The role of customer delight in creating customer loyalty

(1)

Thesis Master Business Administration – Marketing Track

A customer at the local supermarket explores an offering

“The role of customer delight in creating customer

loyalty”

Frédérique Beatrice Hanselaar

5940494

Final version – January 31, 2015

MSc. Business Administration – Marketing track

Amsterdam Business School – University of Amsterdam

Supervisor: Dr. A. Krawczyk

(2)

Statement of originality

This document is written by Student Frédérique Beatrice Hanselaar who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

Table of contents 1. Abstract………..p.4 2. Introduction………p.5 3. Literature review………p.7 3.1. Customer Equity p.7 3.2. Loyalty p.7 3.3. Customer Delight p.8

3.4. Recession and loyalty p.10

3.5. Another point of view on loyalty strategy p.11 3.6. The supermarket environment in The Netherlands p.12 3.7. Interview with marketing director Jumbo, Mr. M. Moeken p.13 3.8. Research gap, research question, hypotheses and conceptual model p.14

4. Method………..p.18 4.1. Sample p.18 4.2. Measurement of variables p.19 4.3. Statistical procedure p.22 5. Results………...p.25 5.1. General results p.25 5.2. Correlation analysis p.26

5.3. Conditional effect: moderation p.28

5.4. Conditional effect: moderation for specific groups p.34

6. Discussion……….p.37

6.1. Theoretical and practical implications p.37

6.2. Limitations and further research p.41

7. References……….p.43 8. Appendixes………p.48

Appendix 1: Atmosphere impression of Trader Joe’s

Appendix 2: Marketing campaign Jumbo 2014: Roy Donders, #samenvierenwekerst and

Top-3 ranked ‘Gouden Loeki’ commercial ‘Moestuin’

Appendix 3: Questionnaire in Dutch

(4)

1. Abstract

What should major food retailers in the Netherlands nowadays do to attract and remain loyal customers? Should they for example follow the US model and delight customers with joy, surprise and humor? To answer these questions, the moderating role of the customer delight factors; joy, surprise and humor are investigated on the effects of three types of customer loyalty strategies. These strategies are value equity (VE), brand equity (BE) and relationship equity (RE), together called the customer equity drivers (CED). The hypotheses are developed based on marketing theories and a correlation and a moderation analysis are used to test them. The results show that customer delight partly impacts customer loyalty intentions towards the retailer. Using humor in a VE- strategy will increase loyalty intentions for customers under the age of 30. Using humor in that strategy for other age groups will have an opposite effect and lower the loyalty intentions. That opposite effect is also observed for surprise in a VE-strategy and humor and BE. These findings contribute academically by attaching a new variable to the customer delight research. Humor has not been investigated in this setting before and has proved its effect. From a managerial perspective, the insight in customer delight shows that for food retailers investing in humor, next to a good price, good quality and convenience for customers under the age of 30, will be fruitful.

Keywords

Customer delight, joy, surprise, humor, customer equity drivers, customer loyalty, moderation, food retailing

(5)

2. Introduction

There was a time that convenient locations, special or unique assortments, greater or better services than competitors and store credit cards were enough to keep loyalty from customers. This time has changed (Kotler & Keller, 2009). Companies need to do more to remain and attract loyal customers.

True loyalty affects profitability and is therefore important to measure for future firm growth. Delighting customers seems necessary to produce consequences such as loyalty. Especially, in times of recession, having loyal customers would save firms from financial concerns. Ou et al. (2013) suggest that since low confident customers are more cautious and selective in times of recession, companies should adapt a value equity strategy; a strategy which is primarily focused on price and quality.

However, some retail specialists argue that major retailers in the Netherlands nowadays are not able to surprise their customers anymore (Droge, 2014). They suggest that trying to make customers loyal works better than reducing prices (Den Hollander, 2014). An example of a grocery store in the USA that is good with customer delight is Trader Joe’s. Trader Joe’s surprises its customers with their products, atmosphere and humor. The focus is on binding customers, creating customer delight and consequently loyalty. It seems to work tremendously; in 2014 Trader Joe’s was ranked highest as favorite grocery store chain in North America (Anderson, 2014).

Therefore, another strategy that focuses more on customer delight creation might be more effective for major food retailers in the Netherlands. Accordingly, it is interesting and important to investigate the following research question: “How does customer delight in food retailing impact customer loyalty intentions towards the retailer and which variables can moderate this relationship?”

To investigate this research question the moderating role of customer delight on the effects of the customer equity drivers will be used: value equity, brand equity and relationship equity on customer loyalty intentions. Positive surprise and joy are the key ingredients in creating customer delight. Next to surprise and joy, humor will also be used as a variable to measure customer delight. The control variables are age, gender, education level, family- and employment situation. The research has been carried out with the help of grocery chain Jumbo and the market research bureau Kien. The results are interpreted with the software Statistical Package for Social Sciences (SPSS). A correlation analyses and the tool Process are used to test the relationships.

(6)

This research will be relevant from an academic and managerial perspective. First, little research has been done so far on the moderating effect of customer delight on loyalty and the variables underlining customer delight. This thesis will fill that academic gap and investigate the addition of a variable, namely humor. Second, these findings can be of magnificent relevance for management. If it turns out that the positive moderating effect of ‘customer delight’ is present and if the outcome can tell which variable within ‘customer delight’ is responsible for that outcome, managers can act accordingly and invest more money and effort in ‘customer delight’ creation.

The thesis will first discuss the relevant literature. Customer equity, loyalty, customer delight, recession and loyalty and another view on loyalty strategy will be explained. The supermarket environment in the Netherlands will be discussed and the marketing director of Jumbo, Mr. M. Moeken is interviewed. This part concludes with a research gap, a research question, hypotheses and a conceptual model. Second, the method section will discuss the sample, how the variables are measured and how the research is executed statistically. Third, the results section will show the outcome of the different statistical tests of correlation and moderation. Fourth, the discussion with theoretical and managerial implications, limitations and suggestions for further research will follow. References and appendixes can be found at the end of the document.

