Internal and external influence to store
performance – a study of Dutch clothes
retailing industry
by
Zhao Lei
University of Groningen
Faculty of Management and Organization
Abstract
Table of Contents
1 INTRODUCTION...3
2 LITERATURE REVIEW ...6
Research concepts and constructs...6
Research questions and hypotheses ...9
3 METHODOLOGY ...14
Explanation of variables ...14
Selection of variables and the research process...16
Data collection and database description ...20
Calculation of variables ...21
4 DATA ANALYSIS AND RESULTS ...25
Data reduction...25
Statistical models...26
Outcome interpretation ...30
5 DISCUSSION ...32
Outcome discussion...32
Comparison with Asian countries...34
Limitations and future suggestions ...35
6 CONCLUSION...37
Appendices...38
1 INTRODUCTION
The store performance is a crucial issue considered by many researchers in recent years, especially in the field of marketing. However, any choice of performance indicator could be criticized for inherent limitations, as no performance indicator is a perfect measure (Silvestro and Cross, 2000). In the last decades, two indicators are widely accepted, which are the sales (Reinartz and Kumar, 1999) and the customer loyalty (Bloemer and de Ruyter, 1998). As so far few existing studies concentrate on potential influences to the two indicators simultaneously.
Some researchers link geo-demographic categories to the store performance with the purpose to find out the best location in the market place to maximize the store’s benefit (Pan and Zinkhan, 2006). These studies are concerned with the characteristics of local customers (demographic categories), for instance, the influence of age, the income level, or the family structure of citizens. These categories are from the
environment of the store, or in other words, they are external categories. Baker (2002) empirically examined how store environment cues influence consumers’ store choice decision criteria, and he found three kinds of environmental issues have impact on store patronage intentions (social influence, design, and ambient). The social construct includes characteristics of local customers (demographic factors), which was found to be most crucial (Baker, 2002). Besides, because geo-demographic influences are the initial driver for the enhancement of the store performance (Mittal and Lassar, 2001), it is interesting to focus more specifically and precisely on the geo-demographic categories alone as the external influence.
size, the product assortment can explain a sizeable variance in the store performance.
Some studies also conclude that the competition and the market position have impact on the store performance (Reinartz and Kumar, 1999), also some researchers concentrate on promotion activities (van Heerde, and Bijmolt, 2005). But due to limitations of our database, these issues will not be discussed in this paper, but they are strongly recommended to be studied in the future. For years, the debate on potential influences to the store loyalty and sales is going on (Reinartz and Kumar, 1999). Some researchers emphasize the importance of either geo-demographic categories or store attributes. But few of them put them in the same level, and they study the influences to the store performance in an imbalanced manner. For instance, some researchers only focus on customer characteristics and underestimate the importance of the store attributes (Reinartz and Kumar, 1999), some others only pay attention to the store attributes and ignore geo-demographic factors (Heald, 1972). Till now, no existing study explicitly reports if the external or internal influences have more impact on the store performance than the other. The purpose of this study is to find out the internal and external influences to 2 indicators of the store performance (the customer loyalty and sales) simultaneously. Therefore, the main research question is: how do geo-demographic categories and store attributes affect the
store performance in the Dutch clothes retailing industry, and which categories within them have the most impact on the store loyalty and sales, respectively?
China is affected by personal relationships which gains the store extra customer access or business networks (Ashley-Cotleur, 2000). Kim and Yu (2005) evaluated and compared the influence of the store attribute to store performance between South Korea and the US, and they found the nationality of stores (domestic operated or foreign operated), price of products, and the assortment of products have significant differences with contributing to the store performance. As a result, at the end of the paper, we are going to compare the differences between the Netherlands and East Asian countries based on empirical outcomes of our research. The comparison is only on the qualitative level due to the lack of a consistent database. A comparison with empirical tests between Eastern and Western countries is heavily suggested for the future study.
2 LITERATURE REVIEW
Research concepts and constructs
Store performance is usually measured by the store loyalty (how loyal do customers regarding to a specific store) and the store sales (Morphet, 1991). Store loyalty is generally defined as the strength of the relationship between an individual’s relative attitude and repeat patronage (Dick, and Basu, 1994). In this paper, the store loyalty is represented by the Penetration Rate of the Loyalty card (PRL1) (Mitchell and Kiral, 1998). The loyalty card measure approach was firstly used by O’Malley (1998) who focused on the ability of loyalty programs to encourage fidelity to a particular retail brand. And the loyalty card program has been frequently used in recent years and experienced rapid expansion all around the world. Over 150 kinds of such programs had been implemented in UK along by 1998 (Mitchell, and Kiral, 1998). From the roots in grocery sector, loyalty card programs have diffused across virtually all sectors of UK multiple retailing industries (Worthington, 1998). The phenomenon is not restricted to the UK, as loyalty card programs have been established across Europe (Boedeker, 1997) and in other parts of the developed world (Brookman, 1998). Nowadays, loyalty programs have drawn much attention and interests of marketing practitioners and academics alike. Although some researchers argue that the loyalty card scheme can not increase stores’ benefit, more empirical evidence reveals stores which carry out the loyalty card scheme have more effective patronage rate than others (Byrom, 2001) due to the availability of rewards, thus the number of card holders should be closely associated with the store performance. However, a problem with using loyalty cards for evaluating store loyalty is not the most accurate approach, because card holders may not always be the visitors. This will be discussed at the end of the paper.
Store sales, on the other hand, stands for the amount of products sold to customers from stores. Both of the store loyalty and the store sales can be considered as
indicators of the store performance (Reinartz and Kumar, 1999). However, sales are not always in line with the loyalty. Liebermann (1999) reports in his paper that through empirical tests, membership is found to enhance three kinds of store performances: image, sales, and marketing. However, “sales” is the least factor that is affected by membership of a store. With establishing membership programs, 95% of his interviewees claim that they are literally enthusiastic about recommending store products/services to friends. 25% claim to always prefer the store's products over those offered by competitors, only 20% report that they significantly buy more due to their membership, whereas 36% report that they buy only slightly more. Therefore we intend to pick up sales independently as an indicator in this paper. Sales as a more visualized indicator, is easier to be perceived and measured in numbers (Applebaum 1966). Recent papers often concern the term of “category sales” (Raju, Sethuraman, and Dhar, 1995), which accounts for the sales of each specific product category or brand within a store. However, when concentrating on the firm level instead of the outlet level, sales should account for the total expenditures that customers spend in all outlets of the firm (Reinartz and Kumar, 1999) as the indicator of the store performance. Both customer loyalty and sales are shown as the dependent variables in this paper, but the raw data of dependent variables that we collected need to be transformed to match independent variables so that the calculation of their relationship is manageable. Details of the variable transformation can be found in the chapter of Methodology.
