Sylwia Maria Juchnowicz
23 November 2009
Customer Loyalty Programs
How to design a preferred customer loyalty program?
Customer Loyalty Programs 2
Sylwia Maria Juchnowicz
1under supervision of
Dr. Sonja Gensler
2Prof. Dr. Peter S. H. Leeflang
Marc Hoogenberg
3
1 Sylwia Maria Juchnowicz is MSc student at Faculty of Economics and Business, University of
Groningen, The Netherlands. The research was conduced as a graduation project for the studies Business Administration in Marketing Research. Address for correspondence: Sylwia Juchnowicz, Meulmansweg 8K, 3441 AT Woerden, The Netherlands; Tel. +31 617405404; E-mail: sjuchnowicz@gmail.com; student number: s1752944.
2 Sonja Gensler is Assistant Professor of Marketing and Peter S. H. Leeflang is Professor of Marketing
at the Department of Marketing, Faculty of Economics and Business, University of Groningen, The Netherlands.
3 Marc Hoogenberg is Senior Manager at Accenture. Visiting address: Accenture, Gustav Mahlerplein
90, 1082 MA Amsterdam, The Netherlands.
Customer Loyalty Programs:
How to design a preferred customer loyalty program?
Customer Loyalty Programs 3
Management summary
Loyalty programs are critical relationship instrument aimed at capturing and growing customer loyalty. Diverse loyalty programs are mostly present in the retail sector. Popularity of loyalty programs has also surged among customers in the last years.
The goal of the present research is to investigate heterogeneity in customer preferences for loyalty program concepts. Varying customer preferences for loyalty program concepts stem from loyalty program design elements, customer characteristics and retail category characteristics.
In order to investigate the link between loyalty program design and customer preferences for loyalty programs, moderated by customer characteristics and retail category, we conduct an empirical study of the Dutch market. The final sample consists of 920 respondents and is representative for the Dutch population. The choice-based conjoint analysis is performed in order to examine customer preferences for existing and hypothesized loyalty programs. The most prevalent and desired loyalty program elements are economic rewards, privileges, card type and redemption options. Further, latent class choice modeling is used to identify homogenous customer segments of different sizes 45%, 33% and 22% of respondents, and corresponding loyalty program concepts. The archetype loyalty program concepts were found at the cross-retail category level. Customer characteristics: age, gender and shopping behavior, and retail category have moderating effect on customer preferences for loyalty program design elements.
Customer Loyalty Programs 4
Table of Contents
1. Introduction... 5
2. Customer loyalty and customer loyalty programs... 8
3. Conceptual framework... 10
3.1. Loyalty program design elements ... 11
3.2. Moderating effects of customer characteristics and retail categories... 15
4. Research design... 19
4.1. Measurement of constructs ... 19
4.2. Data collection method ... 21
4.3. Research method ... 21
5. Results ... 22
5.1. Sample description ... 22
5.2. Customer preferences for loyalty program design elements ... 25
6. Discussion ... 35
6.1. Conclusions... 35
6.2. Managerial implications... 36
6.3. Limitations and further research ... 37
References ... 39
Appendix 1 – Loyalty Programs Questionnaire... 43
Appendix 2 – Comparison of loyalty programs... 50
Appendix 3 – Loyalty card possession per retail category ... 51
Appendix 4 – Choice of latent class model ... 52
Appendix 5 – Latent class model with all covariates... 53
Appendix 6 – Comparison of latent class models ... 54
Appendix 7 – Post hoc test for differences between age groups ... 55
Appendix 8 – Post hoc test for differences between shopping behavior groups ... 56
Customer Loyalty Programs 5
1. Introduction
In the last couple of years, loyalty programs have become one of the most popular relationship marketing instruments for increasing customer loyalty. Diverse loyalty programs are mostly present in the retail sector. 37% of the store chains in the Netherlands offer a loyalty program (Bijmolt, 2008). The customer participation in loyalty programs, measured in terms of loyalty card possession, is also a widespread phenomenon. Between 80% and 90% of the Dutch households own at least 1 loyalty card and 35% of the households have 4 or more loyalty cards (Bijmolt, 2008).
Changing customer behavior, increasing expectations and eroding loyalty characterize the current business environment. To achieve a consistent growth pattern, organizations have to deliver on their loyalty promise by creating and preserving value for their customers. Getting closer to the customers and responding to their preferences, wishes and needs is a path for enhancing loyalty, encouraging a deeper business relationship and indirectly boosting profitability (Accenture report, 2008). However, the proliferation of loyalty programs in some retail categories (supermarkets, petrol stations, and do-it-yourself retailers) raises concerns whether customer loyalty programs are actually effective and create a sustainable competitive advantage for the organization. As stated by Koslowsky (1999), “none of the loyalty programs can result in a perfect world, but each of them can generate that little extra that can provide the retail marketer with potential tactical weapons.”
Customer Loyalty Programs 6 Both researchers and practitioners believe that a properly executed loyalty program can deliver significant and long-lasting benefits to a company. The main objective of a loyalty program is to create or increase customer loyalty by influencing customer preference for the loyalty program i.e. the perceived value of the loyalty program to the customer. The value of the loyalty program is determined by its attractiveness which is dependent upon two factors: (1) loyalty program design and management, the process through which customers become loyalty program members and the specification of the marketing actions (Leenheer, 2004; Kivetz and Simonson, 2003), and (2) customer characteristics, attitudes and purchase levels (Liu and Yang, 2009). Customers evaluate alternative loyalty programs based on their perceived value and select the most attractive program, which is further labeled as the preferred one. The preference toward loyalty program is demonstrated by the loyalty program adoption and/or increased usage, which eventually impacts the company’s performance measured in terms of loyalty program effectiveness and profitability.
The effectiveness of the loyalty program is determined to a great extent by the specific loyalty program design elements: timing of the rewards (Dowling and Uncles, 1997), types of the rewards (Kim, Shi and Srinivasan, 2001; Roehm, Pullins, and Roehm 2002; Yi and Jeon 2003) and their congruency with the effort required from a customer to obtain them (Kivetz and Simonson, 2002). Additionally, the cash value of the redemption rewards (costs of purchases to points ratio), the perceived likelihood of achieving the rewards and the scheme's ease of use (O’Brien and Jones, 1995) determine the value of a loyalty program to the customer. Moreover, loyalty programs are viable strategies only in certain markets and contexts, due to heterogeneity in customer’ responses to loyalty program concepts. Customers evaluate options before choosing the most preferred one. During these evaluations customer characteristics and shopping behavior play a role (Kivetz and Simonson, 2003). Multiple empirical researches provide evidence of the moderating role of customer characteristics: socio-demographics (age, gender, income, education level, etc.) and shopping behavior (shopping orientations, variety-seeking, and price sensitivity) on customer preference for the loyalty programs design elements.