(7)

3. Literature review

The literature review describes the three drivers of customer equity (CE): value equity (VE), brand equity (BE) and relationship equity (RE). The concept of loyalty is described and the way loyalty is measured. Loyalty can be a consequence of customer delight, which is in turn caused by excitement factors. Different points of view on loyalty strategy are discussed. To better understand the background of food retailing, the supermarket environment in the Netherlands will be discussed. And to get more managerial and real-live marketing insight, marketing director Jumbo, Mr. M. Moeken is interviewed. The literature review concludes with a research gap, a research question, hypotheses and a conceptual model.

3.1 Customer equity

The key drivers of firm growth are value equity (VE), brand equity (BE) and relationship equity (RE) (Lemon et al., 2001). VE is about the customer’s objective assessment of the utility of a brand influenced by price, quality and convenience. BE is the customer’s subjective assessment of the brand influenced by brand awareness, attitude towards the brand and corporate ethics. RE is defined as: “The tendency of the customer to stick with the brand, beyond this objectively and subjectively perceived value” (Lemon et al. 2001). For RE the key drivers are loyalty programs, special recognition and treatment, affinity programs, community-building programs, and knowledge-building programs. Together, this strategic marketing approach is called customer equity (CE). It puts the customer at the center of the corporation and is based on the total of the discounted lifetime values of all the firm’s customers. To evaluate this marketing approach a framework is formulated to measure this return on marketing: strategic marketing trade-offs are made on the basis of projected financial impact (Rust et al. 2004).

3.2 Loyalty

A long-term sustainable advantage is achieved mostly by the firm’s ability to retain, sustain, and nurture its customer base. This has to be done by looking beyond repurchase behavior alone. The concept of loyalty does that. According to Oliver, Rust & Varki (1997) loyalty is:

“A deeply held commitment to rebuy or repatronize a preferred product/ service consistently in the future, thereby causing repetitive same-brand or same brand-set purchasing, despite situational influences and marketing efforts having the potential to cause switching behavior.”

(8)

Loyalty is different from repeat purchases. A customer can just be indifferent and buy the same product over and over again (repeat purchases) or a consumer can be very committed but not make frequent repeated purchases since there is no need for a new product (loyalty). True loyalty affects profitability (Reichheld, 2003). True loyal customers buy more over time and make recommendations to friends and relatives. Therefore, it is important to measure loyalty for future firm growth. Loyalty can be measured in a variety of ways; as repeat purchase frequency (Tellis, 1988), as repurchase intentions (Reynolds & Arnold, 2000), as likelihood of switching or likelihood of buying more (Selnes & Gønhaug, 2000). Consultant Reichheld (2003) claims that the Net Promotor Score, which can be defined as customer’s intention to recommend the firm to others, is the only number that counts.

3.3 Customer delight

“It will not suffice to have customers that are merely satisfied” (Deming, 1986). Delighting customers seems necessary to produce consequences such as loyalty (Oliver, Rust, & Varki, 1997). Rust & Oliver (2000) define customer delight as:

“A positive emotional state resulting from having one’s expectations exceeded to a surprising degree.”

Where customer satisfaction is about exceeding one’s expectations, customer delight is based on receiving a positive surprise that is beyond their expectations (Berman, 2005). The loyalty curve is relatively flat within the zone of satisfaction and climbs rapidly as a result of delight (Dick & Basu, 1994). In the literature, a distinction is made between three groups of requirements: must-be, satisfier and delight (Kano et al. 1984). The ‘must-be’ requirement is a basic criterion that consumers take for granted. However, if the requirement is not included in the product or service, it will result in extreme customer dissatisfaction. The elements of the ‘satisfier’ can further a consumer’s satisfaction beyond the basic product. ‘Delight’ will be generated if excitement factors are met. Excitement factors add utility beyond that is expected and are unexpectedly and surprisingly pleasant. Yet, if these factors are not met, no feelings of dissatisfaction will remain. This model helps to explain the difference between satisfaction and delight. Competitive advantage can in turn be generated by consistently providing delightful experiences in a way in which key competitors are not able to match this accurately.

(9)

Fuller and Matzler (2008) argue that because customers have different expectations, it is necessary to separate which factors fall in which category. This will result in a better knowledge of the preferences of customer segments. The product attributes that generate delight and “must” will also be different among consumer segments. The study shows that there is a clear difference between the lifestyle groups, which builds on Oliver’s expectation-disconfirmation model (Oliver, 1980). Oliver’s model explains that satisfaction exists in a state of disconfirmation when performance exceeds expectations.

Positive surprise and joy are seen as the key ingredients in customer delight according to some academics and consultants (Berman, 2005). Vanhamme (2000) investigates the emotion of surprise and its influence on satisfaction. Since ‘features that are surprisingly pleasant’ are able to delight, the research of surprise is interesting for the delight research (Rust et al. 2000). Lindgreen and Vanhamme (2003) examine in a follow-up article when and how surprise can be applied in retaining companies’ customers. The authors conclude in the article with the statement that surprise can be a useful marketing tool, but some situations are more suited for surprise than others.

There are certain ways to measure customer delight, yet there is no commonly accepted scale (Berman, 2005). First, it has been measured by emotional responses (Kumar, 1996). By asking customers in a specific setting to suggest an appropriate name for the emotion (for example: exhilarated, thrilled, delighted or exuberant) and rank them on a five-point scale ranging from very little to very much. Second, customer delight has been measured with in-depth interviews (Arnold et al. 2005). Third, researcher’s assumed a positive effect between performance and expectations to generate surprise and measured accordingly on scales of positive affect and ‘pure arousal’ (Oliver et al. 1997).