The external influences defined in this study are the geo-demographic (geographic information, GDI) categories. The use of GDI in marketing has been greatly utilized in the last decade, and research in the area is rapidly evolving (Longley and Clarke, 1995). Some researchers (Bronnenberg and Mahajan, 2000) emphasize the
importance of the ‘location’ dimension, meaning that characteristics of customers in different locations heavily influence the store performance. Bronnenberg and
is substantial for improving its market share. Studies by Hoch et al. (1995),
Montgomery (1997) and Mulhern et al. (1998) demonstrate that the geo-demographic profiles of a store's trading area may strongly affect consumers' reaction to item prices and promotional offers. Although these studies link local profile of customers to promotional performance or market share, a more direct approach is suggested by Grewal (1999) that in an era of hypercompetition, retailers should exploit local differences and through appropriate local strategies, turn them into sources of profit. Applebaum (1966) suggests the term of sales is the best indicator of the store
performance, because usually it closely links to the profit. Also Reinartz and Kumar (1999) consider sales and sales productivity as crucial measures for a store’s success. We tend to link geo-demographic categories to sales because it is the most visualized approach to investigate the contribution of local differences to the store performance. On the other hand, Byrom (2001) and his colleagues suggest in their study that the data gleaned from loyalty card transactions can be related to an individual’s geo-demographic characteristics. Nevertheless, since more existing studies concentrate on the characteristics of the store itself, the linkage between GDI
categories and the customer loyalty thereby have not been utilized to the same degree (Kumar, 1999). Yeh (2004) implies in his paper that the store loyalty can be
associated closely to the age of customers and the family structure. McDougall et al, (1997) suggests that to establish store loyalty, the store must tailor marketing and communication strategies according to different family size and structure to the most loyal customers in order to maximize revenue and profit growth. In conclusion, lots of researchers agree upon or suggest store loyalty is a useful measure for store success, and they suppose that it has relationship with geo-demographic categories. However, we must pay attention to the change of customer characteristics along with time and regions (Clarke, 1975), which will be discussed qualitatively in the chapter of Discussion.
sites in his article that “in the face of a deluge of new products and the substantial profit opportunities available through slotting allowances, researchers have powerful incentives to make this decision correctly”. Slotting allowances means the fees that retailers charge from the distributors. The purpose of slotting allowances is to increase the exposure of the distributor’s products in the store’s shelf. Chen suggests it is advantageous for the store performance, because slotting allowance leads to larger floor size of the store. His demonstration was later quoted by a number of researchers who argue the store space allocated to a product category can have a positive impact on this category’s performance (Corstjens; Bultez; Bultez and Desmet, 1999). However, Campo and Gijsbrechts (2001) argues that Chen’s (1999) claim is
misleading, since the store attribute does not only involve the store space, but also the product assortment, or even the location where the product category put on the store floor. They also argue the store attribute should be considered in combination with geo-demographic categories (Campo and Gijsbrechts, 2001) which are described as the local customer characteristics. Besides, Desmet and Renaudin (1998) found in stead of the absolute store space, increased visibility of the product category
positively associated with its performance, as the products within a store also compete with each other. However, it only accounts for the performance of a single product category, not for the store as a whole. Therefore this paper does not concern about the product location on the store floor, and only focuses on the store size and the
assortment of products.
Research questions and hypotheses
The main research question is: how do geo-demographic categories and store attributes affect the store performance in the Dutch clothes retailing industry, and which categories within them have the most impact on the store loyalty and sales, respectively? In line with Andressen’s (1998) argumentation and other mentioned studies, 4 sub-questions are drawn as:
clothes retailing industry, and which category is the most influential?
The loyalty could be influenced by the demographics varieties from both the age and family structure (Yeh, 2004). Huddleston (2004) found changes in family structure have a dramatic impact on the amount of time consumers spend on purchase decisions, as well as their repetitive patronage to the stores. This is backed up by Bell (2005) who found the presence of children alters purchasing behaviors of households. Parents with children have been proven to shorten the time for shopping, and they tend to lock on only a few stores, save their time for looking after the children. Besides, the
children themselves are easier to be loyal to specific stores. School-age children in the household were positively associated with store loyalty in Carman’s (1970) study, and Mason (1991, 1996) found that household size and number of children were also positively associated with the store loyalty. Many researchers found significant
relationships between socioeconomic status and the store loyalty (Reinartz and Kumar, 1999). A recent study by McGoldrick and Andre (1997) found that there is negative relationship between the high income group and the store loyalty. Also, in studies by Carman, Enis and Paul (1970), and by Dunn and Wrigley (1984), it appeared that only households with low income have a positive correlation with the store loyalty.
Whereas Koutouvalas (2006) suggests in his study that educational level (out of all demographic variables) is most crucial since a higher educational level is related to an increased tendency for information seeking regarding competitive
products/services/providers. As a result, our first hypotheses are:
H1a: The family structure has the most impact on store loyalty. H1b: Socioeconomic status has the most impact on store loyalty. H1c: Educational level has the most impact on store loyalty.
A number of researchers regard store attributes as the fundamental influence on the store loyalty. Van Kenhove (1999) and his colleagues surveyed a group of shoppers who mostly agree that their store choice is closely related to the store attributes. Earlier researchers tend to link the store loyalty with store attribute variables such as the type of product purchased (Hansen and Deutscher, 1978). Sinha and Banerjee (2004) studied the attribute of the stores and customers at the same time, their study reports the store size has the dominant effect on the store choice of customers, followed by situational attributes (time pressure and gift-versus self-shopping). Osman (1993) also found the store size is a key factor that influences the store image, which directly links to the customers’ patronage behaviors. Besides, he points out that the product assortment is crucial for enhancing the service efficiency, which is also influential to the store loyalty. Morschett and Swoboda (2005) studied the store attributes’ influence on the US shoppers. However, they found different store attributes across different market regions in the USA have varied impacts on the store loyalty. This supports our assumption that potential influences to the store performance can be different with the change of culture and local customer characteristics. Nevertheless, this issue is only discussed qualitatively in this paper. Therefore, our second hypotheses are as follows:
H2a: The store size has the most impact on store loyalty.
H2b: The product assortment has the most impact on store loyalty.
(3) How do geo-demographic categories influence store sales in the Dutch clothes retailing industry, and which category is the most influential?
in-house geo-demographic application system in Tesco. It is found by using the application, geo-demographic information is visually represented, thereby aiding management decision making. It can also be used to determine stocking requirements for individual stores thus improving both order efficiency and sales. More specifically, Yeh (2004) reports the characteristics of the family members change the purchasing behavior. Especially young couples tend to have the most expenditure in new stores (Leung and Hui, 2003). Dowling and Uncles (1997) suggests the income level and average savings have dramatic impact on sales. High income groups often have a smaller portion of savings and most expenditure in stores. Reinartz and Kumar (1999) further confirmed the households’ income is typically found to be positive related to sales performance, and negatively related to price elasticity. Koutouvalas (2006) reports a negative correlation between the educational level of customers and sales since as the increase of information availability, people tend to carry out their
purchasing behaviors in many stores, instead of in only one store intensively. Rowley (2003) argues the educational level has positive impact on sales of stores with high quality products. Therefore:
H3a: The family structure has the most impact on sales. H3b: Socioeconomic status has the most impact on sales. H3c: Educational level has the most impact on sales.