Customer Loyalty Programs 7 the perspective of the customers and how customers’ preferences for these loyalty program elements are affected by customer characteristics and retail category. Different customer segments demonstrate varying preferences for different loyalty programs (Bell, et al. 2008). Customers might even avoid certain loyalty programs because they do not perceive the combination of the design elements as being attractive enough (Kivetz and Simonson, 2003). Thus, as design of the loyalty program is believed to affect its inherent potential and effectiveness (De Wulf, 2002), understanding the relative fit of individual customers to specific offers and options is more important than measuring their “absolute” preference. While understanding how customers perceive design elements of loyalty programs simultaneously, it should be accounted for potential heterogeneity in customer preferences (Jain and Singh, 2002). Obtaining insights, into which loyalty program design elements stimulate preferences of customer segments for loyalty program and what is the effect of customer characteristics and retail category on these preferences, would be of great relevance both for academia and business.
In response to the gaps in the existing academic literature, the purpose of the present research is to contribute to better theoretical and practical understanding of the way loyalty program design elements influence the value perception of a loyalty program. The aim of this paper is to investigate the impact of loyalty program concept elements on customer preferences for loyalty programs, moderated by different sets of factors: customer characteristics and retail categories. The main question of the present research is:
How to design a preferred customer loyalty program?
In order to answer the main question, additional sub-questions are derived:
1. Which loyalty program concept elements have an effect on customer preferences for loyalty programs?
2. Are customer preferences for loyalty program design elements moderated by customer characteristics or retail categories?
3. What loyalty program concepts can be distinguished based on heterogeneity in customer preferences?
Customer Loyalty Programs 8 a choice-based conjoint analytical approach in order to empirically evaluate customer preferences for loyalty program design elements. Thirdly, to examine heterogeneity in customers responses toward different loyalty programs by means of including two sets of moderating variables (1) socio-demographic characteristics and shopping behavior, (2) retail categories, and comparing the segment-specific preferences toward loyalty program design elements between 2 models. The final objective is to choose the best model explaining customer preferences and to formulate a preferred loyalty program concept per customer segment.
The remainder of the paper is structured as follows. Chapter 2 discusses the link between customer loyalty and customer loyalty programs. Further, chapter 3 develops a conceptual framework on the relationship between loyalty program design elements and customer preferences for loyalty programs, moderated by customer characteristics and retailer categories. Recent academic literature in the area of loyalty program design elements and heterogeneity in customer preferences for loyalty program concepts is reviewed. Chapter 4 discusses the research design, including a questionnaire on customer characteristics, shopping behavior and conjoint analysis design which will measure customer preferences toward loyalty program propositions for 3 different retail categories. In chapter 5, results of the analysis are presented. The paper concludes with a discussion of the results, managerial implications, limitations of the present research and directions for future research.
2. Customer loyalty and customer loyalty programs
Despite the advances in existing research on loyalty, our understanding of the phenomenon remains lacking as a consequence of not explicitly articulating the basic concept of loyalty (Fournier and Yao, 1997). In the existing research, a number of customer loyalty classifications have been identified, with attitudinal and behavioral loyalty gaining the most attention (Dick and Basu, 1994; Reinartz and Kumar, 2002). The most commonly cited definition of customer loyalty, that combines both attitudinal and behavioral conceptualizations, was proposed by Oliver (1997, p. 392).
Customer Loyalty Programs 9 purchasing, despite situational influences and marketing efforts having the potential to cause switching behavior".
This definition contains a two-step character of loyalty: customer’s perceptions and attitudes toward a particular product or service revealed by the level of his commitment and the behavioral action a customer demonstrates toward a particular product or service. Building true customer loyalty always encompasses both elements and is very difficult to achieve. Customers might repurchase the product or service due their attitudinal loyalty, but it might be driven by other factors e.g. natural satisfaction and preference with the products’ features or benefits (Kumar and Reinartz, 2006). Alternatively, behavioral loyalty might be a result of price and convenience. Marketing actions and loyalty programs might stimulate behavioral loyalty as well.
Enhancing customer loyalty is believed to be a critical ability for staying relevant, competitive and profitable (Accenture report, 2008). Loyal customers increase their share-of-wallet (Leenheer, 2004) and purchase frequency at the focal company, develop tolerance towards the downsides of the company and resistance toward competitive offers (Melnyk, 2005). Loyal customers can also attract new customers to the company by means of positive word-of-mouth and social communities. Maintaining a long-lasting win-win relationship with the customers is of high importance to the businesses. Companies aiming at increased customer loyalty implement comprehensive loyalty strategies with loyalty program being one of the most popular tools.
Customer Loyalty Programs 10 where shopping experience is satisfactory and brand commitment weak. This implies that customers are usually loyal not to a single brand but to a number of satisfactory brands. It is a common phenomenon in the retailing industry, referred to as polygamous loyalty (Dowling & Uncles, 1997). From a company’s perspective, behavioral loyalty is reflected in terms of increased retention rates, share-of-wallet, purchase frequencies and spending.
Based on the proposed customer loyalty definition and in line with the descriptions of customer loyalty programs present in existing literature (Sharp and Sharp, 1997; Bolton, et al., 2000; Leenheer, 2004), the following loyalty program definition is proposed:
Loyalty program is a long-term-oriented system of marketing actions that aims at enhancing customer behavioral loyalty by means of offering incentives for cumulative purchases with the firm.
In chapter 3, a conceptual framework will be constructed for studying the effect of the loyalty program concept elements on customer preferences for loyalty programs.
3. Conceptual framework
Customer Loyalty Programs 11 Figure 1. Conceptual framework for the effects of loyalty program design elements on
customer preference for loyalty program.
3.1. Loyalty program design elements
The relationship between customer loyalty and loyalty program effectiveness, measured in terms of profitability, is weaker than expected (Reinartz and Kumar, 2002). Companies recognize the fact that even very loyal customer might exploit discounts and services, and therefore may not be profitable. Improving customer value, i.e. economic value of a customer relationship to the company is critical to the organization. By incorporating the concept of customer value at the core of the decision making-process, a company can both measure and optimize its marketing efforts (Kumar and Reinartz, 2006). Personalized marketing interventions accommodate heterogeneity in customer preferences based on customer characteristics (Rust and Verhoef, 2005). Personalized marketing focuses on identifying and matching the preferences of each individual customer with an amount of targeted marketing actions undertaken by the company. Loyalty programs can be used to establish a personalized dialogue with customers (Dowling and Uncles, 1997). Customer value differentiation is visible
Loyalty program design elements
1. Rewards – Economic rewards
- Discount on repeat purchase - Present from a catalogue
- Rebate
2. Rewards - Privileges - Priority treatment - Special services
- Priority treatment & Special services - No privileges 3. Tiers - Normal members - Gold membership - Platinum membership 4. Redemption options - Combined-currency - Single-currency Customer preference for loyalty program Retail category: 1. Apparel 2. Do-it-yourself 3. Petrol stations Customer characteristics 1. Customer socio-demographics - gender - age - income
2. Customer shopping behavior - economic
Customer Loyalty Programs 12 in the form of compelling and personalized loyalty program concepts, which improve customer preferences and enhance customer loyalty, thus loyalty program effectiveness.