The study of customer delight has some critical notes. First, how long does the customer delight take? At a certain point the customer is used to the attribute or service and will expect it the next time. The delight has turned into a must. Second, if customer’s get used to being surprised, they will be disappointed if there is no new surprise effect anymore. Although, if the company is able to communicate to its customers that they cannot take the surprising feature for granted, this negative effect will not necessarily appear (Rust et al, 2000). And third, delighting the customer is costly.

(10)

3.4 Recession and loyalty

According to Reinartz and Kumar (2000), having loyal customers would save firms from financial concerns during economically hard times. However, during times of recession consumers tend to be less loyal (Lamey et al. 2007). An important question therefore is how firms can retain customers during recessions. Different suggestions are present: focus on offering firm value, focus on a good brand or focus on customer relationships. These strategies however do not offer a clear answer.

Ou et al. (2013) answer the question by investigating the moderating role of consumer confidence of the customer equity drivers on loyalty. Consumer confidence measures customers’ expected changes and variance in their household finances and the economic climate. This can range from low customer confidence with uncertainty and a pessimistic outlook to high confidence with certainty and a positive outlook. The results of the meta-analyses show that the positive effects of customer equity drivers on loyalty intentions are partly dependent on different levels of customer confidence. For low confident customers value equity is more important, consumers are more cautious and selective. The authors do not find significant impact of customer confidence on the link between relationship equity and customer loyalty. They explain this outcome by the fact that relationship equity is built on a solid partnership between customers and firms and is an important determinant of customer loyalty (compared to value equity and brand equity). External forces such as customer confidence therefore have a narrower effect on relationship equity. The managerial advice is to adapt strategies during recessions since customers with different levels of customer confidence have diverse preferences. The authors mention the Dutch retailer Albert Heijn for adapting their strategy during the recession. They decreased price levels while trying to improve service quality (Van Kampen, 2013), so they focused strongly on value equity. Also, they increased the number of AH-to-go stores for increasing convenience in buying. Albert Heijn made growing turnover and lower profit. The conclusion of the academic article is that firms should consider consumer confidence as an important driver of effectively adjusting customer loyalty strategies to their specific situation. In particular, during recessions, when consumer confidence is relatively low, value equity is effective for retaining customers.

(11)

3.5 Another point of view on loyalty strategy

Dröge (2014), Sloot (Rijlaarsdam, 2014) and Rutte (Den Hollander, 2014) suggest another point of view on Albert Heijn’s strategy. Instead of valuing them for their survival during the recession, Dröge (2014) suggests that Albert Heijn is starting to lose its strong position as market leader. Chains as Jumbo and Lidl are driving up steam and are getting closer to fill the gap. An important element for this growth is that Lidl is for example able to surprise its customers; with their fresh fruits and vegetables they surpass expectations constantly while AH customers have overstated expectations. If expectations are not fulfilled, a negative image remains.

Laurens Sloot, professor retail and marketing at the University of Groningen, says that Jumbo has, in contrast with Albert Heijn, lots of space and freedom for marketing. “They keep strictly to their own rules, such as the low pricing guarantee and short check-out lines, but outside of that there is freedom for marketing campaigns.” (Rijlaarsdam, 2014) They have guts, are creative and innovative according to supermarket specialist Gerard Rutte. They ask themselves: “What would the consumer really want?” and just do it. A campaign like Roy Donders would never have happened at Albert Heijn. In the Roy Donders campaign customers could collect five coupons and by paying €9,99 a specially designed ‘Roy Donders juichpak’ could be ordered (please see appendix 2). The campaign was an unexpected big success. The fact that Jumbo is a family business plays a major role in the overall success as well. It is more authentic than other retail businesses.

In the months November and December 2014, another development was observed at Albert Heijn. According to the website Inprijsverhoogd.nl (who did the research for Distrifood) AH reduced the prices of 3.700 products in silence, in the overall assortment of 26.000 products (Den Hollander, 2014). The supermarkets and producers are worse off; the consumer will be the only winner. “A disastrous route”, as supermarket specialist Gerard Rutte calls it. “Albert Heijn tries to change its image of expensive supermarket, but in this way they are only throwing away money. The customer won’t take notice of it at all since the changes are constantly very little.” Rutte has a better advice: “Try to make loyal customers, cut the Bonuscard into pieces and replace it by a savings card. That will work way better than reducing prices which lead nowhere.” The newly introduced AH Bonuscard gives registered customers extra discount by email based on their previous buying habits. It offers the personal service from before, but in a renewed way (Deibel et al. 2015). Clearly, the concept

(12)

of surprise and creating loyalty seems to play an important role in the market position of food retailers.

A good example of a store that is good with customer delight in the USA is Trader Joe’s (Orr, 2006). Trader Joe’s devoted customer like the food, the prices and the sense of humor. “When you look at food retailers, there is the low end, the big middle and then there is the cool edge – that’s Trader Joe’s”, says Richard George, professor food marketing at Saint Joseph University (PA, USA). Joe Columbo created the chain in 1958. He tried to adopt a tropical theme and focused on building no-frills stores with hard to find gourmet products at impossibly low prices. A product only enters the store when it is approved by a regional tasting panel (Wu, 2003). Other aspects that stand out and are able to create customer delight at Trader Joe’s are: the continuing surprise of new products, the friendly staff who pack your shopping bag, the ability to enter a raffle for free shopping if your brought your own bag. All taken together, it creates an atmosphere which makes customers fans of Trader Joe’s.

In a survey with 6.200 North Americans by Market Force Information, Trader Joe’s was ranked highest as favorite grocery store chain (Anderson, 2014). According to the Chief Marketing Officer of Market Force, Janet Eden-Harris, Trader Joe’s created loyal customers because of its quirky branding, unique private label products such as Speculoos Cookie Butter and a constant rotating array of merchandise. About the grocery chains in general she states: “We have found that delighted customers are three times more likely to recommend a grocery store than those who had an OK experience. This tells us that chains that really ‘wow’ their customers on their first visit can establish brand advocates who go on to recommend the grocery to friends and family.” (Please see appendix 1 for an atmosphere impression of Trader Joe’s.)

