a positive effect on brand loyalty. Therefore the following hypotheses are formulated:
H5a. The mean category price has a direct positive relationship with store loyalty. H5b. The mean category price has a direct positive relationship with brand loyalty.
18 3 Method
The aim of this study is to compare characteristics for different categories and determine which one will lead consumers to be more brand loyal or more store loyal. To discover what characteristics differ between categories a quantitative research will be conducted. By using quantitative data a large sample of consumers can be targeted to test which characteristics influence brand or store loyalty. This has a positive effect to the extent to which the
conclusions are generalizable (Lewis and Sanders 2012). The appropriate research design will be a survey. This will give the possibility to collect data about attitudinal loyalty from a large number, is cost-effective and the external validity is relatively high. To test which
characteristics and which categories to choose a list will be provided and a pre-test will be conducted to be confident that the best possible categories and indicators are tested. Another reason to pre-test the categories is to ensure categories are included for both store loyal consumers and brand loyal consumers. Finally, in the pre-test the questions are already tested on reliability and validity.
The selection of the characteristics are based on previous research (Gijsbrechts et al. 2008; Bijmolt et al. 2005; Briesch et al. 2009), who all suggest different characteristics effect consumers’ shopping behaviour. The research of Gijsbrechts et al. (2008) initially discusses the importance of purchase frequency. Subsequently, Briesch et al. (2009) stress that prices are significant influencers on consumers’ shopping route and in the research of Bijmolt et al. (2005) the product life cycle has an effect on consumers’ brand loyalty. Additionally, Heerde et al. (2004) and Ailawadi et al. (2001) examined the promotion effectiveness on store and brand loyalty. Finally, DelVecchio (2001) argues that also the quality variance of a category matters and can influence consumers’ choice. Since none of the abovementioned research measured all these characteristics together and if it effects store or brand loyalty, this study will give insights in the effect of all those category characteristics. Although convenience
19
(distance to store) has a large impact on store loyalty, this will not be measured in this study, as it is not a characteristic of a category. The different category characteristics will either be determined for each category beforehand or asked during the pre-test and survey. For
example, the mean price is determined in advance, since category prices are already available. The same holds for the number of sales promotions, which will be calculated by actual retail data. Next, the average purchase frequency can be measured in advance. Then again this may differ among consumers and that is why the pre-test and survey also contain questions about consumers’ perceptions of purchase frequency.
Because almost every consumer shops at the supermarket, the sample can be broad. Nonetheless, preferable are differences in age, gender and education. Most of the research questions will be questioned on a 7-point Likert scale (1= strongly disagree; 7 = strongly agree) to discover if consumers led specific category characteristics drive them towards the store or towards the brand. Most of the surveys will be collected online and if necessary the respondents can also be reached anonymously outside or inside different supermarkets. Finally, the study will consist of two parts. In the first part all hypotheses will be tested and their direct relationship with either store loyalty or brand loyalty. The second part of the study will determine which category characteristics are more important for store loyalty or brand loyalty. Therefore, the ratio between store and brand loyalty will be computed.
3.1 Conceptual Framework
The conceptual framework visualizes the different relationships between the variables and both store loyalty and brand loyalty. Based on the independent variables and dependent variables (store loyalty and brand loyalty) the aforementioned hypothesis will be tested.
20
Figure 1. The relationship of category characteristics with store loyalty and brand loyalty.
To test these hypotheses different categories will be presented to the respondents. These categories differ in terms of mean category price, average purchase frequency and promotion frequency. A pre-test is used to indicate if the categories also show differences in perceived category quality variance and stage of product life cycle. In order to diminish respondents biasing their answers due to order effects, the questions are asked randomly. And the study does not present all categories for each respondent. In the pre-test the following eight categories are tested randomly:
1. Coffee; 2. Beer; 3. Liquid detergents; 4. Chips; 5. Ketchup; 6. Chocolate bars; 7. Mayonnaise; 8. Soda.
Some categories are chosen because they may contain of more loyal brands. For instance, the categories ketchup and mayonnaise are both typical examples of categories that contain loyal brands for Dutch consumers. Beer and liquid detergents on the other hand may be more promotion driven and consumers select different stores because their favourite brand is discounted somewhere else. Soda, coffee and chips are three frequently purchased categories and chocolate bars are more purchased as an impulse. All these categories have different
Promotion frequency in category Average purchase frequency Store Loyalty Brand Loyalty Perceived category quality variance Stage of product life cycle
21
brands and therefore have higher chances of different outcomes towards store and brand loyalty.