(4) How do store attributes influence store sales in the Dutch clothes retailing industry, and which category is the most influential?
performance. As Davies’ application, Larson used multiple statistical models, and they yield identical results. However, Larson’s models show negative correlation between the sales performance and the store size, which is quite different from Davies’ outcomes. He argues that it is because the customers’ interests are altered from single stores to specific brands. Although less previous researchers focus on the association between product assortments and sales performance, Simonson (1999) synthesizes empirical evidence indicating that product assortment can play a key role, not only in satisfying wants, but also in influencing buyers’ preferences. He suggests male customers are more likely to buy female products in a store which specializes in female products, rather than a store with mixed product assortments. So that our last hypotheses are:
H4a: The store size has the most impact on sales.
3 METHODOLOGY
Explanation of variables
As shown in the framework (Figure 1), three geo-demographic categories are involved within the external influence to the store performance. The first category is the family structure. It concerns the data about the percentage of households with children (%CHH), households with couples (%COH), and households with single living people (%SP) within each zipcode area. All these three percentages make 100% of the total population of such area (However, in the process of data analysis, one of them has to be excluded due to collinearity problems). The second geo-demographic category is the socioeconomic status, which concerns about the income level of households in each zipcode area. 5 variables are included in the assumed model: the percentage of high income households (%HI), above-average income households (%AAI), average income households (%AI), low income households (%LI), and minimum income households (%MI). As the geo-demographic category of the family structure, one out of the 5 variables must be eliminated in order to avoid collinearity problems. The last category is the educational level of customers in each zipcode area. Three educational grades are considered. But due to database limitations, the educational grade is measured by ordinal data (from 1-8) about the overall condition of inhabitants within an area, instead of percentage numbers. The higher grade represents more people with corresponding educational level (elementary/secondary/high) within the zipcode area, vice versa.
In the internal index, there are two store attribute categories. The first one accounts for the total floor size of the store which covers specific zipcode areas (FS in M²)2
.
FS in M² is firm specific data, so that it is assumed to be associated with total store sales and the loyalty card penetration in all zipcode areas that the firm covers. The second category is about the product assortment of stores3. Desmet and Renaudin
2 FS in M²: floor size in square meters
(1998) found the visibility of products is positively related to the store performance, but different combinations of product categories can also influence the performance. Huddleston (1990) studied the relationship between the store attributes and mature female purchasing behaviors, and he found female customers between 25 and 35 are the largest purchasing group of stores. Moye (2000) claims large female product exposure is the most important chain to attract female consumers’ first visit, and then repetitive patronage is much easier to be built. Besides, Carman (1970) points out school-aged children are positively related to the store loyalty, large kid product exposure should be advantageous. As a result, the product assortment in this study is differentiated into the percentage of the floor size for the female assortment (%FA), the floor size for the male assortment (%MA), and the floor size for the kids assortment (%KA).
Three multiple regression models are applied to test the association between the internal and external influence and 2 indicators of store performance separately. The first indicator is the store loyalty measured by the PRL, which stands for the penetration rate of loyalty card holders. It represents the degree of store loyalty that the customers have. When concerning the influence from GDI categories, PRL is on outlet basis4 since GDI data is area specific; whereas when concerning the influence from store attributes, PRL is on firm basis5 since the independent variables are firm specific. On the other hand, store sales can be measured in a more salient approach (in €). The idea is to survey the total expenditure that all customers spend in the store which covers such area. Sales is zipcode area specific when testing the relationship between GDI categories and sales performance; however, when concerning store attribute influences, the customer expenditures from all zipcode areas the firm covers have to be added in order to work out the total sales of each firm. The reason to test the association with the two indicators separately is that the customer loyalty is not always in line with the customer satisfaction (Yeh, 2004) and sales (Liebermann,
1999). Applying the outcome for customer loyalty to the sales mechanically could be misleading and inaccurate (Andressen and Lindestad, 1998).
(Figure 1, Conceptual framework of the research)
Selection of variables and the research process
Both external and internal influences on store performance may involve numerous variables, but it is not realistic to collect and calculate a weighted index of all relevant variables. The idea of this paper is to select the most significant variable from certain variable sets. The selected variable is the representative of its own category. The influential categories are exclusively discussed after the empirical test.
Socioeconomic Status (%HI, %AAI, %AI, %LI, %MI)
Educational Level (EE, SE, HE)
Store Performance Store Loyalty (PRL) Store Sales (in €) Product Assortment (%FA, %MA, %KA) Internal Influence Size of Total Store Floor
(in M²) External Influence
As shown previously in the conceptual framework, the store loyalty and sales are shown as the dependent variables in the empirical test. Nevertheless, the measurement of variables is on each zipcode area (outlet) basis for external categories, but on firm basis for internal categories. Therefore, the dependent variables should be also on zipcode basis for external influences tests (store loyalty and sales in each zipcode area), and on the firm basis for internal influence tests (total firm loyalty and sales of each firm).
The reason of using data of zipcode area for external categories is firstly because of the data collection limitations (only information of zipcode areas is provided). Visits of “occasional” customers may make the outcome misleading (for instance, a foreign customer may not consider the store’s reputation in the local place, and he is not restricted by characteristics of local people, he just visits the store, or in other words, affects the store performance by chance). Also, people do not always visit the store located right in the zipcode area where they live, thus the approach has inherent limitations anyway, but due to less data access, this problem can not be solved, and it will be discussed in the part of limitation. The external influences which are tested base on zipcode areas account for geo-demographic characteristics of households within each zipcode area. In previous studies, three categories are mostly included, which are the family structure (Huddleston, 2004), socioeconomic status (Cunningham, 1961), and the educational level (Koutouvalas, 2006) of local citizens within in a specific area.
influences, internal influences and the store performance are going to be tested with multiple regression models.
The research process is illustrated in Figure 2. 15 independent variables are considered, and they are classified into four variable sets (VS1 – VS4). In our conceptual model, five categories are involved (External: Family structure, Socioeconomic status, and Educational level; Internal: Store size, and Product assortment), but “size of total store floor” has only one variable, thus it is not considered as a VS in the data analysis. For Family structure (VS1) and Socioeconomic status (VS2), the component variables make 100% of the population in the local area, it is necessary to eliminate at least one variable from VS1 and VS2 in order to avoid collinearity problems. The method is to use a pilot regression model examine the least influential variable and eliminate it. The raw data of variables in VS4 also has this problem, but through transformation, they are measured by square meters, so that we do not have to exclude any of them.