3.6 The supermarket environment in The Netherlands

To better understand the background of food retailing in the Netherlands, the supermarket environment in the Netherlands will be discussed. On a yearly basis, the Dutch spend €42,2 billion on food, beverages and tobacco (CBS). 78,9% (€33,3 billion) of this total amount is spend in supermarkets. Counted up, it comes to an average spending of €84,60 a week per household, which is €38,06 a week per inhabitant. According to a Deloitte report called ‘Bedrijfsvergelijking 2014, zelfstandige levensmiddelen detailhandel’, Albert Heijn had a market share of 33,8% in 2013, followed by Jumbo (20,6%), Lidl (9%) and Aldi (7,4%). In that same year Albert Heijn had 939 branches, Jumbo 422 branches, Aldi (502), Lidl (385)

(13)

and C1000 (223). Albert Heijn still is the market leader, however Jumbo is growing enormously; the amount of stores grew to 499 in 2014 and turnover to 4,8 billion in the same year (Driessen, 2014). From the general numbers of all supermarkets can be said that the continuous difficult economical times let consumers to focus on price and sales items (Op Heij, 2014). The volume of supermarket spending decreased in 2013, while the population grew a bit (CBS). This volume decrease can be seen as a consequence of an increase in online spending outside the supermarket environment, an increase of sales abroad, a decrease in throwing out groceries and the fact that boundaries between branches are fading (Op Heij, 2014).

3.7 Interview with Marketing Director Jumbo, Mr. M. Moeken

After formulating the theoretical framework, it is interesting to get more managerial and real-live marketing insight in the Dutch supermarket environment. Questions such as to what extent are marketing managers aware of customer delight? What do they think of humor as another driver of customer delight? Are marketing managers actually busy executing loyalty programs? Do marketing managers look at the retail environment in other countries? And have they heard of Trader Joe’s before?

Many successful marketing campaigns in food retailing in 2014 came from Jumbo (see appendix 2); the campaign with Roy Donders during the World Cup, the #Together we celebrate Christmas campaign, one of their TV commercials achieved top-3 rankings in the ‘Gouden Loeki’ and Jumbo was the most discussed topic on Twitter in combination with commercials in 2014 as well (Beemster, 2015). Therefore, it was very interesting to interview Marketing Director Mr. M. Moeken about the current developments in food retailing. Moeken has been Marketing Director at Jumbo since July 2013. Before this job he has, amongst other jobs, worked as marketing manager at Danone, at Ahold and at C1000.

Moeken knows Trader Joe’s well. He thinks it is so successful because of its authenticity, the genuineness and the people behind the brand; they have 100% the Trader Joe’s DNA. Trader Joe’s possesses the core characteristic of a strong brand; it is impossible to copy and therefore unique. He therefore praises Aldi for keeping the company as it was and not changing the Trader Joe formula to Aldi’ stores when they took over the company in 1979. American companies in other branches such as Starbucks, Abercrombie & Fitch and Holister, are quite successful in the Netherlands. “Wouldn’t Trader Joe’s be feasible in the Netherlands?” Moeken: “Definitely a little corner in the supermarket, but not a whole store.

(14)

Simply because Trader Joe’s customers are mostly highly educated people and since the amount of people is so much higher in the USA, with especially highly educated people at the East and West coast where the stores are located, it is more feasible there than it is here.” Another important difference between the USA and the Netherlands, is how the country and the supermarket climate are constructed. Most Dutch people have to ride their bikes for 200 meters only to find a supermarket. Americans and for example the French have to drive miles for their groceries. This results in differences in grocery habits. An element of service at the checkout, which plays a crucial role at Trader Joe’s, is not very feasible to adapt in the Netherlands as well. Minimum wage in the Netherlands is twice as much as in the US. In case Jumbo would change this, competition could cut their price margins and that would make the competitor cheaper. So, the feasibly of a whole Trader Joe’s store in the Netherlands is according to Moeken not very high because of the specific target group of highly educated customers, the differences in the supermarket climate and the differences in minimum wage and therefore the possibility of creating high service levels.

Moeken concludes the interview with an important take-away about the power of a successful brand. The core of building a brand is making the brand humane; adding some ‘joie de vivre’ to the brand. The power of a successful brand therefore lies internally and is intrinsic. The brand has to be real and the brand has to be authentic. In that sense, a question Moeken asks himself is: “Would I want to have Jumbo or another supermarket as a friend, if it were a person?” “For some grocery stores in the Netherlands, I wouldn’t answer that question affirmative.”

3.8 Research gap, research question, hypotheses and conceptual model

After reviewing the current academic literature and the current trends, the following research gap can be drawn. True loyalty affects profitability and is therefore important to measure for future firm growth. Delighting customers seems necessary to produce consequences such as loyalty. Positive surprise and joy are seen as the key ingredients in creating customer delight. Especially, in times of recession, having loyal customers would save firms from financial concerns. Ou et al. (2013) suggest that since low confident customers are more cautious and selective in times of recession, companies should adapt a value equity strategy. However, some authors (Droge, 2014) argue that major retailers nowadays are not able to surprise their customers anymore and focusing on loyalty creation works better than reducing prices (Den Hollander 2014). Therefore, another strategy that focuses more on customer delight creation

(15)

might be more effective. Accordingly, it is interesting and important to investigate the following research question:

“How does customer delight in food retailing impact customer loyalty towards the retailer and which variables can moderate this relationship?”

To answer this question, the moderating role of customer delight on the effects of the customer equity drivers will be investigated: value equity, brand equity and relationship equity on customer loyalty intentions. The variables that cause customer delight are excitement factors. Positive surprise and joy are seen as the key excitement factors (Berman, 2005). These two variables will therefore be used in the research. The factor ‘humor’ has not been measured in the context of customer delight yet. However, in the context of viral marketing, it seems to play a relevant role (Dobele et al. 2007). As well as in the context of TV commercials; the award winning Dutch commercials “Gouden Loeki” were especially so successful because they made almost three quarters (73.5%) of the people laugh (Oude Elferink & Dam, 2015). This, in combination with the findings of humor at the supermarket Trader Joe’s, makes it an interesting variable to research as well.