3.2 Sample
This study focuses first and foremost on Dutch consumers. As the population of interest is large and the sampling frame unidentified, a non-probability convenience sampling method is permitted to reach respondents (Saunders et al., 2012). Respondents will be requested either online or offline to fill in a Qualtrics designed survey. During a time period of three weeks as many respondents as possible will be reached. Nevertheless, at least 200 respondents are needed for analysing the data. Despite of difficulties calculating the response rate, the survey will be distributed via Facebook in order to reach a suitable number. Since the respondents reached via Facebook are expected to be between 20 and 35 years old, personal e-mail and offline methods are used to approach different groups.
3.3 Measures
The survey includes individual demographics, such as age (ratio variable), gender (nominal variable) and highest completed education (ordinal variable). To measure attitudinal store and brand loyalty, multiple validated researches will be used and adapted to the survey. To
measure attitudinal store loyalty previous research (e.g. Bloemer and Ruyter 1998; Osman 1993; Zentes et al. 2008) stresses intention to recommend and brand commitment as two important indicators. To investigate both brand loyalty and store loyalty, the 7-point Likert scale of Yi and Jeon (2003), which is also used by Liu-Thompkins and Tam (2013) in combination with the overview of Rundle-Thiele (2005) will be used (see Appendix I). All Cronbach’s alphas are between .788 and .88 (E.g. "I have a strong preference for this store,"). Respondents are requested to answer on a scale from “strongly disagree” to “strongly agree”
22
how they feel about a positively formulated statement. In addition, both statements from Chadhuri and Holbrook (2001) will be added to the survey, to ensure brand loyalty is
measured sufficiently. Those statements are: “I would be willing to pay a higher price for this brand over other brands” and “I am committed to this brand”, with Cronbach’s alphas of .90 and .83. The adapted version to measure the category characteristics (see Appendix I) will be translated into Dutch. Besides these measurement scales, the questions for purchase frequency and stage of product life cycle will be added. These scales will be tested in the pre-test if their Cronbach’s Alphas will meet the threshold of .70. Furthermore, the pre-test is performed to see if respondents interpret the questions as intended. The pre-test was spread before the final survey among a small group (n=40) and all aforementioned categories were tested. After the pre-test a few adjustments were made to reduce respondents’ bias. Prior to the questions of the pre-test and the final survey, confidentiality and anonymity are ensured. Besides the principle of categories is explained briefly, to ensure respondents understand the definition of a supermarket category.
23 4 Results
The results of this study consist of two parts. In order to understand the meaning of the category characteristics, both their relationship with either store or brand loyalty and the relationship store versus brand loyalty is examined. First, the following section will present all variables and their influence on both store loyalty and brand loyalty individually. The
hypotheses will be tested in this first part of the study. In the second part of the study the ratio between store and brand loyalty will show which category characteristics are more important for one of the two outcome variables.
4.1 Analytical strategy
Before analysing the dataset a quick check for missing data is executed. All missing cases will be excluded from analysing. After excluding the missing cases, the counter-indicative items are recoded into new variables. As there is only one counter-indicative item in the dataset, this item is recoded into rQualityVariance. This item is recoded for all different categories, so in total there are 8 recoded items. Once excluding cases and recoding items is done, the dataset of 225 cases is ready for analysis. First of all, when looking at the reliability for each
category, the only variable below a Cronbach’s alpha of .70 is stage of product life cycle (.670). All other variables score higher than the threshold of .70, except for quality variance in the category chocolate bars (α = .695). All those variables have a high reliability, as they all have Cronbach’s alphas higher than .70. Besides all items have a good correlation with all corrected item correlation scores above .30. After all items are final, the means for the
variables were computed and this results in Quality_VarianceTOT, Purchase_FrequencyTOT, PLCTOT, Store_LoyalTOT and Brand_LoyalTOT (see table 1).
The correlation table shows a significant (p < .01) tendency to a positive correlation between store loyalty and purchase frequency (r = .199) and stage of product life cycle (r = .206).