The value of a loyalty program to the customer is determined by five elements: cash value of the rewards, aspirational value of the rewards, range of choice of the rewards, relevance of the rewards to the customer and convenience in obtaining the rewards (O’Brien and Jones, 1995). For a loyalty program to be effective, loyalty program design elements have to be relevant and attainable to the customer and aligned with the desired goals of the organization (Kumar and Reinartz, 2006). Loyalty program design elements enhance customer preference toward loyalty program through a number of concept elements: economic benefits, privileges, psychological benefits e.g. status and convenience of the loyalty program usage e.g. redemption options.
Customer Loyalty Programs 13 A common practice is that after reaching the saving threshold the customer directly receives a coupon with a cash value that can be further redeemed for a reward. The most frequent economic rewards offered through a points saving medium include: presents i.e. products and services, not present in the assortment offered by a retailer; discount on repeat purchases at a retailer, and rebate i.e. receiving a percentage of money spent back. Discount is presented in a form of targeted discount where each customer receives a customized discount via: e-mail, by issuing the discount electronically at the checkout or by controlling discounts via a loyalty card. Since the focus of the present study are long-term loyalty programs and based on the above discussion, it is argued that economic rewards offered via a saving points medium have a positive effect on the customers’ preference for loyalty program.
The privileges (also called non-economic rewards) offered through a loyalty program can be classified as priority treatment, special services or a combination of both. Customers evaluate the rewards and the individual effort invested to obtain these rewards not only in an absolute sense but also relative to the effort of other customers (Kivetz and Simonson, 2002). The realization of a customer that he is provided with better value than others (to have an effort advantage), elicits a feeling of being preferred or special customer, hence impacts the preference for a loyalty program. Secondly, the amount of effort required to achieve a reward shifts the customer’s preference from necessary rewards in favor of luxury rewards. The customer effort cannot be understood in terms of monetary cost of participating in a loyalty program, which when increased negatively influences preference for privileges. To sum up, it is argued that privileges have a positive effect on customer preferences toward loyalty programs.
Customer Loyalty Programs 14 Figure 2. Loyalty program rewards
Several psychological benefits generated by loyalty program design elements enhance the customers’ preference toward a loyalty program. As already mentioned, loyalty program rewards might stimulate a feeling of getting a good deal (smart shopper feelings; Kivetz and Simonson, 2002). The simple fact of belonging to the loyalty program and participating in collecting points can also elicit psychological benefits (Dowling and Uncles, 1997). Customers typically participate in the loyalty program to obtain prestige or recognition (Wirtz, Matilla, Lwin, 2007). Existing research shows that loyalty programs which operate with three tiers: normal members, Gold members and Platinum members, elicit a feeling of status among all customers, even those who do not qualify for elite status (Dreze and Nunes, 2009). Therefore, it is argued that loyalty programs working with tiers have a positive effect on the customers’ preference for a loyalty program. Tiers are most commonly incorporated in the loyalty program design in the form of different types of loyalty cards available to the program members.
Loyalty programs need to be simple, well structured and reliable (Bijmolt, 2008) in order to be preferred by the customers. The customers prefer loyalty programs with little participation effort i.e. the activities a customer is expected to undertake in order to get the promised benefits (De Wulf et al., 2002). The convenience and ease of loyalty program usage can be improved by using combined-currency prices, where customers can pay for their rewards in
LOYALTY POINTS
Priority treatment & Special services
Coupons
Special services
REWARDS
Privileges Economic rewards
Rebate Priority
treatment
Customer Loyalty Programs 15 various combinations of point’s currency and cash. The customers who use different payment currencies experience a positive feeling of getting a good deal. Combined-currency prices either lower the psychological or perceived cost associated with a particular revenue objective (receiving a reward) or increase the amount of revenue collected (value of a reward) given a particular perceived cost (Dreze and Nunes, 2004). The customers provided with a possibility to choose a redemption method experience greater value obtained through a loyalty program. Based on the above discussion, it is argued that a number of redemption options have a positive effect on the customers’ preference for loyalty programs.
3.2. Moderating effects of customer characteristics and retail categories
Loyalty programs are the basic pillar of customer centricity and customer experience enhancement. However, customers being ever more knowledgeable, demanding, and diverse, demonstrate stable levels of satisfaction and at the same time increasing levels of expectations (Accenture, 2008). Further, customers are not homogeneous in their responses to interactions with a company. The customers’ perceptions of a company’s value proposition, stimulated by preferences toward the loyalty program design elements, are dependent upon the customer socio-demographic characteristics (age, gender, income, etc.) and shopping behavior (economic, personalizing and variety-seeking). Heterogeneity in customer responses to a company’s value proposition arises also from differences between retail categories offering loyalty programs.
Customer Loyalty Programs 16 demonstrate the same buying behavior pattern while shopping at different types of retailers and customer characteristics have a reported moderating effect on customer preferences. Further, the retail category characteristics do not impact loyalty program effectiveness (Leenheer, 2004). Therefore, it is argued that heterogeneity in customers’ preferences for the loyalty program concepts should be dependent upon a combination of customer characteristics.
Gender is expected to moderate preferences for loyalty program design elements. Female customers are more favorable toward loyalty program elements imitating one-to-one relationships, e.g. rewarding on special occasions while male customers favor loyalty program elements facilitating one-to-group communications e.g. making their status visible (Melnyk, 2005). The impact of idiosyncratic fit on customers’ responses to a loyalty program is more pronounced among women than men. Thus, women are more prone to the perceived effort advantage, which can be exhibited by means of highlighting e.g. proximity to the store, greater purchase frequency or perceived exclusiveness of double-points offer (Kivetz and Simonson, 2003). Moreover, gender has an influential role on shopping behavior. Female customers desire more variety in the retail context and there are no gender differences in relation to economic shopping behavior (Noble et al. 2006). Older customers attach greater importance to social benefits like recognition and sense of personal connection to the retailer e.g. priority over non-members. On the other hand, younger customers are more oriented toward special benefits and deals, like additional services (Patterson, 2007). Further, lower-income customers demonstrate preference for direct monetary rewards, e.g. discounts, and higher-income prompts customers to participate in greater number of loyalty programs (Van Doorn, Verhoef and Bijmolt, 2007).
Customer Loyalty Programs 17 Economic shoppers evaluate shopping in terms of price, quality and variety. Therefore, economic shoppers are expected to be bargain hunters who spread purchases across a number of retailers in search of the best program. Economic shoppers choose programs that offer them frequent discounts, bonus promotions and large pool of companies participating in a loyalty program. Having possibility to purchase additional points is also of great value to the economic shopper (Points Priority Survey, 2005). Economic rewards are more effective in case of a small heavy user segment of economic shoppers (Kim, Shi and Srinivasan, 2001). Based on the above, it can be hypothesized that both economic rewards and redemption options have higher relevance in influencing preference of the economic shopper for a loyalty program than additional privileges and tiers.
Further, the personalizing shopper has a tendency to individualize the customer role in the shop, where close personal relationship and attachment between a customer and the shop personnel is crucial for store patronage i.e. customer behavioral loyalty. Personalized treatment and special services (e.g. personal shoppers or travel planners) are of higher importance to the personalizing shopper than price, quality and selection of merchandise. Personalizing shoppers usually join loyalty programs with no point expiry date and are simply saving points for a personal, luxury experience. In addition, personalizing shoppers desire recognition and upgrades (Points Priority Survey, 2005). Thus, we can hypothesize that receiving non-monetary privileges and loyalty programs working with tiers have higher relevance in influencing the preference of the personalizing shopper for a loyalty program than monetary rewards and redemption options.