These findings lead to the development of the following hypotheses:

Joy

These hypotheses are based on findings in previous research that have shown that the customer equity drivers are positively related to loyalty intentions (Rust, Lemon, and Zeithaml, 2004). Delighting customers seems necessary to produce consequences such as loyalty (Oliver, Rust, and Varki, 1997). Customer delight is a positive emotional state resulting from having one’s expectations exceeded to a surprising degree (Rust and Oliver, 2000) and joy and positive surprise are seen as the key ingredients in customer delight (Berman, 2005). Combining the customer equity drivers with joy is therefore expected to positively impact loyalty intentions.

H1 The strength of the relationship between Value Equity and Loyalty Intentions is positively effected by joy.

H2 The strength of the relationship between Brand Equity and Loyalty Intentions is positively effected by joy.

H3 The strength of the relationship between Relationship Equity and Loyalty Intentions is positively effected by joy.

(16)

Surprise

Next to the findings discussed above, are features that are surprisingly pleasant able to delight (Rust et al., 2000). Surprise can, in some situations, be a good marketing tool (Lindgreen and Vanhamme, 2003). Albert Heijn is according to some authors loosing its strong positions as market leader because they are not able to surprise their customers anymore and firms as Lidl are (Droge, 2014). Therefore, the combination of the customer equity drivers with surprise is expected to positively impact loyalty intentions.

H4 The strength of the relationship between Value Equity and Loyalty Intentions is positively effected by surprise.

H5 The strength of the relationship between Brand Equity and Loyalty Intentions is positively effected by surprise.

H6 The strength of the relationship between Relationship Equity and Loyalty Intentions is positively effected by surprise.

Humor

Humor plays a major role in viral marketing (Dobele et al., 2007), in TV-commercials (Oude Elferink and Dam, 2015) and at American supermarket chain Trader Joe’s (Orr, 2006). Marketing director Jumbo, Sir Moeken, confirms the importance of humor in customer loyalty creation. Humor has not been measured in the context of customer delight yet. However, humor does in the contexts mentioned above, have a positive impact on loyalty intentions. Therefore, the combination of the customer equity drivers with humor is expected to positively impact loyalty intentions.

H7 The strength of the relationship between Value Equity and Loyalty Intentions is positively effected by humor.

H8 The strength of the relationship between Brand Equity and Loyalty Intentions is positively effected by humor.

H9 The strength of the relationship between Relationship Equity and Loyalty Intentions is positively effected by humor.

Customer Delight

Customer delight is a positive emotional state resulting from having one’s expectations exceeded to a surprising degree (Rust and Oliver, 2000). Customer delight has mostly been measured as an overall construct (Berman, 2005), therefore it could not be missed here as well. According to previous research, there are differences between men and women in their intentions to remain loyal. There can be differences across age groups regarding their loyalty

(17)

intentions too (Melnyk, 2009). Sir Moeken stressed that Trader Joe’s customers can be characterized mostly as highly educated people. Therefore, can be expected that gender, education and age will play a major role with regard to loyalty intentions.

H10 The strength of the relationship between Customer Equity and Loyalty intentions is positively effected by Customer Delight.

H11 The respondents’ gender plays a major role with regard to loyalty intentions.

H12 The respondents’ education level plays a major role with regard to loyalty intentions. H13 The strength of the relationship between Value Equity and Loyalty Intentions for

people under 30 years of age is positively effected by humor.

The research will contribute academically and managerially. Little research has been done so far on the moderating effect of customer delight on loyalty. This thesis will fill that academic gap and add an extra variable to the measurement of ‘customer delight’ as well, namely humor. These findings can be of magnificent relevance for management. If it turns out that the positive moderating effect of ‘customer delight’ is present and if the outcome can tell which variable within customer delight is responsible for that outcome, managers can act accordingly and invest more money and effort on ‘customer delight’ creation.

(18)

4. Method

This chapter will discuss how the online research is executed, the representativeness of the study and will provide information of the sample. The description is given of the measurement of dependent, independent and control variables and how the procedure is executed statistically with a description of the correlation analysis and the use of the PROCESS tool.

4.1 Sample

For the online research of this thesis the research panel of Panelwizard Direct is used. Panelwizard Direct is the research panel of ‘Kien Onderzoek’ and consists of more than 20.000 members of 16 years and older. The panel members have indicated beforehand to be willing to participate in several types of research. The panel members are rewarded for their efforts and are given incentives on a continuous basis. Panelwizard employs the Golden Standard, developed by MOA (Center for information based decision making and marketing research) and CBS (Dutch Central Bureau for Statistics), and can be described as the recognized instrument for samples.

The study is representative for the population of the Netherlands on gender, age, education level, family setting and work participation. With a reliability of 95% can be concluded that the output of the sample will maximally deviate 3.6% from the real situation. This means that the percentages that the research will find are 95% reliable, and won’t in reality be more than 3.6% higher or lower.

The target group for this study consisted of Dutch persons above 16 years of age who are (jointly) responsible for doing the daily groceries. The selection of the sample was made by selecting customers who do their groceries primarily and/or secondarily at Albert Heijn, C1000 and/ or Jumbo for the minimum of a year. The questionnaire was send to 7358 respondents; 2942 persons did not provide cooperation for filling out the questionnaire; 307 questionnaires were deleted after the data check; 1090 persons fell outside the target group because they weren’t Albert Heijn, C1000 or Jumbo customers for more than a year; approximately 294 respondents were bouncers and the quota was already reached before 1970 respondents responded. From the 755 respondents who started the questionnaire, 729 completed the questionnaire (response rate is 58%).