24
Furthermore, as can be seen, brand loyalty positively correlates on a significant level (p < .01) with quality variance (r = .305), purchase frequency (r = .404), stage of product life cycle (r = .191) and store loyalty (r = .322). Finally, both quality variance and stage in product life cycle (PLC) significantly correlate with purchase frequency. As this correlation is below .20
multicollinearity is not an issue.
Table 1 (Means, Standard Deviations, Correlations)
**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).
4.2 Results I
In the first part of the study the hypotheses 1a – 5b will be tested. For the independent variables promotion frequency and mean price, the dataset from previous research is used (Guyt and Gijsbrechts 2014). These two variables were computed into different variables, so the mean price and mean promotion could be computed. For each respondent the mean price of their chosen retailer is used. All categories consist of a mean size, except for the beer category. All respondents had to describe their last beer purchase. With the exception of 6 respondents, all respondents described if they purchased a crate of beer, a single beer or a six-pack. For each respondent the mean price is adapted to their description. For those 6
exceptions the mean price between a single beer, six-pack and crate of beer is calculated. For the promotion frequency, the average of one year is calculated for all categories individually.
Variables M SD 1 2 3 4 5 6 7 8 1. Gender 1.54 .500 2. Age 26.87 8.957 .070 3. Education 3.58 1.167 .087 .128 4. Quality variance 4.98 .987 -.097 -.073 -.016 (.872) 5. Purchase frequency 3.846 1.328 -.036 .041 -.025 .175** (.817) 6. PLC 3.333 1.0467 .036 .040 -.110 .100 .199** (.670) 7. Store loyalty 3.5288 1.348 .080 .147* -.037 .097 .192** .206** (.774) 8. Brand loyalty 5.1818 1.0914 .118 .039 .063 .305** .404** .191** .322** (.904)
25
Those promotions are transformed to weeks, so the number of promotions per week for each category is computed.
Table 2: Average promotions per week per category
First, the relationship between all independent variables and store loyalty will be measured. As all variables can be measured as scale variables to test the hypotheses, a multiple
hierarchical regression was performed to examine the relationship between the independent variables and store loyalty. The equation for this regression is as follows:
Y = α + β1 X1 + β2 X2 + β3 X3 + β4X4 + β5X5 + β6X6+ β7X7 + β8X8 + ε
Store loyalty is the outcome variable (Y), α the constant and ε is the number that is not explained by this model (residual or error). For the independent variables (Quality variance β1; Purchase frequency β2; Stage of product life cycle β3; Mean Price β4; Promotion frequency β5) and the dependent variable (Store loyalty) the control variables were gender (β6), age (β7) and education (β8). In the first block of the multiple hierarchical regression the three
predictors are gender, age and education. By controlling for these demographic variables, they do not affect the relationship between the independent and dependent variables. For the first step with the three predictors the model was not statistically significant (F3, 219) = 2.286; p > .05 (.079) and this explains 3 % of variance of consumers’ store loyalty. For the second step of the model, the independent variables (Quality variance; Purchase frequency; PLC; Mean price; Promotion frequency) are included and this leads to 32,7 % explanation of the variance in the model as a whole. This model was statistically significant (F8, 214) = 13,012; p < .001. After the entry of the independent variables, these caused an increase of 29,7% variance explained in the model. This is after controlling for gender, age and education (R2 Change =.073; F (3, 214) = 5.795; p = .001). In the final model (see table
Category Beer Chips Coffee Soda Ketchup Liquid Detergents Mayonnaise Cholate bars
Average promotion
26
3) five out of eight variables are statistically significant. Promotion frequency has the highest Beta (β = .476; p < .001), in comparison with age (β = .143; p < .05), mean price (β = .111; p < .05), stage of product life cycle (β = .159; p < .05) and purchase frequency (β = .205; p < .01). This means that the independent variable promotion frequency out of these eight
variables has the highest effect on store loyalty, followed by purchase frequency and stage of product life cycle.