Customer Loyalty Programs 18 program design elements offering wide variety of choice, i.e. redemption options, economic rewards and non-monetary privileges, have a higher relevance in influencing preference of the variety-seeking shopper for a loyalty program.
The following table presents the summary of the expected importance of the loyalty program concept elements for economic, personalizing and variety-seeking shoppers.
Table 1. Expected importance of loyalty program design elements (hypothesized effects)
Economic Personalizing Variety-seeking
Economic rewards ++ - +
Privileges -- ++ -
Tiers - + --
Redemption option + -- ++
* ++ the most important element; + important element; - not important element; -- the least important element
Based on the above discussion, the following hypotheses are formulated:
Hypothesis 1a: Customer preferences for loyalty program design elements are moderated by customer characteristics.
Hypothesis 1b: Customer preferences for loyalty program design elements are moderated by retail categories.
Hypothesis 1c: The moderating effect of customer characteristics is stronger than moderating effect of retail categories.
Statement 1a: Economic shoppers attach the highest importance to economic rewards, followed by redemption options, tiers and privileges.
Statement 1b: Personalizing shoppers attach the highest importance to privileges, followed by tiers, economic rewards and redemption options.
Customer Loyalty Programs 19
4. Research design
4.1. Measurement of constructs
In order to conduct an empirical study, we have designed a questionnaire. The questionnaire consists of 4 parts: general information on loyalty cards possession and loyalty programs usage, stimulus presentation, customer shopping behavior questions and information on the socio-demographic characteristics of respondents. The questionnaire can be found in the Appendix 1 – Loyalty Programs Questionnaire.
Customer Loyalty Programs 20 program usage, is represented by two redemption options available to the customers: single-currency redemption (only points) and combined-single-currency redemption (points and cash). The following table presents loyalty program design elements and their levels, which have an impact on the customers’ preference for a loyalty program (the proposed list of design elements is for the purpose of the present research and therefore not exhaustive).
Table 2. Loyalty program design elements
Rewards Privileges Tiers Redemption option
Discount on repeat
purchase Priority over non-members The same card for every member Single-currency payment Product from a
catalogue available only to Special services members
Gold membership after
saving enough points Combined-currency payment Rebate Priority over
non-members & Special services
Platinum membership after paying a fee, only
for Gold members No additional privileges
(open-Customer Loyalty Programs 21 ended question) and household composition (5 categories), general card possession (7 categories) and card usage (6 categories).
4.2. Data collection method
To assess the effect of a loyalty program design elements on the customers’ preferences toward loyalty programs, moderated by customer characteristics and retail categories, we conduct an empirical study of the Dutch retail market. The data is collected by means of an online questionnaire distributed by Global Market Insite (research agency) within its commercial customer database. We have an initial sample of 1,207 Dutch customers, representative for the Dutch population. Two screening criteria were applied in order to determine the initial sample. First of all, respondents under the age of 18 are excluded from the research. We believe that people who are 18-year-old or younger do not have developed preferences toward loyalty program design elements, nor distinctive shopping behavior. Secondly, respondents have to possess at least one loyalty card in the pre-specified retail categories: do-it-yourself, apparel or petrol stations. Respondents are assigned to the retail category in which they possess the most of the loyalty cards. The limit is at least 400 respondents per retail category. These respondents receive a version of a questionnaire for the retail category to which they have been assigned.
The questionnaire was translated into Dutch language and pre-tested for comprehension and readability among a group of 30 respondents. Small adaptations to the initial questionnaire were made accordingly.
4.3. Research method
Customer Loyalty Programs 22 As already stated, the primary goal of the present research is to measure whether heterogeneity in customer responses to loyalty program design elements is customer-characteristic-specific or retail-category-specific. To determine that, we use latent class choice modeling that accounts for heterogeneity in customer preferences. The data analysis comprises of few subsequent steps. First, we derive homogeneous benefit segments. Form the academic literature it is evident that the use of benefits outperforms the other segmentation bases, because only that base can satisfy all the target segments criteria: responsiveness, sustainability, accessibility and actionability (Vriens, 1994). We utilize a latent class choice model for the purpose of benefit segmentation. A starting solution is generated by clustering respondents into segments based on their stated preferences toward design elements. Further, segment-level parameters are estimated via maximum likelihood and respondents are placed in the segment with the highest membership likelihood. This process is repeated until each respondent is placed in the segment in which he maximizes his utility (Moore, Gray-Lee and Louviere, 1998). We use inactive and active covariates, in order to assess if the solution accuracy can be improved. Significant active covariates represent moderating effects and therefore are included in the subsequent analysis. We derive two models: (1) with customer characteristics as active covariates and (2) with retail categories as active covariate. The purpose of employing both models is to assess which of those models generates more distinctive customer segmentation with significant preferences toward design elements. Finally, we compare the customers’ preferences for the design elements between different segments and derive unique loyalty program concepts.
5. Results
5.1. Sample description
Customer Loyalty Programs 23 of these respondents toward the loyalty program design elements. Consequently, the final sample consists of 920 respondents.
At a first step, we examine loyalty card possession and usage for the pre-specified retail categories and retailers. The most popular loyalty card within the do-it-yourself category is Gamma Voordeelpas with 49% of the respondents possessing one. The IKEA Family Card follows it with 43.4% of the respondents and the Karwei Kaart is owned by 30.1% of the respondents. Only 3.1% of the respondents indicated to possess a loyalty card for a do-it-yourself retail not specified in the questionnaire. Within the apparel retail category, the V&D VIPkaart is the most widespread loyalty card with 26.2% of the respondents having one. Further, 13.6% of the respondents hold the Esprit Club Card and 5.9% has the Mexx Connect Card. 14.3% of the respondents possess a loyalty card for an apparel retail not specified in the questionnaire. The loyalty cards for the petrol stations are not as popular as the loyalty cards for the retailers in the other categories. 21.6% of the respondents have the FreeBees offered by BP, further 12.5% of the respondents collect the Shell Zegels and 6.1% hold the Esso Extra Kaart. 8.1% of the respondents specify to have a loyalty card for a petrol station not listed in the survey. Air Miles, a multi vendor loyalty card, is the most common loyalty card, with 83% of the respondents possessing one (Appendix 3 – Loyalty card possession per retail category).