From the 3815 respondents 45.3% were male and 54.7% were female (M=1.55, SD= .49 with 1=male and 2=female). The ages ranged from age group 1 to 5. 16.4% were younger

(19)

than 30, 18.6% were between 30 – 39 years of age, 21.7% fell within the age range of 40 – 49 years, 17.3% were 50 – 59 years and 26.1% were 60 years and older. The education level was also very diverse; with 34.5% of the people with a low education level, 23.7% with an education in the middle and 26.1% were highly educated. The largest part of the respondents did not work (41.8%), 23.7% worked part-time and 34.5% were working on a fulltime basis. In the family settings, the more-person-families without children under 18 years of age were mostly present (45.2%). Followed by the more-person-families with the youngest child under 13 years of age (23.5%) and one-person families (22.5%). Only 8.8% of the respondents were more-person-families with children where the youngest was between 13 and 17 years of age.

4.2 Measurement of variables

Translation-back-translation

Some items used for the measurements derived from English studies. Since the respondents of the questionnaire had Dutch as their first language, those questions were translated in Dutch. To ascertain that the content of the questions didn’t change; the questions were back translated in English by a third person. On the basis of this English back-translation, some Dutch items were adapted accordingly (please see Appendix 3 for the Dutch questionnaire, please see appendix 4 for the English questionnaire).

Customer Equity (CE)

For the measurement of the Customer Equity Drivers – Value Equity, Brand Equity and Relationship Equity – can be relied upon prior studies (Rust et al., 2004).

Value Equity (VE)

Value Equity measures the price/quality ratio (Rust et al., 2004). The item ‘The price-quality ratio is good’ was used to measure VE and could be answered on a 5-point Likert scale ranging from totally disagree (1), neutral (3) to totally agree (5). ‘I don’t know/I don’t have an opinion’ was the sixth option (6). Of the 745 respondents to this question, 3 (.4%) responded ‘don’t know, no opinion’. These responses were deleted from the data set. The ‘don’t know/ no opinion’ -option was added in the answer possibilities to prevent respondents to disrupt the data. In this way, choosing the 1 to 5 options can be interpreted as totally valid.

(20)

Brand Equity (BE)

Brand Equity measures the perceived strength and innovativeness of the brand (Verhoef et al., 2007). The item ‘I would rather like to wear clothing or a bag with the logo of the (supermarket) brand on it, than I would like wearing the logo of another brand’ measures BE (Johnson et al., 2006). It was measured on a 5-point Likert scale ranging from ‘totally disagree’ (1), ‘neutral’ (3) to ‘totally agree’ (5). ‘I don’t know, I don’t have an opinion’ is the sixth option (6). From the 745 respondents, 65 responded (8.7%) ‘don’t know, no opinion’. These responses were deleted from the data set.

Relationship Equity (RE)

Relationship Equity is about perceived commitment, feeling ‘at home’ and connected to the firm (Verhoef et al., 2007). RE is measured by two items (Cronbach’s α= .72), an example item is ‘If the supermarket were a person, I would like to have him or her as a friend’ (Johnson et al., 2006). RE was measured on a 5-point Likert scale ranging from ‘totally disagree’ (1), ‘neutral’ (3) to ‘totally agree’ (5). ‘I don’t know, I don’t have an opinion’ is the sixth option (6). From the 760 respondents, 105 (13.8%) responded ‘don’t know, I don't have an opinion’. These responses were deleted from the data set.

Customer delight

Customer delight is a positive emotional state resulting from having one’s expectations exceeded to a surprising degree (Rust and Oliver, 2000). Positive surprise and joy are seen as the key ingredients in customer delight (Berman, 2005). There is no commonly accepted scale to measure customer delight (Berman, 2005). In this research is decided to measure the constructs ‘joy’ and ‘surprise’ separately. The construct ‘humor’ has not been measured in the context of customer delight yet, but since it plays an essential role in for example viral marketing (Dobele et al., 2007), as well as in the context of TV commercials (Oude Elferink & Dam, 2015), it may also contribute to customer delight and is therefore investigated as a separate construct as well.

Joy

Joy can be described as ‘feelings of pleasantness and happiness’ (and as ‘any ongoing activity which brings an individual into contact, physically and/ or mentally, with some aspect of the world around him/her. It is based on an activity that triggers a feeling of relatedness between a person and a stimulus, e.g. customer and a firm or product’ (Schachtel, 1959). Six items are

(21)

used to measure ‘joy’ in a supermarket setting (Cronbach’s α= .75). Two examples of items are: ‘The supermarket staff takes time for me as a customer’ and ‘The supermarket staff is wearing an outfit that looks nice’. It is measured on a five-point semantic scale ranging from ‘clearly unpleasant’ (1), to ‘neutral’ (3) and ‘clearly pleasant’ (5). ‘I don’t know, I don’t have an opinion’ is the sixth option (6). The option ‘don’t know/ no opinion’ was deleted for the following items Joy1(14) with 728 respondents remaining, Joy2(8), Joy3(3), Joy4(15), Joy5(22), Joy6(10).

Surprise

Surprise is: ‘A neutral and short-lived emotion elicited by either unexpected or misexpected products/services/ attributes, or more precisely, by a schema discrepancy’ (Ekman and Friesen, 1975). Five items are used to measure ‘surprise’ in a supermarket setting (Cronbach’s α =. 77). Two examples of items are: ‘There are regularly new products to purchase in the supermarket’ and ‘At the supermarket it is possible to participate in a cooking or baking competition with several house brand products.’ It is measured on a five-point semantic scale ranging from ‘very negatively surprised’ (1) to ‘neutral, not surprised’ (3) and ‘very positively surprised’ (5). ‘I don’t know, I don’t have an opinion’ is the sixth option (6). The option ‘don't know/ no opinion’ was deleted for the following items; Surprise1 (8), surprise2 (49), surprise3 (36), surprise4 (36), surprise5 (34).