Table 3: Hierarchical Regression Model of Store Loyalty and Brand Loyalty
Note. Statistical significance: *p <.05; **p <.01; ***p <.001
Brand Loyalty
To measure the effect of the independent variables on brand loyalty, a multiple hierarchical regression is performed as well. Only the outcome variable changes in the regression
equation. For the first step with the three predictors the model was not statistically significant (F3, 219) = 1.294; p > .05 and this explains only 1,7 % of variance of consumers’ brand loyalty. For the second step of the model the independent variables (Quality variance; Purchase frequency; PLC; Mean price; Promotion frequency) are included and this leads to 30,7 % explanation of the variance in the model as a whole. This model was statistically significant (F8, 214) = 11.859; p < .001. After the entry of the independent variables, these caused an increase of 29 % variance explained in the model. This is after controlling for gender, age and education (R2 Change =.272; F (3, 214) = 27.163; p < .001). In the final
Store Loyalty Brand Loyalty
R2 Change B SE β R2 Change B SE β Step 2 .297** .290** Gender (.216) (.182) .079 (.292) (.128) .132* Age (.023) (.010) 1.43* (-.001) (.007) -.004 Education (.023) (.068) .020 (.083) (.056) .087 Quality variance (.129) (.081) .092 (.328) (.067) .289*** Purchase frequency (.212) (.061) .205** (.309) (.050) .369*** PLC (.206) (.086) .159* (.068) (.062) .064 Mean Price (.116) (.060) .111* (-.025) (.049) -.030 Promotion frequency (.780) (.095) .476*** (.183) (.078) .138*
27
model (see table 3) four out of eight variables are statistically significant. Purchase frequency has the highest Beta (β = .369; p < .001), in comparison with gender (β = .132; p < .05), promotion frequency (β = .138; p < .05 and quality variance (β = .289; p < .001). This means that the independent variable purchase frequency out of these eight variables has the highest effect on brand loyalty, followed by quality variance.
To understand and interpret the regression, the standardized coefficients (β) will be used, so different measurements for each variable can be compared (Field, 2013). Nevertheless for each hypothesis individually, the unstandardized coefficients will indicate the actual increase or decrease. For hypotheses 1a and 1b, table 3 shows for promotion frequency a statistically significant effect (p < .05) on both store loyalty and brand loyalty. If the promotion frequency increases with one, the store loyalty will increase with .780 on a 7-point Likert Scale.
Therefore hypothesis 1a is supported. For brand loyalty, exactly the opposite of the hypothesis is statistically significant. On a 7-point Likert scale, an increase in the promotion frequency with one will show an increase of .183 on brand loyalty. Secondly, purchase frequency has a coefficient of .212 for store loyalty and .309 for brand loyalty. So on a 7-point Likert scale, if the purchase frequency goes up by one, store loyalty will increase with .2.12 and also for brand loyalty an increase of .356 will occur. These effects are statistically significant (p < .05) and therefore both hypotheses 2a and 2b are supported. Additionally, for stage of product life cycle (PLC), the coefficients are .206 for store loyalty and .068 for brand loyalty. Both have positive relationship, however only the effect of PLC on store loyalty is statistically
significant (p < .05). As a result only hypothesis 3a is supported. The fourth hypotheses about quality variance have unstandardized coefficients of .129 for store loyalty and .328 for brand loyalty. The positive effect of quality variance on brand loyalty is statistically significant (β = .289; p < .001) and therefore hypothesis 4b is supported. Although, there is no significant support for hypothesis 4a. Finally, if the mean price goes up by one on a 7-point Likert scale,
28
store loyalty will increase by .116 (β = .111; p <.05) and the coefficient of brand loyalty will decrease (β = -.030; p > .05). Hypothesis 5a is supported, but then again the relationship between the mean price of a category and brand loyalty is not significant.
Control variables
For both store loyalty and brand loyalty the control variables were age, gender and education. There were only two significant effects in the model. First of all, for store loyalty a positive relation came forward for the control variable age (β = 1.43; p < .05). If the age of a
respondent goes up by one year, store loyalty will increase with .023 on a 7-point Likert scale. A possible explanation for this effect might be that if consumers getting older, their salary will go up and therefore have more money to spend. They are less vulnerable for bargains and have higher preferences for the store. A second explanation might be that the older consumers get, the less spare time they have. Therefore they will more easily choose for convenience and stores nearby, instead of switching their shopping route to purchase their products.