(Figure 2, Research process)
*VSs: Variable sets
most significant variable prior to the other two representatives, it therefore represents within the Family structure category, the percentage of households with couple is more associated with store loyalty than the percentage of households with single-living people. Further, it would indicate that the Family structure is more influential than Socioeconomic and Educational level.
A more simplified option is to enter all independent variables into a regression model simultaneously, even though it may result in collinearity problems. For instance, Glick and Miller’s study (1956) shows the increase in annual income associates with an increase of year of schooling, so that the percentage of high income households (%HI) may closely associate with the number of high educated people (HE), and the percentage of minimum income households (%MI) associates with the number of elementary educated people (EE). This problem can be diagnosed and solved by the factor analysis. However, if two or more independent variables from the external and internal indices respectively are classified into the same factor, it may be difficult to define this factor and the research process can be barely continued.
Data collection and database description
Data from 4013 zipcode area in the Netherlands is selected, which are covered by 28 Dutch clothes retailing firms. All the firms participating in the survey are anonymous. The data were collected in the year of 2004. Identified and differentiated by 4013 zipcode areas, the database includes the number of loyalty card holders of firms which cover an area; total expenditures of customers in the local outlet; number of
information in 747 zipcode areas, only 3266 zipcodes are available for the test, with 18% missing data. All the regression tests are two tailed, and based on 95%
confidence interval (α=0.05).
Calculation of variables
Before running regression models, it is necessary to transform variables into measurable and logical numbers in order to be correspondent to our conceptual framework. This section explains all necessary transformation of some variables. The formula of data processing can be found in Appendix I.
Firstly, the store loyalty is measured by the penetration rate of loyalty card holders (PRL). It represents the ratio of loyalty card holders for each firm out of all
households across a zipcode area. Therefore, it equals to the number of loyalty card holders divided by the number of households, so that we know what percent of local households that have at least one loyalty card. We use PRL instead of the absolute value of card holders because of the store scale difference. The population density of zipcode areas makes difference of PRL. Assume a zipcode area with small population and thereby small number of card holders does not always have lower loyalty card coverage than a zipcode area with a larger population. In this case, we concentrate on the penetration rate of cards. Therefore we use the PRL instead of the absolute number card holders. Nevertheless, a problem with using the PRL is that it ranges from 0 to 1, but in the regression equation, we observe the variables influential power in the term of “one unit change effect”, means the change of PRL should be
penetration rate of the loyalty card in all zipcode areas where a firm covers. The data processing formula of all variables is shown in Appendix I.
Sales as another indicator of the store performance, is much easier to be measured. We directly use the number of money (in €) spent in each local outlet by local customers in each zipcode area (Ysa), and the amount of money spent by all customers of the specific firm (Ysf):
Secondly, we define the percentage of households with children (%CHH) as X1. It is the sum of households with younger children, young and old children, and older children (presented in the database as %HHYO, %HHYOOLD, and %HHOLD, respectively). The percentage of households with couples (%COH) is defined as X2 and follows the same approach of calculation. It equals to the sum of households with young couples, mid-aged couples, and senior couples (presented in the database as %YCOUPLE, %MCOUPLE, and %OCOUPLE, respectively). At last, the percentage of households with single-living people (%SP) is defined as X3, equals to the sum of households with young single living people (%YSP), mid-aged single living people (%MSP), and old single living people (%OSP).
%HI, %AAI, %AI, %LI, and %MI are directly given in the database, and there is no need to carry out any transformation. In this case, they are defined from X4 to X8. However, the sum of X1, X2, and X3 makes 100% of the total population in each zipcode area, as well as X4 to X8, so that it makes collinearity problems in the regression model. We need to exclude at least one variable from each VS1 and VS2. This process is going to be done in the next chapter.
whereas “8” stands for the most people. For instance, if the zipcode area A has a “1” for elementary education and “8” for high education, it means the zipcode area A has the smallest portion of people with elementary while most people with high education out of all zipcode areas. Since ordinal numbers do not make problem of collinearity, it is no need to reduce any variable from this variable set (VS3). We define EE, SE, and HE as X9, X10, and X11 respectively.
Thirdly, there is only one variable involved in the category of the firm’s store size, which is expressed as Floor Size (FS) in M² (provided in the database). FS only accounts for the total size of a firm, not outlets, and it is not zipcode area specific. In this paper, it is expressed as X12.
The product assortment for female, male, and kids are defined as X13, X14, and X15 respectively. The database provides the percentage of product assortment for each store, by multiplying by the total floor size of the firm, we can work out the floor size of each product assortment. Transformation from percentage numbers into square meters is to avoid collinearity problems.
(Table 1, explanations of concerned variables and abbreviations)
Name of variables (Abbreviations)
Expression in the regression equation
Belonging category Meaning
PRLa Yla Store performance Penetration rate of loyalty cards in specific zipcode areas
PRLs Yls Store performance Penetration rate of loyalty cards in all zipcode areas
Ysa Ysa Store performance sales of a store in specific zipcode areas Yss Yss Store performance sales of a store in all zipcode areas %CHH X1 External influence % households with children in specific
zipcode areas
%COH X2 External influence % households with couples in specific zipcode areas
%SP X3 External influence % households with single living people in specific zipcode areas
%HI X4 External influence % households with high income
%AAI X5 External influence % households with above-average income
%AI X6 External influence % households with average income
%LI X7 External influence % households with low income
%MI X8 External influence % households with minimum income
EE X9 External influence Number of customers with elementary education
SE X10 External influence Number of customers with secondary education
HE X11 External influence Number of customers with high
education
4 DATA ANALYSIS AND RESULTS
Data reduction
Although variables are determined and transformed, the sum of X1, X2, and X3 makes 100% of the total population in each zipcode area, so that it makes collinearity problems in the regression model. We use regression analysis as the pilot test to remove the least significant variable from variable set 1. The left variables will be used further for answering the research questions. Table 2 shows the result of the regression analysis which tests the relationship between the family structure and the store loyalty (model I) and sales (model II). The outcomes reveal %CHH (X1) is the least influential variable (p=.383) to the store loyalty (Yla) whereas %COH (X2) is the least influential variable (p=.760) to sales (Ysa). X1 and X2 are excluded from the tests for Yla and Ysa respectively. Therefore, for the influence of family structure to the store loyalty, we only consider X2 and X3. And for the influence of family structure to sales, we only consider X1 and X3.