Humor

Humor in this research focuses on the extent of which something makes someone happy or even makes someone smile. It is measured on a five-point semantic scale ranging from ‘absolutely does not make me smile’ (1), to ‘neutral’ (3) and ‘definitely makes me smile’ (5). ‘I don’t know, no opinion’ is the sixth option (6). The option ‘don’t know, no opinion’ was deleted for the following items; humor1(19), humor2(121), humor3(28), humor4(41) and humor5(38). Five items are used to measure humor (Cronbach’s α =0.78). Two example items are: ‘The TV commercials of the supermarket are humorous’ and ‘Several house brand products in the supermarket have funny names’.

Loyalty intentions

Loyalty is the willingness of someone – a customer, an employee, a friend – to make an investment or personal sacrifice in order to strengthen a relationship. Loyalty will be

(22)

measured by using the Net Promotor Score (NPS), the willingness to recommend a product or service to someone else, a measure developed by Reichheld (2003). The item used is: ‘How likely would you recommend this supermarket to your family, friends, acquaintances or colleagues, based on the experience(s) you have had with this supermarket? It is measured on a scale ranging from zero to ten, by 0=not at all likely, 5=neutral and 10=extremely likely.

Control variables

There will be accounted for age, gender, education level, family setting and work participation as control variables. According to previous research, there are differences between men and women regarding their loyalty intentions. There can be differences across age groups as well in their intention to remain loyal (Melnyk et al. 2009).

4.3. Statistical procedure

The respondents were invited for the questionnaire on December 5th_{2014 and on December} 11th_{2014 the possibility to respond was closed. The results were interpreted with the software} Statistical Package for Social Sciences (SPSS).

First counter-indicate items were recoded. Then, there was dealt with the missing values (the ‘Don’t know/no opinion’ option was deleted from the data which is precisely described at the different variables). Computing reliabilities revealed good scales, since all items showed a Cronbach’s α above >.70 (please see reliabilities on the diagonal). Computing scale means formed the total scores of joy, surprise, humor, VE, BE, RE and loyalty intentions. The averages of the items were used to form the variables CE en Customer Delight. After completing those steps the full correlation matrix was established (The results can be found in table 1). By checking the histogram charts and skewness and kurtosis frequencies, the normality of the variables could be observed. The variables VE, BE, RE, joy, surprise and humor are normally distributed. Only the variable ‘loyalty intentions’ is negatively skewed (skewness = -1.55, kurtosis = 5).

Relatively high scores were observed for joy for the item ‘The supermarket staff takes time for me as a customer’; 76.6% of the respondents found it to be pleasant. Respondents were quite to very positively surprised (76%) with the item ‘In the supermarket there are regularly new products to purchase’. For humor (making people smile) the highest score (70.8%) was achieved for the item ‘The TV commercials of the supermarket are humorous.’ Relatively low scores for joy were observed for the item ‘The supermarket is decorated in the

(23)

style of the location/place where it is established’. Only 26.4% of the respondents stated that it would give them a pleasant feeling. More extreme was the score for surprise on the item ‘The supermarket staff gives a performance in the supermarket (for example morning exercises or a live-performance)’. 53% of the respondents were negatively surprised, 24.4% were not surprised and only 16% was positively surprised. Also for the humor measurement the score for one item was extreme. The question ‘In the supermarket a big funny board is present at which you can put through your head and can take pictures’ wouldn’t make 55.8% of the respondents smile, 27.3% would be neutral and only 9.4% would smile. These extreme high and extreme low scores can be explained by the fact that the respondents appreciate what they know, what they have experienced and what they are in some cases familiar with. Taking time for a customer, new products in the supermarket and humorous TV adds are examples of the current reality. Decoration in the style of the location/place where the supermarket is established, a live-performance in the supermarket and a big funny board where you can take pictures are mostly examples customers are quite unfamiliar with. Therefore one could conclude that respondents tend to answer more favourably to situations they are familiar with.

The correlation analyses looks at the interaction between two variables only. If one is interested in measuring the combined effect of two or more predictor variables on an outcome, a different analysis is done. The combined effect of variable X and variable M on outcome Y is conceptually called moderation and statistically known as the interaction effect. This moderation analysis is executed with the help of the PROCESS tool. PROCESS is a computational procedure for SPSS that implements moderation and mediation analyses and their combination in an integrated conditional process model (Hayes, 2012). To test moderation for this thesis, model 1 is chosen. Statistically, the model takes the following form:

Yi= (b0+b1Ai+b2Bi+b3ABi) + Ei

In the analyses all relations possible are tested; as predictor variable customer equity as overall construct, as well as the separate constructs VE, BE and RE. Different moderators are used: customer delight as separate construct, joy, surprise and humor. The outcome variable, loyalty intentions, is the same in every test. The predictor variables are grand mean centred. This means that the variables are transformed into deviations around a fixed point, which

(24)

typically is the grand mean (Field, 2012). This transformation has no effect on the b-values for the combined effect, but it will affect the b-values for the predictors where only one variable is involved (as is the case with b1 and b2). In that case the b-values represent the effect of the predictor when the other predictor is its mean value (instead of the value of zero).

In case of significant moderation, the nature of the interaction will be investigated as well. This will be done using simple slope analysis in which the relationship between the predictor and the outcome at low, average and high moderation levels will be compared (Aiken & West, 1991) and by interpreting the model of Johnson and Neyman (1936) at which the relationship between predictor and outcome is tested at lots of different values of the moderator.

(25)

5. Results

In this section, the general remarking results from the questionnaire will be discussed first. Then a closer look will be given to the correlation analyses: which variables are related and how large is the effect? After that, the hypotheses about moderation will be analyzed in the PROCESS analyses.