Finally, the significant relationship between age and brand loyalty (β = .148; p < .05) shows that females are somewhat more loyal towards brands than men.
Figure 2: Overview of hypotheses
H1a The promotion frequency in a category has a direct positive effect on store loyalty.
Supported
H1b The promotion frequency in a category has a direct negative effect on brand loyalty.
Not supported H2a The average purchase frequency of a supermarket category has a direct
positive relationship with store loyalty.
Supported
H2b The average purchase frequency of a supermarket category has a direct relationship with brand loyalty.
Supported
H3a Products in early stages of the product life time cycle have a direct positive relationship with store loyalty.
Supported
H3b Products in early stages of the product life time cycle have a direct positive relationship with brand loyalty.
Not supported H4a The category quality variance has a direct positive effect on store loyalty. Not
supported H4b The category quality variance has a direct positive effect on brand loyalty. Supported
29 H5a The mean category price has a direct positive relationship with store
loyalty.
Supported
H5b The mean category price has a direct positive relationship with brand loyalty.
Not supported
In sum, the overall results are to a certain extent satisfying. Six out of 10 hypotheses are supported and indicate that these category characteristics do have an effect on both store loyalty and brand loyalty. Less satisfying is the fact that these independent variables explained for the store loyalty model only 29,7 % of variance and 29 % of variance for the brand loyalty model. An explanation can be that not only category characteristics play a role. As shown in other studies (Meyer-Waarden 2007; Liu 2007) distance to the store and loyalty programs also have enormous impact.
4.3 Results II
To measure the differences between store loyalty and brand loyalty, the ratio between the two variables is computed for each consumer. So for each consumer the mean of store loyalty is divided by the mean of brand loyalty (see table 4). All ratios are below 1.0, indicating that for all categories the mean of brand loyalty is higher than store loyalty. The minimum for all ratios is almost similar, but the maximum varies from 1.25 (Chips) to even 4.00 (Coffee). This suggests that both the coffee and soda category consists of some extreme loyal consumers towards brands and are willing to shop at different stores for these brands.
Table 4: the ratio for store loyalty and brand loyalty
Category Total ratio Minimum Maximum Std. Deviation Store mean Brand mean
Beer .7550 .14 2.43 .40573 3.6859 5.2308 Liquid detergents .7530 .19 1.80 .34844 3.5503 4.9259 Chips .6758 .18 1.25 .26197 3.5556 5.3363 Ketchup .7478 .17 1.50 .32162 3.6061 5.1455 Soda .8059 .15 3.60 .54240 3.7581 5.3118 Chocolate bars .6929 .16 1.58 .29693 3.4645 5.2077 Coffee .7585 .14 4.00 .54949 3.4354 5.0884 Mayonnaise .6974 .17 2.00 .36232 3.3548 5.0591
30
For all ratios a multiple hierarchical regression is conducted with the ratios as dependent variables. These regressions are controlled for age, gender and education. The only significant models after controlling for age, gender and education are for the categories soda and coffee (see Appendix II for all tables). All other categories did not show any significant differences between the five independent variables (Quality variance; Purchase frequency; PLC; Mean price; Promotion frequency) and the dependent variable (ratio per category). First of all, for the category soda the model was not significant with the three predictors (F3, 54) = 1.228; p > .05 and explains 6,6 % of variance of the ratio (Store loyalty / Brand Loyalty). Including the five independent variables lead to an increase of 19,7 % and this explains in total 26,4 % variance in the model. This model is statistically significant (F8, 49) = 2.922; p < .01. Table 5 shows both quality variance (β = -.345; p < .01) and purchase frequency (β = -.267; p < .05) are statistically significant and both have a negative relationship with the ratio between store loyalty and brand loyalty. Moreover, the independent variable promotion frequency has a positive relationship with the ratio (β = .361; p < .001).