(Table 2, regression analysis for the category of the family structure (VS1))
Unstandardized Coefficients Standardized Coefficients B Std. Error Beta t Sig. (Constant) -4.244 1.778 -2.388 .017 %CHH (X1) 1.559 1.788 .077 .872 .383 %COH (X2) 1.997 1.801 .065 1.109 .268 I, Relationship between Family Structure and Store
Loyalty %SP (X3) -5.987 1.841 -.212 -3.252 .001 (Constant) 2843.782 8584.469 .331 .740 %CHH (X1) 6966.498 8807.455 .034 .791 .429 %COH (X2) -2773.513 9063.472 -.009 -.306 .760 II, Relationship between Family Structure and Sales
%SP (X3) 16218.752 9552.325 .057 1.698 .090 * Dependent Variable: Yla for model I, and Ysa for model II
Variables in the category of socioeconomic status have the same problem as
(VS2) and the store loyalty (model I) and sales (model II). Outcomes show that for the store loyalty (Yla), %AI (X6) is the least influential variable (p=.022). And for sales (Ysa), %HI (X4) has the least influential power (p=.570). Therefore, X6 and X4 are excluded from the tests for Yla and Ysa respectively, which means for the influence of socioeconomic status to the store loyalty, we only consider X4, X5, X7, and X8. However, for the influence of socioeconomic status to sales, we only consider X5, X6, X7, and X8.
(Table 3, regression analysis in the category of the socioeconomic status (VS2))
Unstandardized Coefficients Standardized Coefficients B Std. Error Beta t Sig. (Constant) -1.382 .691 -2.001 .045 %HI (X4) -3.091 .772 -.149 -4.006 .000 %AAI (X5) -2.414 .708 -.192 -3.410 .001 %AI (X6) -1.617 .704 -.129 -2.297 .022 %LI (X7) -4.262 .753 -.230 -5.660 .000 I, Relationship between Socioeconomic Status and Store Loyalty %MI (X8) -6.740 .971 -.174 -6.941 .000 (Constant) 162.358 4619.308 .035 .972 %HI (X4) 3115.882 5491.348 .016 .567 .570 %AAI (X5) 7911.890 4837.093 .062 1.636 .102 %AI (X6) 6020.958 4793.628 .047 1.256 .209 %LI (X7) 15327.403 5460.193 .079 2.807 .005 II, Relationship between Socioeconomic Status and Sales
%MI (X8) 4825.274 7697.090 .013 .627 .531
* Dependent Variable: Yla for model I, and Ysa for model II
Statistical models
Association between external variables and the store performance
Table 5 links external variables with sales. Within the VS1, both %CHH (X1, p=.040, β=.045) and SP (X3, p=.003, β=.065) significantly associate with Ys in a positive direction, the influential power of X3 is higher than X1. Only %LI (X7, p=.001, β =.067) has significant relationship with Ysa in a positive direction. In the VS3, it is found EE (X9’, p=.009, β=-.078) and HE (X11’,p=.000,β=-.108) have negative relationships with Ysa, whereas SE (X10, p=.020, β=.052) has positive impact. X11 has the most influential power than other two variables.
As a result, %SP (X3), %LI (X7), and EE (X9) are variables that being picked up for the second step test between the external index and Yla; %COHSP (X3), HE (X7), and HE (X11) are representative variables for the second step regression between the external index and Ysa.
(Table 4, Coefficients table for the external variables (Store loyalty))
Unstandardized Coefficients Standardized Coefficients Model 1.1 B Std. Error Beta t Sig. (Constant) -2.706 .218 -12.405 .000 X2 (%COH) .493 .517 .016 .953 .340 Family Structure (VS1) X3 (%SP) -7.537 .477 -.267 -15.807 .000 (Constant) -2.938 .133 -22.055 .000 X4 (%HI) -1.541 .374 -.074 -4.119 .000 X5 (%AAI) -.893 .250 -.071 -3.572 .000 X7 (%LI) 2.759 .373 .149 7.403 .000 Socioeconomic Status (VS2) X8 (%MI) 5.220 .711 .135 7.342 .000 (Constant) -3,958 ,905 -4,371 ,000 X9 (EE) ,145 ,055 ,086 2,647 ,008 X10 (SE) -,021 ,096 -,005 -,215 ,830 Educational Level (VS3) X11 (HE) -,129 ,053 -,078 -2,426 ,015
* Dependent Variable: Yla
(Table 5, Coefficients table for the external variables (Sales))
Unstandardized Coefficients Standardized Coefficients Model 1.2 B Std. Error Beta t Sig.
X1 (%CHH) 9280.810 4513.284 .045 2.056 .040 (VS1) X3 (%SP) 18439.938 6209.007 .065 2.970 .003 (Constant) 2367.194 2497.586 .948 .343 X5 (%AAI) 5851.981 3196.318 .046 1.831 .067 X6 (%AI) 3837.532 2858.370 .030 1.343 .179 X7 (%LI) 13112.618 3817.807 .067 3.435 .001 Socioeconomic Status (VS2) X8 (%MI) 2669.818 6693.829 .007 .399 .690 (Constant) 13751,248 8723,468 1,576 ,115 X9 (EE) -1385,872 533,754 -,078 -2,596 ,009 X10 (SE) 2046,712 881,871 ,052 2,321 ,020 Educational Level (VS3) X11 (HE) -1868,778 533,607 -,108 -3,502 ,000
* Dependent Variable: Ysa
Association between internal variables and the store performance
In table 6, the FS in M² has found no relationship with Ylf. In the VS4, only FA (X14,p=.039,β=1.148) has a positive relationship with Ylf. Table 7 shows the FS in M² (X12, p=.000, β =.887) has a positive relationship with Ysf. Same as the relationship with the store loyalty, only FA (X14, p=.001,β=.871) has a positive relationship with Ysf.
(Table 6, Coefficients table for the internal variables (Store loyalty)) Unstandardized Coefficients Standardized Coefficients Model 2.1 B Std. Error Beta t Sig. (Constant) -3.141 .202 -15.582 .000 Store Size X12 (FS) 3.843E-04 .000 .323 1.739 .094 (Constant) -3.200 .237 -13.482 .000 X13 (FA) -3.130E-03 .002 -.856 -1.542 .136 X14 (MA) 3.369E-03 .002 1.148 2.182 .039 Product Assortment (VS4) X15 (KA) 3.001E-04 .001 .121 .521 .607
* Dependent Variable: Ylf
(Table 7, Coefficients table for the internal variables (Sales))
Unstandardized Coefficients Standardized Coefficients Model 2.2 B Std. Error Beta t Sig. (Constant) -122577.032 149093.534 -.822 .418 Store Size X12 (FS) 1603.664 163.525 .887 9.807 .000 (Constant) -432748.994 151939.265 -2.848 .009 X13 (FA) 176.714 1299.412 .032 .136 .893 X14 (MA) 3881.760 988.309 .871 3.928 .001 Product Assortment (VS4) X15 (KA) 638.002 368.461 .170 1.732 .096 * Dependent Variable: Ysf
Association between internal and external indices and the store performance
(Table 8, Coefficients table for the internal & external index (Store loyalty)) Unstandardized Coefficients Standardized Coefficients Model 3.1 B Std. Error Beta t Sig. (Constant) -3.897 .176 -22.196 .000 X3 (%SP) -6.559 .500 -.232 -13.130 .000 X7 (%LI) 1.836 .339 .099 5.422 .000 GDI Categories (VS5) X9 (EE) .279 .029 .166 9.607 .000 (Constant) -3.221 .209 -15.398 .000 Store Attributes (VS6) X14 (MA) 1.110E-03 .001 .378 2.083 .047
* Dependent Variable: Yla for GDI categories and Ylf for store attribute.