The average number of loyalty cards possessed is 4.64 (SD=1.69). 90% of the respondents carry more than 1 loyalty card with themselves while shopping and almost 50% indicated to carry 4 and more loyalty cards. This number is much higher than the results reported by Bijmolt (2008), where only 35% of the Dutch households possess 4 and more loyalty cards. The respondents belonging to the do-it-yourself, apparel and petrol stations categories differ significantly on the number of loyalty cards carried while shopping (Mdiy = 4.76, Mapp = 5.04,
Mps = 4.30, F2,917= 15.912, p<.01). Further, there are significant differences in the average
amount of loyalty cards held between the male and female respondents (Mmale = 4.35, Mfemale
= 4.99, t = -5.864, p<.01). Thus, respondents belonging to the apparel retail category, where the number of female respondents is almost twice as high as the number of male respondents (F = 209, M =117), in general tend to hold more loyalty cards than the respondents at do-it-yourself and petrol stations. There are also significant differences in the average number of loyalty cards owned between the usage frequency groups (Mat least once a week = 5.05, Monce in 2
Customer Loyalty Programs 24 users tend to participate in greater number of loyalty programs. Further, there are significant differences in frequency of loyalty cards usage between retail categories (Mdiy = 2.47, Mapp =
2.26, Mps = 2.14, F2,917= 4.534, p<.05). The respondents belonging to the petrol station
category are in general the most frequent loyalty cards users, followed by the respondents belonging to the apparel and do-it-yourself categories. There were no significant differences in frequency of loyalty cards usage between the male and female respondents. However, the male and female respondents different significantly on variety-seeking shopping behavior (Mmale = 3.13, Mfemale = 3.28, t = -3.240, p<.01). No significant differences in economic and
personalizing shopping behaviors between the male and female respondents were identified. Further, there are significant differences in age between genders (Mmale = 4.92, Mfemale =
4.45, t = 6.470, p<.01), indicating that male respondents are older. The correlations between main variables are presented in Table 3.
Table 3. Correlations between the customer characteristics and shopping behaviors
Age HH_ size HH_ income Nr_cards Use_ frequency Economic Variety-seeking Age 1 HH_size4 -,118** 1 HH_income5 -0,003 ,289** 1 Nr_cards6 -,065* ,088** 0,054 1 Use_frequency7 -0,015 -0,022 -0,035 -,321** 1 Economic 0,039 0,03 -,079* ,071* 0,013 1 Variety-seeking -0,056 -0,033 0,037 ,132** -,067* ,102** 1 Personalizing ,199** 0,004 -0,038 0,008 -0,057 ,072* ,107**
**Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed).
As expected, the customer characteristics related to the household: size and income, were highly correlated with each other (Table 3). Further, number of cards carried while shopping is significantly correlated with age and household size, as well as, with economic and variety-seeking shopping behavior. Not surprisingly, household income is negatively correlated with economic shopping behavior and frequency of loyalty card usage with variety-seeking shopping behavior. It was expected that customers with higher levels of disposable income
4 HH_size – household size 5 HH_income – household income
6 Nr_cards – number of loyalty cards carried with while shopping
7 Use_frequency – frequency of a loyalty card usage. The ‘frequency of loyalty card usage’ has a reversed scale, because of the
Customer Loyalty Programs 25 are less price-quality oriented. Variety-seeking shoppers tend to favor a variety of options for collecting and spending their points, thus they use loyalty cards more frequently. Moreover, age is significantly correlated with personalizing shopping behavior. Shopping behaviors are also positively correlated with each other.
In order to determine whether heterogeneity in customers’ responses to the loyalty program design elements is customer-characteristics-specific or retail-category-specific, first we have to allocate respondents between different groups. Distribution between retail categories was performed while conducting the questionnaire. Division of respondents between economic, variety seeking and personalizing shopping behaviors is determined. First, reliability and internal consistency scores for each construct are measured by Cronbach’s alphas. Economic shopping behavior has a Cronbach’s alpha = 0.848, personalizing shopping behavior has a Cronbach’s alpha = 0.811 and variety seeking shopping behavior has a Cronbach’s alpha = 0.806. Hence, high internal consistency scores allow for item reduction into 3 one-scale factors that will further be applied in the analysis. For this purpose we undertake the following steps: (1) 12 shopping behavior items were standardized per respondent; (2) the items belonging to one construct were summed up; (3) the membership probabilities for each respondent for each shopping behavior were allocated; (4) each respondent is assigned to the shopping behavior for which his membership probability is at least 0.5. 160 respondents with equal membership probabilities for two shopping behaviors were identified and consequently assigned to both groups. Table 4 presents the distribution of respondents between shopping behaviors and retail categories.
Table 4. Distribution of respondents between shopping behaviors and retail categories.
ECONOMIC
VARIETY-SEEKING PERSONALIZING
sum per retail category
DO-IT-YOURSELF 217 40 86 343
APPAREL 219 83 81 383
PETROL STATIONS 191 70 93 354
sum per shopping behavior 627 193 260 1080 5.2. Customer preferences for loyalty program design elements
Customer Loyalty Programs 26 included in the conjoint study. The utility range is calculated by subtracting the part-worth utility of the least preferred level of each attribute from the part-worth utility of the most preferred level of each attribute. The relative importance (on the basis of inter-attribute trade-offs) is expressed in percentage of the total range (De Wulf, et al. 2002).
To begin with, we identify the number of segments to be further investigated in terms of customer preferences for loyalty program concept elements. The most popular set of model selection tools in latent class choice analysis are the information criteria: AIC, BIC and CAIC indices8 (Fraley and Raftery 1998). Those indices are parsimony-based because all other
things being equal the model with the fewest parameters is better. Thus the minimum rule applies. The three statistics are based on the value of -2*log likelihood of the model, the number of parameters and the sample size. Since AIC9 does not consider the sample size, we
will look at the BIC10 and CAIC11 indices (which are almost identical) because of the big
sample size of 920 respondents. The minimum BIC and CAIC values are with 15 latent classes model, which is not a manageable solution. Therefore, we investigate the graph presenting the BIC and CAIC values for the models with different numbers of segments (Appendix 4 – Choice of latent class model). We choose a solution for which the difference between the indices’ values levels off to achieve a manageable and distinctive solution and at the same moment not to loose valuable information. We decide for a solution with 3 benefit segments. We assess the significance of the variables in differentiating between customer preferences between latent classes. We include retail category, shopping behavior, loyalty card possession and usage, as well as, demographic variables (age, gender, household size, household income and household composition) as active covariates into the model. The examination of the significance of the covariates reveals that retail category, shopping behavior, age and gender, all with p-values below 0.05, are significant in moderating customer preferences toward loyalty program designs (Hypotheses 1a and 1b, without income, accepted; Appendix 5 – Latent class model with all covariates). Thus, these variables will be included as active covariates in the subsequent analysis. The remaining variables will be employed as inactive covariates only for description of the segments. We run the following 3 latent class models to
8 BIC – Bayesian Information Criterion, AIC – Akaike Information Criterion, CAIC – Consistent Akaike
Information Criterion
9 AIC = -2ln(L) + 2p 10 BIC = -2ln(L) + p*ln(N)
Customer Loyalty Programs 27 assess changes in the goodness of model prediction as measured by R²: (1) without covariates, (2) with shopping behaviors, age and gender as active covariates and (3) with retail categories as active covariate. The three models have relatively similar R²: 0.2837, 0.2840 and 0.2841 respectively (the standard aggregate model has an R² of 0.0891). In addition, the ‘hit rates’ for the models are similar as well and equal to 61.27% for the model with no covariates, 61.24% for the model with customer characteristics and 61.44% for the model with retail category. Furthermore, the L² p-values for the 3 models are comparable and all models are significant at the 0.01 level. Therefore, we investigate the differences in attribute importance and part-worth utilities between latent classes for the 3 models. There are major within-model differences between attribute importance and part-worth utility scores for the loyalty program design elements. However, the between-models differences pertain only to little changes in magnitude of attribute importance and part-worth utilities (Appendix 6 – Comparison of latent class models). Therefore, it can be stated that the benefit segmentation results in a distinctive and manageable 3 latent class solution. It is supported by the significance of all attributes in determining customer preferences toward loyalty program elements, with p-values < .01. Inclusion of active covariates merely improves the accuracy of the results. Hence, no support for the main statement of the present research is found. It cannot be acknowledged that heterogeneity in customer preferences toward loyalty program design should be exclusively contingent upon customer characteristics and shopping behavior or retail category (Hypothesis 1c, rejected). Accordingly, both models offer similar and homogenous 3 latent class solutions. However, since our primary interest is a model explaining heterogeneity in customer preferences by customer characteristics, we will focus on a 3 latent class choice model with shopping behavior, age and gender as active covariates. As already stated, our focus is to assess the moderating effect of customer characteristics on the customer part-worth utilities. For this purpose, we employ the covariates table (Table 5) and the posterior mean estimates for the individual coefficients.