Table 5: the ratio store/brand loyalty for the categories soda and coffee
Soda Coffee R2 Change B SE β R2 Change B SE β Step 2 .197** .181* Gender (-.053) (.142) -.047 (.053) (.153) .048 Age (.010) (.010) .131 (.014) (.009) .236 Education (-.025) (.072) -.045 (-.213) (.074) -.416** Quality variance (-.130) (.047) -.345** (-.137) (.080) -.267 Purchase frequency (-.075) (.038) -.267* (-.088) (.044) -.323* PLC (-.003) (.060) -.006 (.062) (.063) .143 Mean Price (.057) (.063) .133 (.013) (.116) .015 Promotion frequency (.123) (.022) .361*** (.130) (.047) .357**
Note. Statistical significance: *p <.05; **p <.01; ***p <.001
Secondly, for the category coffee the three predictors (gender; age; education) are not statistically significant (F3, 54) = 1.436; p > .05 and explain 9,1 % of the variance. By including the five independent variables (Quality variance; Purchase frequency; PLC; Mean
31
price; Promotion frequency) an increase occurred of 18,1 %. So in total 27,2 % variance in the model is explained. Including the five independent variables ensures the model becomes significant (F8, 49) = 2.487; p < .05. Both the control variable education (β = -.416; p < .01) and the independent variable purchase frequency (β = -.323; p < .05) negatively influence the ratio significantly. Additionally, promotion frequency has a significant positive relationship with the ratio between store and brand loyalty (β = .357; p < .01). Third, a regression is performed for the mean of all category characteristics and the ratio between store and brand loyalty as well (see Appendix II). This model is statistically significant (F8, 214) = 5.686; p <.001 and shows only significance for promotion frequency (β = .3.62; p < .001). Overall, promotion frequency is the only characteristic that significantly influences consumers to be more loyal towards a store than to a brand.
Finally, for all categories is established if they score high or low on store loyalty and brand loyalty and if they score high or low on the category characteristics. As the mean results of brand loyalty and store loyalty already showed, the differences between the means are relatively small. For that reason, a median split will divide those groups into low and high. Figure 3: Categories divided into high vs. low brand loyalty and store loyalty
Low store loyalty High store loyalty Low brand loyalty Mayonnaise
Liquid detergents Coffee
Ketchup
High brand loyalty Chocolate bars Chips Beer Soda
Each category is defined by category characteristics and the mean of these category
characteristics is presented in table 6. Also those characteristics are divided into low or high. The three independent variables purchase frequency, quality variance and stage of product life cycle were measured on a 7-point Likert scale, with 4 as neutral (median). Every score
32
below neutral is defined as low and every score above as high. For the independent variable quality variance all categories score above 4, indicating that on average every category has a high variation of quality. On the other hand, stage of product life cycle is evaluated for all categories somewhat low. Finally, the purchase frequency score equally high and low for both four categories. The categories that have a high score on this variable are: beer; chips; liquid detergents and coffee. Ketchup, mayonnaise, soda and chocolate bars have a low perceived purchase frequency. For the other two independent variables (Mean price; Promotion
frequency) a median split divides the categories into low and high. Beer scores the highest on mean price and fits together with liquid detergents, coffee and chocolate bars in the high price group. The categories chips, ketchup, mayonnaise and soda score low on the variable mean category price. As a final point, liquid detergents, chips, coffee and soda are the categories with the highest promotion frequency.
Table 6: The mean of all category characteristics
Beer Liquid
detergent
Chips Ketchup Chocolate bars
Coffee Mayonnaise Soda
Mean price 4.96 4.82 1.06 1.24 1.78 2.86 1.20 1.40 Promotion frequency 2.09 3.05 2.63 1.09 1.00 3.25 0.837 2.49 Quality variance 5.53 4.45 5.12 4.73 4.83 5.25 4.99 5.04 Purchase frequency 4.24 4.10 4.21 3.11 3.64 4.25 3.51 3.75 PLC 3.40 3.19 3.50 2.60 3.64 3.71 3.02 3.48
33 5 Discussion
In this section the results will be discussed. First of all, this chapter will elaborate on the findings and secondly those findings will be interpreted based on previous literature. At the end of this chapter a discussion on the theoretical and the managerial implications is given.
5.1 Findings
The main goal of this study was to examine the influence of category characteristics on both store loyalty and brand loyalty. This study sheds light on these relationships and provides therefore useful insights. The core research question of this study was: What characteristics of supermarket categories influence consumers to be more brand or more store loyal? First of all, the findings of the survey will be briefly evaluated, to give a concrete answer to the research question. In the first part of the study, the direct relationship between each category characteristic and store loyalty and brand loyalty is examined. The two outcome variables store loyalty and brand loyalty are measured separately. Consequently, this resulted in 10 hypotheses.