(Table 9, Coefficients table for the internal & external index (Sales))
Unstandardized Coefficients Standardized Coefficients Model 3.2 B Std. Error Beta t Sig. (Constant) 11164.282 1598.034 6.986 .000 X3 (%SP) 10088.471 4792.440 .036 2.105 .035 X7 (%LI) 3727.248 3366.509 .019 1.107 .268 GDI Categories (VS5) X11 (HE) -1208.088 284.530 -.070 -4.246 .000 (Constant) -260353.560 145558.222 -1.789 .086 X12 (FS) 99.023 605.096 .055 .164 .871 Store Attributes (VS6) X14 (MA) 3825.166 1491.335 .859 2.565 .017
* Dependent Variable: Ysa for GDI categories and Ysf for store attribute.
Outcome interpretation
Outcomes shown in Table 8 indicate the variable from the family structure (X3, %SP) has the most impact on the store loyalty in a negative direction, which means a large percentage of households with single living people is disadvantageous for loyalty card penetration in specific areas. Besides the family structure, socioeconomic status and educational level of customers are also associated with the store loyalty (shown in table 4). Low income people are easier to be loyalty card holders, and lower educated customers tend to be loyal with specific stores. Nevertheless, they are less influential than the factor of the family structure. Therefore, H1a is supported by our tests.
product assortment category has the most positive influential power on the store loyalty, meaning more female products exposure, the store loyalty is easier to be built. There is no relationship found of the store size, as well as the male and kid product assortment as shown in table 6. Thus H2b is supported.
For sales performance, table 9 indicates high educated people (X11, HE) from the educational level of customers category appears to be the dominant factor in a negatively direction, meaning the larger number of well educated people tend to lower the sales of stores. Although large portion of households with single living people (X3, %SP) leads to higher sales of stores, it appears the influential power of the family structure is weaker than which from the educational level. Therefore, H3c is supported by the tests.
5 DISCUSSION
Outcome discussion
From the data analysis, it is found that for the store loyalty, the family structure category has predominant influential power rather than socioeconomic status and educational level of customers. Our findings are supported by Yeh (2004) and Huddleston’s (2004) study which all proves the family structure is closely associated with repetitive patronage to stores. Huddleston (2004) found the change of family structure heavily influences the amount of time consumers spend on purchasing decisions, as well as the amount of time they are willing to spend. He also found customers spending less time on purchasing are more advantageous for building the store loyalty because consumers have less time to compare goods store by store. Instead, they tend to visit the right store they most trust repetitively.
However, lots of researchers suggest that within the family structure category, the households with children are the key influence to the store loyalty (Bell, 2005, and Carman, 1970). Whereas through the empirical test, we found the households with single-living people are mostly correlated with the store loyalty in a negative direction.
Drèze and Vanhuele (2003) state in their paper that older single people is the most difficult group to be loyal to specific stores, because single people usually have more adequate time to perceive price differences, comparing goods, and collect substitute product information. But in Huddleston’s (2004) study, however, he points out that single-living people are more likely to be the loyalty card holders, as their life style is simpler than couples whose characteristics can be easier influenced by each other. Simple life style relies on lower diversity of purchased grocery goods and it even lowers their choice of brands or stores.
study. Drèze and Vanhuele (2003) collect the data from clothes stores while Huddleston (2004) surveyed food stores. Therefore Drèze and Vanhuele’s findings are more relevant for this paper.
On the other hand, our empirical test reports the educational level category has the most significant association with the sales performance. In our findings, people with high education have the most impact on sales performance by comparing with the family structure and socioeconomic status.
Larger well educated group tends to make smaller expenditures in the store. This outcome is supported by Koutouvalas (2006) who reports in his paper a negative correlation between the educational level of customers and sales. It is because higher income level widens the information seeking approach, and with the increase of information availability, people tend to less intensively visit only one store. High income people tend to spend money on investment, education, etc. rather than living expenses. Dowling and Uncles (1997) point out at the end of their paper that high income households often have small savings, and therefore more manageable money which whereas is mostly used on real estate or reinvestment rather than in stores.
In the internal index, for both store loyalty and sales, the product assortment is the most influential category, which is contrary with Sinha and Banerjee’s (2004) outcome, in which they argue the store size has the dominant effect on the store choice of customers in a positive way. Larson (2001) also suggests the store size is related to sales performance, but differently with Sinha and Banerjee, the relationship he found is in a negative direction. This variance can be caused by Sinha and Banerjee’s focus on durable goods, but Larson mainly concentrates on the store inventory instead of the product categories.
product assortment can play a key role for enhancing the likelihood that a purchase will be made and to affect sales. Bearden (1977) suggests the product assortment and atmosphere is a more crucial factor than price, quality of the merchandise, and the location of the store. Odekerken’s (2001) study provides our findings strong backup, in which he found both the technical and functional quality on the store loyalty is stronger for female than for male consumers. He suggests more female product exposure can greatly enhance the functional quality on store loyalty for female consumers, because there would be more product diversity, and the purchasing convenience is improved. According to his survey, female consumers make 72.2% out of the total visitors of clothes stores. Therefore, large floor size for female products can more effectively build the store loyalty than other product assortments.
Comparison with Asian countries
Retailing industry is expanding in Asian countries in a dramatic speed in the last two decades. With static retail markets in the US and Europe, retailers have targeted Asia where markets are growing twice as fast as at home (Wright and Sparks, 2005). Some loyalty strategies have been translated into the expanding Asian retail sector, including the loyalty cards. One advantage to carry out loyalty card strategies in the Asian market is that customers have not had the opportunity to acquire loyalty card fatigue (Nordhoff, Pauwels, and Odekerken-Schro¨der, 2004). Because of the generalization of loyalty cards in Korea, China, Singapore, and Hong Kong, now it is possible to compare its effect and influencing factors with western countries.
cultural shopping patterns.