Customer Loyalty Programs 28 customer characteristics on the customer preferences for particular levels within the attributes, we conducted an independent samples t-test (gender) and analysis of variance (shopping behavior, age groups and retail categories) on the posterior mean estimates for individual coefficients.
Table 5. Class profiles – mean probabilities.
COVARIATES CLASS 1 CLASS 2 CLASS 3 p-value
SHOPPING BEHAVIOR 0,0170 Economic 0,5708 0,5415 0,6565 Variety-seeking 0,1641 0,2219 0,1449 Personalizing 0,2652 0,2366 0,1987 AGE 0,0022 18 - 24 & 25 -34 0,1946 0,2182 0,1434 34 – 44 0,2372 0,2820 0,2575 45 – 55 0,2925 0,2889 0,2368
55 - 64 & above 65 y/o 0,2757 0,2110 0,3623
GENDER 0,0028
Male 0,3823 0,4380 0,5353
Female 0,6177 0,5620 0,4647
As can be observed from Table 6, female customers are more favorable toward product from a catalogue (Mmale = -.6777, Mfemale = -.5764, t = -2.177, p<.05) and male customers prefer
to receive a rebate (Mmale = .7726, Mfemale = .5771, t = 3.295, p<.01). Furthermore, female
have higher preference for special services (Mmale = .0349, Mfemale = .0535, t = -3.772,
p<.01) or even no privileges option (Mmale = -.2828, Mfemale = -.2416, t = -3.223, p<.01). In
contrast, male customers demonstrate higher preference for priority over non-members (Mmale = .1193, Mfemale = .0930, t = 3.502, p<.01), combination of priority and special
services (Mmale = .1285, Mfemale = .0951, t = 3.337, p<.01), and Gold card (Mmale = .1504,
Mfemale = .1316, t = 3.098, p<.01). Thus, supporting the results of Melnyk (2005), we can
Customer Loyalty Programs 29 Table 6. Independent samples t-test - differences in mean customer preferences between genders.
t-test
Sig.
(2-tailed) Mean male
Mean female REWARDS
discount on repeat purchase -3,820 ,000 -,0949 -,0006 product from a catalogue -2,177 ,030 -,6777 -,5764
rebate 3,295 ,001 ,7726 ,5771
PRIVILEGES
priority over non-members 3,502 ,000 ,1193 ,0930 special services -3,772 ,000 ,0349 ,0535 priority & special service 3,337 ,001 ,1285 ,0951 no additional privileges -3,223 ,001 -,2828 -,2416 CARD TYPE normal card -3,842 ,000 ,2950 ,3539 Gold card 3,098 ,002 ,1504 ,1316 Platinum card 3,943 ,000 -,4454 -,4855 REDEMPTION OPTIONS combined currency -1,645 ,100 -,1563 -,1299 single currency 1,645 ,100 ,1563 ,1299
In the table 7, the differences in customer preferences toward loyalty program elements moderated by customer age are presented. In order to allow for an explicit comparison, we combined the existing 6 age groups into 3 categories: below 35 y/o, 35-54 y/o and above 55 y/o. As can be observed from table 7, older customers are more favorable toward receiving a rebate (M<35 = .5368, M35-54 = .6303,M>55 = .8132, F = 5.520, p-value < .01). Moreover,
older customers have stronger preferences for priority over non-members (M<35 = .0892, M 35-54 = .1010, M>55 = .1221, F = 4.778, p-value < .01), combination of priority over
non-members and special services (M<35 = .0889, M35-54 = .1044, M>55 = .1348, F = 5.383,
p-value < .01), and combined currency redemption (M<35 = -.1537, M35-54 = -.1549,M>55 =
-.1063, F = 3.657, p-value < .05). Contrary, younger respondents exhibit higher preference for receiving a product from a catalogue (M<35 = -.5231, M35-54 = -.5834, M>55 = -.7626, F =
7.585, p-value < .01), special services (M<35 = .0547, M35-54 = .0469, M>55 = .0359, F =
3.437, p-value < .05), no additional privileges (M<35 = -.2328, M35-54 = -.2524, M>55 =
-.2928, F = 5.740, p-value < .01) and single currency redemption (M<35 = .1537, M35-54 =
.1549, M>55 = .1063, F = 3.657, p-value < .05). Further investigation of the differences in
Customer Loyalty Programs 30 age group > 55 y/o (Appendix 7 – Post hoc test for differences between age groups). Hence, supporting the hypotheses of Patterson (2007), we can state that older customers attach greater importance to social benefits like recognition and sense of social connection to the retailer e.g. priority over non-members and younger customers are more oriented toward special benefits and deals, like additional services.
Table 7. ANOVA - Differences in mean customer preferences per age group.
F statistic Sig. Mean age 18-34 Mean age 35-54 Mean age 55 and above REWARDS
discount on repeat purchase ,535 ,586 -,0155 -,0469 -,0506 product from a catalogue 7,585 ,001 -,5231 -,5834 -,7626
Rebate 5,520 ,004 ,5386 ,6303 ,8132
PRIVILEGES
priority over non-members 4,778 ,009 ,0892 ,1010 ,1221 special services 3,437 ,033 ,0547 ,0469 ,0359 priority & special service 5,383 ,005 ,0889 ,1044 ,1348 no additional privileges 5,740 ,003 -,2328 -,2524 -,2928 CARD TYPE normal card ,593 ,553 ,3452 ,3254 ,3216 Gold card ,830 ,436 ,1385 ,1432 ,1342 Platinum card 1,666 ,190 -,4836 -,4686 -,4558 REDEMPTION OPTIONS combined currency 3,657 ,026 -,1537 -,1549 -,1063 single currency 3,657 ,026 ,1537 ,1549 ,1063
Customer Loyalty Programs 31 Table 8. ANOVA - differences in mean customer preferences per shopping behavior.