Promotion frequency
First of all, the relationship of promotion frequency lead to a significant positive effect on store loyalty and brand loyalty as well. These results are in contrast with previous literature (Ailawadi et al. 2001; DelVecchio et al., 2006). The latter suggests that sales promotions only have a short-term effect and do not result in an increase of brand preferences. This study’s multiple regression contradicts those findings and shows that an increase in the category’s promotion frequency effects brand loyalty in a positive manner. One explanation for this result can be that consumers became more loyal to brands with higher promotion frequencies, due to the fact that those brands are perceived relatively low-priced compared with base
34
prices. Additionally, van Heerde et al. (2004) give an explanation for the positive relationship between promotion frequency and store loyalty. An increase in promotion frequency can lead to category expansion. So this might lead to consumers adjust their shopping route for the store with the highest number of promotion frequencies or the deepest discounts. Once inside the store, they will expand more easily to other categories. This study confirms their results as in increase in promotion frequency has a significant effect on store loyalty. And even though the article of Gijsbrechts et al. (2008) states there is little empirical evidence, this study shows there is an actual increase in attitudinal store loyalty.
Purchase frequency
The second two hypotheses concern the category perceived purchase frequency. The survey specifically asks for respondents their perception on how many times they purchase products from this category. This gives insights in their attitude towards the category and is therefore a great measurement to examine attitudinal loyalty. Although Liu (2007) based her findings on longitudinal data and did not measure attitudinal loyalty, she also stressed the importance of purchase frequency. One of her outcomes is that loyalty programs have less effect on heavy buyers than on moderate and light buyers. She argues that if consumers already purchase frequently, they will not increase their purchasing behaviour. Which is not in line with this study, because an increase in purchase frequency will not only lead to an increase in store loyalty, but in brand loyalty as well. One of the explanation for this effect might be that consumers have purchased many different brands in the past. Consumers have their preferences for specific stores and brands for the products they purchase frequently
(Venkatesan and Kumar 2003). When you have to buy a product multiple times a week, you are more willing to investigate which store offers the best products. And once decided which store to visit, consumers will also made a thorough decision on which brand they will
35
purchase. On the other hand, if a store or product immediately satisfies the consumers’ needs, why should they switch? The multiple regression shows a statistically significant effect of purchase frequency on both store loyalty and brand loyalty. This effect is stronger for brand loyalty than for store loyalty. So the increase in purchase frequency will lead to a higher increase in brand loyalty than in store loyalty.
Stage of product life cycle
Third, previous literature (Bijmolt et al. 2005; Johnson et al. 2006) has contradicting findings about the product life cycle. The former stated that products in early stages of the product life cycle have benefits differentiating themselves compared with other. This differentiation should lead to more loyal consumers. As the industry of consumer packaged goods (CPG) does not have many new innovations, the survey asked the respondents about their opinion. A category can be seen as innovative if there are many brands that constantly are trying to create new features to their product. Ketchup and mayonnaise for example try to differentiate with their bottle cap, so you do not spill any sauce anymore. Moreover, in the beer category, brands differentiate with twistable beer bottles. And also the current focus on fair trade cacao has led to many new brands in the chocolate bar category. On the other hand, the more mature a brand is, the more chances the manufacture had to build a strong relationship with the consumer (Johnson et al. 2006). The survey consists of questions about how new and
innovative the category is and is all about the consumers’ perception. In this line of reasoning the stage of product life cycle should have a negative relationship with brand loyalty
according to Johnson et al. (2006). This study did not show any significant results between the stage of product life cycle and brand loyalty. Overall the stage of product life cycle scored relatively low for all categories. It can be beneficial to further examine this relationship between categories that have more variation in their outcomes. The category coffee scored the
36
highest on average for this variable with 3.70 on a 7-point Likert scale. This means that on average none of the product categories is perceived as innovative. Besides the non-significant effect on brand loyalty, there is a significant relation between stage of product life cycle and store loyalty. This indicates that if categories are perceived newer or more innovative, they attract and retain consumers. This is in line with the formulated hypothesis and can be explained by the fact that if not many stores offer those products, the consumer gets automatically more loyal to that specific store.