Wright and Sparks (2005) found in most eastern countries, the age of household members is the most significant factor to the store loyalty, also the types of industry influence loyalty. Their paper reports young consumers in Asia normally have quite low loyalty of grocery stores. But for clothes stores, the loyalty is slight higher. Senior people in Asian countries have dramatic high store loyalty especially for the food retailing stores, normally some restaurant or supermarkets. These conclusions are different with our outcomes which prove the family structure is the most crucial element for building store loyalty in the Netherlands.
Another interesting finding of Nordhoff (2004) in Singapore is that the store loyalty is not truly correlated with the store satisfaction. But the store satisfaction is crucial for buying decision of customers. Therefore, the age is not the most significant influence for sales in Singapore.
Chang (2006) found in her study that in Taiwan and mainland China, product assortments have significant relationship with sales performance, but in his research, the products are assorted by brands or nationality (domestic or import products). However, with more international chain stores emerging in Asian countries, the categorization of products with “nationality” within the same store is more ambiguous. Therefore the influence of such product assortment on sales is turning weaker. The gender of visitors is becoming the most influential factor in East Asian.
Limitations and future suggestions
Second, for external analysis between geo-demographic categories and the store performance, the data is collected from zipcode areas, which greatly increased the complexity of defining and transforming variables. Besides, some zipcode areas are covered by more than one outlet so the influence from these areas is inflated on each local outlet. Using data at the firm level would make the research easier to be understood and the research questions can be answered more clearly and accurately.
Third, the loyalty card measure may not really reveal the store loyalty. Nordhoff (2004) studied the effectiveness of loyalty card schemes in the Netherlands with differentiating the store loyalty into behavioral loyalty and attitude loyalty. He found there was no correlation between card ownership and behavioral loyalty. That is to say, just because Dutch customers had loyalty cards did not mean that they spent more of their shopping budgets in those retail outlets. Interestingly, card ownership was related to attitudinal loyalty. This only meant that customers liked the shop. It did not mean that they shopped there exclusively, or even spent significant amounts of money there. Wright and Sparks (2005) suggest that loyalty cards have a life cycle. Although they may be an effective sales strategy when introduced into a market, over time they lose their effectiveness as saturation occurs and customers become habituated to the cards. Therefore, it is suggested for the further study that the store loyalty can be more specific and differentiated when expressed by the loyalty card, or other measures of the store loyalty is suggested.
6 CONCLUSION
Appendix I
Data processing
(Table 10, Data processing formula of transformed dependent variables)
Variables Formula
At area (outlets) level (for association between external influences and store loyalty): Yla = ln[PRLa/(1-PRLa)] = ln [(CHOLDERSa/HHa6)/ (1-CHOLDERSa/HHa)] Loyalty
At firm level (for association between internal influences and store loyalty): Yls = ln[PRLs/(1-PRLs)] = ln [(CHOLDERSs/HHs7)/ (1-CHOLDERSs/HHs8)] At area (outlest) level (for association between external influences and store sales): Ysa = local outlet specific sales (in €)
Sales
At firm level (for association between internal influences and store sales): Yss = overall sales of each firm (in €)
(Table 11, Data processing formula of transformed independent variables)
Variables Formula
X1 = %CHH = %HHYO + %HHYOOLD + %HHOLD X2 = %COH = %YCOUPLE + %MCOUPLE + %OCOUPLE Family structure
X3 = %SP = %YSP + %MSP + %OSP Socio-economic
status
No transformation is needed Educational level No transformation is needed
Geo-demographic
Index
Floor size X12 = FS in M² = Total floor size of a firm
X13 = FS in M² * %FA = Total floor size of a store * %female product assortment
X14 = FS in M² * %MA = Total floor size of a store * %male product assortment
Product assortment
X15 = FS in M² *%KA = Total floor size of a store * %kid product assortment
Store attribute
Index
6 CHOLDERSa/HHa: number of loyalty card holders within each area/ number of households within each area.
Reference
Andreassen TW, and Lindestad B, 1998: “Store loyalty and complex services: The impact of corporate image on quality, customer satisfaction and loyalty for customers”
- International Journal of Service Industry Management
Applebaum W, 1966: “Methods for Determining Store Trade Areas, Market Penetration, and Potential Sales”
- Journal of Marketing Research, 1966
Ashley-Cotleur C, King S, and Brazeal D, 2000: “Marketing family businesses: Do ‘Family Businesses’ attract more business? The impact of ethnicity’”
- ICSB, 45th World Conference, 2000
Babb HW, and Wynne AJ, 1982: “A Geodemographic Market Research and Planning System for Estimating Sales Potential”
- Development in Marketing Science, 1982
Baker J, Parasuraman A, Grewal D, and Glenn B, 2002: “The Influence of Multiple Store Environment Cues on Perceived Merchandise Value and Patronage Intentions”
- Journal of Marketing, April 2002
Bloemer J, and de Ruyter K, 1998:”On the relationship between store image, store satisfaction and store loyalty”
- European Journal of Marketing, 1998
Boedeker M, 1997: “Relationship marketing and regular customer cards daily product retailing in Finland'”
Bronnenberg BJ, and Mahajan V, 1999: “Spatial dependence in multimarket data: an application to endogeneity in promotion and price data”
- Working Paper, No. 341. The John E. Anderson Graduate School of Management at UCLA, Los Angeles, CA.
Bronnenberg BJ, and Sismeiro C, 2000: “Spatial prediction of market shares using a model of the United States retail industry”
- Working Paper, No. 347. The John E. Anderson Graduate School of Management at UCLA, Los Angeles, CA.
Brookman F, 1998: “Rite aid expansion strategy stresses electronic loyalty programmes” - Stores, July, pp. 102-3.
Byrom J, Hernandez T, Bennison D, and Hooper P, 2001: “Exploring the geographical dimension in loyalty card data”
- Planning, 2001
Campo K, and Gijsbrechts E, 2004: “Should retailers adjust their micro-marketing strategies to type of outlet? An application to location-based store space allocation in limited and full-service grocery stores”
- Journal of Retailing and Consumer Serices, 11(6), 369-383
Carman JM, 1970: “Correlates of brand loyalty” - Journal of Marketing Research, 7, 67-76
Chang Y, 2006: “Taiwanese consumers perceptions of American versus Taiwanese apparel brands”
- lib1.wsulibs.wsu.edu
A relevant metric for category management” - Marketing Science 18, pp. 208–229.
Clarke A, 1975: “Sociability and Social Sensitivity: Characteristics of the Stranger” - Annual meeting of the Eastern Psychological Association
Corstjens M, and Doyle P, 1981: “A model for optimizing retail space allocations” - Management Sciences 27, pp. 822–833.