F statistic Sig. mean E mean VS mean P REWARDS
discount on repeat purchase 1,204 ,300 -,0505 -,0601 -,0103 product from a catalogue 5,259 ,005 -,6866 -,5041 -,5802
rebate 3,624 ,027 ,7371 ,5642 ,5906
PRIVILEGES
priority over non-members 3,148 ,043 ,1134 ,0940 ,0949 special services 2,368 ,094 ,0404 ,0498 ,0520 priority & special service 3,533 ,030 ,1222 ,0936 ,0974 no additional privileges 3,772 ,023 -,2759 -,2374 -,2444 CARD TYPE normal card 1,204 ,301 ,3223 ,3179 ,3479 Gold card 1,996 ,136 ,1384 ,1511 ,1339 Platinum card 1,511 ,221 -,4607 -,4690 -,4818 REDEMPTION OPTIONS combined currency 4,040 ,018 -,1278 -,1851 -,1345 single currency 4,040 ,018 ,1278 ,1851 ,1345
As can be observed from the table 8, shopping behavior has a significant moderating effect on customer preferences for several attribute levels. Thus, economic shoppers have strong preference for rebate (Me = .7371, Mvs = .5642, Mp = .5906, F = 3.624, p-value < .05),
priority over non-members (Me = .1134, Mvs = .0940, Mp = .0949, F = 3.148, p-value < .05),
combination of priority over non-members and special services (Me = .1222, Mvs = .0936, Mp
= .0974, F = 3.533, p-value < .05), and combined currency redemption (Me = -.1278, Mvs =
-.1851, Mp = -.1345, F = 4.040, p-value < .05). Further, variety-seeking shoppers favor
product from the catalogue (Me = -.6866, Mvs = -.5041, Mp = -.5802, F = 5.259, p-value <
.01) and single currency redemption (Me = .1278, Mvs = .1851, Mp = .1345, F = 4.040,
p-value < .05). Personalizing shoppers choose for special services (Me = .0404, Mvs = .0498, Mp
= .0520, F = 2.368, p-value < .1). Further comparison of differences in customer preferences between shopping behaviors reveals that for none of the attribute levels there are significant differences between all three shopping behaviors (Appendix 8 – Post hoc test for differences between shopping behavior groups).
Customer Loyalty Programs 32 = .7188, F = 7.735, p-value < .01), combination of priority over non-members and special services (Mdiy = .1283, Mapp = .0843, Mps = .1198, F = 7.931, p-value < .01). Further, apparel
customers have the highest preference for discount on repeat purchase (Mdiy = -.0868, Mapp
= .0615, Mps = -.1103, F = 20.616, p-value < .01), special services (Mdiy = .0365, Mapp =
.0618, Mps = .0380, F = 12.796, p-value < .01), normal card (Mdiy = .3012, Mapp = .3910, Mps
= .2885, F = 20.348, p-value < .01) and combined currency redemption (Mdiy = -.1502, Mapp
= -.0953, Mps = -.1858, F = 11.313, p-value < .1). Petrol stations customers prefer to receive
priority over non-members (Mdiy = .1172, Mapp = .0813, Mps = .1123, F = 9.756, p-value <
.01), Gold card (Mdiy = .1479, Mapp = .1155, Mps = .1581, F = 19.362, p-value < .01) and
single currency redemption (Mdiy = .1502, Mapp = .0953, Mps = .1858, F = 11.313, p-value <
.01). Further analysis of differences in customer preferences between retail categories reveals that a significant variation exists between 1) do-it-yourself customers and apparel customers, and 2) petrol stations customers and apparel customers (Appendix 9 – Post hoc test for differences between retail category groups).
Table 9. ANOVA – differences in mean customer preferences per retail category.
F statistic Sig. Mean DIY APPAREL Mean
Mean PETROL STATIONS
REWARDS
discount on repeat purchase 20,616 ,000 ‐,0868 ,0615 ‐,1103 product from a catalogue 2,156 ,116 ‐,6848 ‐,5697 ‐,6085
rebate 7,735 ,000 ,7716 ,5082 ,7188
PRIVILEGES
priority over non-members 9,756 ,000 ,1172 ,0813 ,1123
special services 12,796 ,000 ,0365 ,0618 ,0380
priority & special service 7,931 ,000 ,1283 ,0843 ,1198
no additional privileges 7,260 ,001 ‐,2821 ‐,2274 ‐,2701 CARD TYPE normal card 20,348 ,000 ,3012 ,3910 ,2885 Gold card 19,362 ,000 ,1479 ,1155 ,1581 Platinum card 17,257 ,000 ‐,4491 ‐,5065 ‐,4466 REDEMPTION OPTIONS combined currency 11,313 ,000 ‐,1502 ‐,0953 ‐,1858 single currency 11,313 ,000 ,1502 ,0953 ,1858
Customer Loyalty Programs 33 sizes: 44.58%, 32.79% and 22.63% of respondents. There are major differences between latent classes for attributes importance (Table 10), rescaled for comparison utility scores (Table 11) and customer characteristics (Table 5). To begin with, the attributes with the highest importance scores are rewards (class 3), card type (class 1) and redemption options (class 2). Privileges receive low attribute importance scores, with the highest one for the class 3.
Table 10. Attribute importance.
ATTRIBUTE IMPORTANCE CLASS 1 CLASS 2 CLASS 3
rewards 0,3408 0,1798 0,6629
privileges 0,0797 0,0505 0,1594
card type 0,4486 0,2294 0,0989
redemption 0,1309 0,5403 0,0788
Our next step is to gain more insights into the moderating effects of age, gender and shopping behavior on customer preferences for specific attribute levels. As can be observed from Table 11, segment 1 has a high preference for normal card, discount on repeat purchase, combined currency redemption and special services. It is a segment dominated by female respondents, with a high percentage of personalizing and variety-seeking shoppers. Hence, normal card is preferred by females belonging to variety-seeking (Mmale = .2634,
Mfemale = .3471, t = -2.531, p<.05) and personalizing shopping behaviors (Mmale = .3150,
Mfemale = .3827, t = -2.285, p<.05). The same groups of female customers also prefer
discount on repeat purchase: variety-seeking shoppers (Mmale = -.1491, Mfemale = -.0124, t =
-2.531, p<.05) and personalizing shoppers (Mmale = -.0619, Mfemale = .0443, t = -2.559,
Customer Loyalty Programs 34 between genders. Latent class 3 has the strongest preference for rebate, followed by a combination of priority over non-members and special services, Gold card and single currency redemption. It is a segment characterized by higher number of male than female respondents, high percentage of respondents above 55y/o and economic shoppers. Rebate and combination of priority and special services is favored by male respondents, especially in the age groups 35-44 y/o and above 55y/o. Further, Gold and Platinum cards, as well as single currency redemption option receives higher preference scores for male respondents. To sum up, loyalty program 3 is aimed at male respondents in the age group > 55 y/o.
Table 11. Latent class – profiles.