Quality variance
In addition, the fourth hypotheses examined the effect of quality variance on the outcome variables. If a category has a broad assortment, this will attract many different consumers (DelVecchio, 2001). Choices can both positively as negatively influence consumers’ purchasing behaviour. Consumers’ want to compare different products, although too many choices can cause an information overload. Moreover, high variance in a product category can lead manufacturers to differentiate and differentiation leads to higher brand loyalty (Bijmolt et al. 2005). For the first line of reasoning this study does not find any support, as hypothesis 4a is not significant. So the quality variance of a category does not influence the effect of store loyalty. On the other hand, the quality variance does significantly effects consumers’ brand loyalty. So if quality variance goes up by one, brand loyalty goes up .328 on a 7-point Likert scale (β = .289; p < .001).
Mean price
Finally, the mean price is an extremely valuable variable for attitudinal loyalty (Dick and Basu, 1994; Bandyopadhyay and Martell, 2007). Although these researches stress the fact that brands with loyal customers have relatively higher prices, the other way around might lead to
37
the same effect. Do consumers become more loyal towards a brand if the mean category prices are higher? It can have multiple reasons why consumers prefer higher priced brands. Examples are higher status and risk reduction. Besides higher priced products can be perceived to have higher quality (Zeithaml 1988; Johnson et al. 2001). In this study the multiple regression shows a significant positive effect of the category mean price on store loyalty. The higher the category price, the more loyal a consumer is towards a retailer. This is in line with consumers switching their shopping route, because a retailer has better prices for high priced categories. The higher the category price, the more easily money can be saved by adjusting your shopping route.
Relative importance of store loyalty and brand loyalty
The second part of the study goes deeper into the relative importance of store loyalty and brand loyalty. First, one of the findings is that respondents on average are more loyal towards brands than towards stores. According to this study, true store loyalty is weaker than brand loyalty and naturally it is only measured how consumers perceive loyalty. It can be beneficial to further investigate if the same effect occurs for actual behaviour. The results per category for store loyalty showed only small differences, which made it harder to interpret the ratios between store loyalty and brand loyalty. Besides, out of the eight performed multiple regressions, only two categories were significant (soda; coffee). For the category soda, the model showed two significant independent variables, namely quality variance and purchase frequency. Both have a negative relationship with the ratio and this means that if the
coefficient will increase for quality variance and purchase frequency the ratio between store loyalty and brand loyalty will decrease. This means that both purchase frequency and quality variance have more impact on brand loyalty in the soda category than for store loyalty. If they go up by one, brand loyalty will significantly increase more on a 7-point Likert scale than the
38
coefficient of store loyalty. In addition, purchase frequency also has the same effect in the coffee category. If consumers purchase more frequently, brand loyalty will increase more than store loyalty. Promotion frequency has the opposite effect. Consumers are significantly more loyal towards the store than towards brands. Hence, if promotion frequency increases, the ratio between store and brand loyalty will also increase. This implies that store loyalty increases significantly more than brand loyalty.
Besides, the coffee category is the only category where education as a control variable
matters. If the coefficient of education increases by one, the ratio (store loyalty/brand loyalty) decreases with .213. This means that for coffee, if consumers have a higher education they become significantly more loyal towards brands than towards stores. So higher educated consumers prefer one brand more than lower educated consumers and are willing to shop around for this brand. Overall, it can be concluded that purchase frequency and promotion frequency both have a positive effect on store loyalty and brand loyalty. Even though,
promotion frequency is more important for store loyalty and the effect of purchase frequency is stronger for brand loyalty. Furthermore, stage of product life cycle and mean price are also more important for store loyalty. As quality variance only significantly effects brand loyalty, indicating that quality variance is more important for brands.
Both parts of the study give relevant insights in order to answer the research question: What characteristics of supermarket categories influence consumers to be more brand or more store loyal? It can be concluded that all five independent variables play a significant role in either brand loyalty or store loyalty. Moreover, purchase frequency is the only product characteristic with support for both hypotheses. And as aforementioned, purchase frequency also plays a significant role in the ratio between store loyalty and brand loyalty for the categories soda and coffee. So overall, the purchase frequency of the category is one of the