Cunningham RM, 1961: “Store loyalty to store and brand” - Harvard Business Review, 39, Nov/Dec, 127-137
Davies R, 1973: “Evaluation of retail store attributes and sales performance” - European Journal of Marketing, 1973
Davies RL, and Rogers DS, 1984: “Store location and store assessment research” - Wiley 1984
Desmet P, and Renaudin V, 1998: “Estimation of product category sales responsiveness to allocated shelf space”
- International Journal of Research in Marketing 15, pp. 443–457.
Dick AS, and Basu K, 1994: “Store loyalty: Toward an Integrated Conceptual Framework” - Journal of the Academy of Marketing Science
Dowling GR, and Uncles M, 1997: “Do Store loyalty Programs Really Work?” - Sloan Management Review
- www-1.gsb.columbia.edu
Dunn R and Wrigley N, 1984: “Store loyalty for grocery products: an empirical study” - Area, 16, 4, 307-14
Dvorak RE, and Paasschen FV, 1996: “Retail Logistics: One Size Doesn't Fit All” - The McKinsey Quarterly, 1996
Glick PC, and Miller HP, 1956: “Educational Level and Potential Income” - American Sociological Review, 1956
Grewal D, Levy M, Mehrotra A, and Sharma A, 1999: “Planning merchandising decisions to account for regional and product assortment differences”
- Journal of Retailing 75, pp. 405–424.
Hansen RA, and Deutscher T, 1978: "An empirical investigation of attribute importance in retail store selection"
- Journal of Retailing, Vol. 53 No.4, pp.59-72.
Heald GI, 1972:” The application of the Automatic Interaction Detector (A.I.D) Programme and Multiple Regression Techniques to the Assessment of Store Performance and Site Selection” - Operational Research Quarterly (1970-1977), 1972
Hoch SJ, Kim B, Montgomery AL, and Rossi PE, 1995: “Determinants of Store-Level Price Elasticity.”
- Journal of Marketing Research 32, (February), 17–29.
Huddleston P, Whipple J, and van Auken A, 2004: “Food Store Loyalty: Application of a Consumer Loyalty Framework”
Huddleston P, Ford I, Mahoney MY, 1990: “The relationship between importance of retail store attributes and lifestyle of mature female customers”
- Blackwell Synergy
Kim SH, Chen-Yu J, 2005: “Discount Store Patronage: A Comparison between South Korea and the United States”
- Clothing and Textiles Research Journal, 2005
Koutouvalas DK, and Siomkos GJ, 2006: “An examination of the relationship between service quality perceptions and store loyalty in public versus private Greek banks”
- International Journal of Financial Services Management, 2006
Larson PD, 2001: “Relationships between inventory, sales and service in a retail chain store operation”
- International Journal of Physical Distribution & Logistics, 2001
Leung B, and Hui WC, 2003: “Retail loyalty system (RLS) with embedded web server” - freepatentsonline.com
Liebermann Y, 1999: “Membership clubs as a tool for enhancing buyers’ patronage” - Journal of Business Research, 1999
Longley P. and Clarke G., 1995: “GIS for Business and Service Planning” - Bell and Bain, Glasgow
Mason N, 1991: “An Investigation into Grocery Shopping Behaviour in Britain, A.C.” - Nielsen, Nielsen House, Headington, Oxford OX3 9RX
growth”
- Managing Service Quality, vol. 10, pp. 347-355
McGoldrick PJ and Andre E, 1997: “Consumer misbehaviour: promiscuity or loyalty in grocery shopping”
- Journal of Retailing and Consumer Services, 4
Mitchell W, and Kiral RH, 1998: “Primary and secondary store-loyal customer perceptions of grocery retailers”
- British Food Journal
Mittal B, and Lassar W, 2001: “Why do customers switch? The dynamics of satisfaction versus loyalty”
- Journal of Service Research, 2001
Morphet CS, 1991: “Applying multiple regression analysis to the forecasting of grocery stores: an application and critical appraisal”
- The international review of retail, distribution, and consumer research, volume 1, issue 3, April 1991, pages 329-351
Morschett D, Swoboda B, and Foscht T, 2005: “Perception of store attributes and overall attitude towards grocery retailers: The role of shopping motives”
- The International Review of Retail, Distribution and Consumer Research, Volume 15, Issue 4 October 2005, pages 423 - 447
Moye LN, 2000:”Influence of Shopping Orientations, Selected Environmental Dimensions with Apparel Shopping Scenarios, and Attitude on Store Patronage for Female Consumers”
- scholar.lib.vt.edu
programs:a comparative study in Singapore and The Netherlands”
- International Journal of Service Industry Management, Vol. 15 No. 4, pp. 351-64, ISSN 0956-4233.
O'Malley L, 1998: “Can loyalty schemes really build loyalty?” - Marketing Intelligence & Planning, Vol. 16 No. 1, pp. 47-55.
O’Malley L, Patterson M, and Evans M, 1995: “Retailing applications of geodemographics” - Marketing Intelligence & Planning, 1995
Osman MZ, 1993: “A conceptual model of retail image influences on loyalty patronage behaviour”
- The International Review of Retail, Distribution and Consumer Research, 1993
Odekerken-Schroder G, De Wulf K, and Kasper H, 2001: “The impact of quality on store loyalty: A contingency approach”
- Total Quality Management & Business Excellence
Pan Y, and Zinkhan GM, 2006: “Determinants of retail patronage: A meta-analytical perspective”
- Journal of Retailing, 82(3), 229-243.
Raju JS, Sethuraman R, and Dhar SK, 1995: “The Introduction and Performance of Store Brands”
- Management Science, 1995
Reinartz WJ, and Kumar V, 1999: “Store-, Market-, and Consumer-Characteristics: The Drivers of Store Performance”
Rowley J, 2003: “Retention: rhetoric or realistic agendas for the future of higher education” - The International Journal of Educational Management, 2003
Silvestro R, and Cross S, 2000: “Applying the service profit chain in a retail environment” - International Journal of Service Industry Management, 2000
Sinha PK, and Banerjee A, 2004: “Store choice behaviour in an evolving market” - International Journal of Retail & Distribution Management, 2004
Tan T, and Lui TJ, 2002: “Globalization and trends in international marketing research in Asia” - Journal of Business Research, 2002
Van Kenhove P, De Wulf K, and Van Waterschoot W, 1999: “The Impact of Task Definition on Store-Attribute Saliences and Store Choice”
- Journal of Retailing, 1999 - ingentaconnect.com
Worthington S, 1998: “Loyalty cards and the revitalisation of the town centre” - International Journal of Retail & Distribution Management, Vol. 26 No. 2, pp 68-77.
Wright C, and Sparks L, 2005: “Loyalty cards and customer behaviour” - Strategic Direction, VOL. 21 NO. 7 2005, pp. 18-20
Yeh YL, 2004: “The Purchasing Effects of Customer Behavior on Customer Satisfaction and Store loyalty through an Empirical Study of Frozen Prepared Foods”