CLASS 1 CLASS 2 CLASS 3 p-value
REWARDS 1,80E-266
discount on repeat purchase 0,4493 0,2455 0,0495 product from a catalogue 0,1634 0,4045 0,0143
Rebate 0,3873 0,3500 0,9363
PRIVILEGES 3,80E-21
priority over non-members 0,2582 0,2614 0,3199 special services 0,2756 0,2652 0,2108 priority & special service 0,2575 0,2529 0,3455 no additional privileges 0,2087 0,2205 0,1238
CARD TYPE 3,80E-193
normal card 0,5357 0,3841 0,3274
Gold card 0,3049 0,4094 0,4051
Platinum card 0,1594 0,2065 0,2676
REDEMPTION OPTIONS 3,60E-57
combined currency 0,5549 0,2611 0,4452 single currency 0,4451 0,7389 0,5548
NONE OPTION 1,10E-112
0 0,6472 0,8959 0,8518
1 0,3528 0,1041 0,1482
Customer Loyalty Programs 35
6. Discussion
6.1. Conclusions
The findings from the present study show a great variation in customer preferences for loyalty program design elements. Supporting our expectations, this variation in customer preferences is to a great extent contingent upon the loyalty program concept elements and is moderated by the customer characteristics and retail categories. Based on a thorough academic research, supported by insights gained from studying the existing loyalty programs, we have identified the following loyalty program design elements to affect customer preferences: economic rewards, privileges, psychological benefits (status, represented by tiers) and convenience of loyalty program usage (redemption options). Specific levels within each of these 4 elements were distinguished, which in the subsequent analysis were found to significantly differentiate between customer preferences. Further, we found significant moderating effects of few customer characteristics and retail categories on customer preferences for loyalty program elements. As hypothesized, shopping behavior, age and gender are moderating customer preferences. There was no support found for the moderating effect of household income. Somewhat contrary to the expected effect, the combination of customer characteristics, age gender and shopping behavior, was not found more significant in moderating customer preferences for loyalty program design elements. One explanation for this finding might be that most of the customers actually do not demonstrate one apparent shopping behavior but exhibit a mix of the shopping behaviors. We have also shown that age and gender are correlated with shopping behavior. Accordingly, customer age is strongly related to the personalizing shopping behavior, male and female respondents differ significantly in their level of variety-seeking shopping behavior and there are significant differences in age between genders. As a result, it is essential to evaluate customers’ preferences for a loyalty program on the basis of a combination of customer characteristics.
Customer Loyalty Programs 36 specific preferences for loyalty program elements can be explained by means of moderating effects of age, gender and shopping behavior. By investigating the non-significant customer variables it is possible to gain further insights into the differences between 3 customer groups.
6.2. Managerial implications
The results of the present research could help managers to improve their decisions concerning loyalty programs effectiveness. Loyalty programs play an important role in retailing. However, the loyalty program implementation is strongly driven by market standards (Leenheer, 2005). Thus, companies follow closely their competitors experience and copy it when designing own loyalty program. Investigation of the existing loyalty programs within three retail categories, do-it-yourself, apparel and petrol stations, reveals little or no differences in loyalty program concepts across retail categories. On one hand, using competitor’s experience might generate valuable information and help to avoid mistakes in loyalty program design and management. However, what is rather prevalent is that copying the competitor’s loyalty program leads to a sub-optimal situation. Firstly, retailers do not take advantage of the whole spectrum of loyalty program design elements, mainly focusing on a limited number of the most common elements. In addition, knowledge of the customer preferences for loyalty program elements is not properly utilized, leading to insufficient focus on the customer side. As a result, ineffective and unprofitable loyalty programs trouble companies.
Customer Loyalty Programs 37 Companies could gain an advantage over their competitors by differentiating loyalty program design. A unique, customer-oriented loyalty program should cause customers to specifically prefer a particular loyalty program, thus increase the usage of this program. We propose 2 paths for loyalty program improvement. One direction is to match the preferences of the target market with one of the archetype loyalty concepts. Secondly, companies that do not offer a loyalty program but are interested in introducing one, have a possibility to design a totally new loyalty program based on the target segment preferences. Following the guidelines of the present research, businesses are able to consciously develop a loyalty programs for the existing and new customer segments oriented on customer preferences. Adapting the loyalty program design to the key factors, namely customer characteristics, might drive the effectiveness of a loyalty program. The benefits to an organization of mastering loyalty program concepts can be considerable and include most importantly building true loyalty, positive WOM and communities, and effectiveness profits. Pursuing a loyalty strategy based on learning from customer preferences leads to strengthened value and brand proposition, and increased long-term profits. The ultimate objective of a loyalty program, that is creating a sustainable competitive advantage for the organization, will be thus achieved.
6.3. Limitations and further research
Customer Loyalty Programs 39
References
Bijmolt, Tammo (2008), “Loyalty programs; Insights from the recent academic research,” Customer Insights Center, Report Nr. 2008-01
Bolton, Ruth N., P.K. Kannan, and Matthew D. Bramlett (2000), “Implications of Loyalty Program Membership and Service Experiences for Customer Retention and Value,” Journal of the Academy of Marketing Science, 28 (1), 95–108.
De Wulf, Kristof, G.J. Odekerken-Schröder, Marie Hélene de Canniere, and Claudia van Oppen (2003), “What Drives Customer Participation to Loyalty Schemes? A Conjoint Analytical Approach,” Journal of Relationship Marketing, 2 (1-2), 69-83
Dick, Alan and Kunal Basu, (1994), “Customer loyalty: toward an integrated conceptual framework,” Journal of Marketing Science, Vol. 22 (2), 99-113.
Donthu, Naveen and David Gilliland (1996), “Observations: The Infomercial Shopper,” Journal of Advertising Research, March/ April 1996, 69-76.
Dorotic, Matilda, Peter C. Verhoef and Tammo H. A. Bijmolt (2009), “Loyalty Programs – Current Knowledge and Research Directions”, draft version.
Dowling, Grahame R., and Mark Uncles (1997), “Do customer loyalty programs really work?”, Sloan Management Review, 38 (4), 71-82.
Drèze, Xavier and Joseph C. Nunes (2004), “Using Combined-Currency Prices to Lower Customers’ Perceived Cost,” Journal of Marketing Research, Vol. XLI (February 2004), 59-72. Drèze, Xavier and Joseph C. Nunes (2009), “Feeling Superior: The Impact of Loyalty Program Structure on Customers’ Perceptions of Status”, Journal of Customer Research, vol. 35 (6), 890-905.
Fournier, Susan and Julie L. Yao (1997), “Reviving brand loyalty: A reconceptualization within the framework of customer-brand relationships”, International Journal of Research in Marketing, Vol.14 (5), 451-472.
Fraley, C., and Raftery, A.E. 1998, How many clusters? Which clustering method? – Answers via model-based cluster analysis. Department of Statistics, University of Washington: Technical Report no. 329.
Gehrt, C. Kenneth and Soyeon Shim (1998), “A Shopping Orientation Segmentation of French Customers: Implications for Catalog Marketing,” Journal of Interactive Marketing, Vol.12 (4), 34-46.