5.1. General results

Almost half of the respondents indicated that they enjoyed doing groceries (41% liked it, 5,6% ‘liked it very much’). Only 12,2% didn’t like it (of which 2,9% didn’t like it at all). The rest (40,8%) was neutral. In the open-ended question people could indicate what they liked best about doing groceries. The most mentioned answers can be categorized in four main groups. (1) The respondents like the sales, the special offers, the cheaper prices and the hunt for this in the supermarket. (2) The respondents simply like to seek out tasty food and discover new products. (3) The respondents indicated that doing groceries is a social thing; they like to meet new people, see acquaintances and just be around people. (4) Going grocery shopping gives the respondents’ new ideas, inspiration, they like the atmosphere and entourage; they think it is ‘gezellig’ (translated it means something like ‘pleasant’) and they enjoy just looking around. In the next open-ended question the respondents could also indicate what they thought was most annoying about doing groceries. The two things that were most annoying, far more than other things, were when the supermarket is overcrowded and when you have to wait in line. Other things that were named as being annoying were: out-of-stock products, the fact that doing groceries costs time and money, lugging around the groceries, small and crowded aisles, not being able to find the products and unfriendly staff.

Interesting to indicate as well are the answers on the appreciation questions. The questions about appreciation are more general and different from the specific example questions that measured joy, surprise and humor. In the appreciation questions, respondents were asked to indicate to what extent they appreciated it when the supermarket visit gave them a pleasant feeling (1), whether they appreciated when the supermarket surprised them (2) and to what extent they appreciated it when their contact with the supermarket brought a smile on their face (3). More than four fifth of the people (84.8%) appreciated the pleasant feeling of the supermarket visit. Humor (78.2%) in a supermarket setting was more appreciated than surprise (71.6%). And that is an interesting outcome since the core assets of

(26)

customer delight developed in the literature, are joy and surprise. Humor therefore as well is an important tool to win the appreciation of customers.

5.2. Correlation analysis

The outcome of the correlation analyses can be found in table 1 and the visualization of the relations in figure 2. Just as prior has indicated all variables are related. The results show additional evidence of the positive link between CED drivers and loyalty intentions. Some variables are more related than others. The largest effect can be found for ‘RE and loyalty intentions’ (r=.49, p<.001) and ‘VE and loyalty intentions’ (r=.47, p<.001). Remarkable is the medium effect for ‘customer delight and loyalty intentions’, for which joy (r=.35, p<.001) has the highest effect of the three customer delight variables; humor the second highest (r=.27, p<.001) and surprise the lowest (r=.24, p<.001). It is also worth mentioning that the relation between the customer equity drivers and surprise and humor is large for RE (r=.44, p<.001), medium for BE (r=.31, p<.001) (surprise) and (r=.33,p<.001) (humor) and small for VE(r=.15, p<.001) (surprise) and (r=.18, p<.001) (humor).

Table 1: Means, standard deviations, correlations

M SD 1. 2. 3. 4. 5. 6. 7. 1. Value Equity 3.70 .77 - 2. Brand Equity 2.86 .99 .32** - 3. Relationship Equity 2.95 .99 .37** .75** (α=.72) - 4. Joy 3.65 .55 .31** .30** .39** (α=.75) - 5. Surprise 3.17 .79 .15** .31** .44** .40** (α=.77) - 6. Humor 3.13 .70 .18** .33** .44** .41** .53** (α=.78) - 7. Loyalty intentions 7.27 1.57 .47** .41** .49** .35** .24** .27** -

Note. Reliabilities are reported on the diagonal

(27)

(28)

5.3.Conditional effect: moderation

In the moderating analyses all relations possible are tested; as predictor variable VE, BE and RE. Different moderators are used: joy, surprise and humor. The outcome variable, loyalty intentions, is the same in every test. The hypotheses are visualized in conceptual models (Figure 3, 4, 5, 6, 7 and 8).

Figure 3: H1, H2 and H3

H1 The strength of the relationship between VE and Loyalty Intentions is positively effected by joy. H2 The strength of the relationship between BE and Loyalty Intentions is positively effected by joy. H3 The strength of the relationship between RE and Loyalty Intentions is positively effected by joy.

In hypotheses 1 (Figure 3) is expected that the strength of the relationship between VE and loyalty intentions is positively effected by joy. In PROCESS model 1, VE is the predictor variable (X), loyalty intentions the outcome variable (Y) and joy the moderating variable (M). Table 2 shows the outcome of the PROCESS analyses. The b-values for Joy and VE alone represent the regression of the outcome when the other predictor is its mean value (Hayes, 2012). According to Hayes (2012), it is the interaction effect (Joy x VE) that explains whether moderation has occurred. The interaction effect doesn’t show a significant p-value. Therefore, no support was found for hypotheses 1. The coefficient for the product is -.17 and statistically not different from zero (p=. 16).

(29)

Table 2: Linear model of predictors of loyalty intentions (H1) B SE b t p Constant 7.31** .05 150.59 p < .001 Joy .67** .10 6.47 p < .001 Value Equity .80** .07 11.46 p < .001 Joy x VE -.17 .12 -1.42 P=.16 Note. R2=. 28

**. Significant at the 0.001 level.

In hypotheses 2 (Figure 3) is expected that the strength of the relationship between BE and loyalty intentions is positively effected by joy. In using PROCESS model 1, BE is the predictor variable (X), loyalty intentions the outcome variable (Y) and joy the moderating variable (M). Table 3 shows the outcome of the PROCESS analyses. The interaction effect doesn’t show a significant p-value. Therefore, no support was found for hypotheses 2. The coefficient for the product is -.17 and statistically not different from zero (p=. 11).

Table 3 Linear model of predictors of loyalty intentions (H2)

B SE b t p Constant 7.29** .05 139.28 p < .001 Joy .67** .11 6.17 p < .001 Brand Equity .54** .06 8.37 p < .001 Joy x BE -.17 .10 -1.6 p = .11 R2=. 23

In hypotheses 3 (Figure 3) is expected that the strength of the relationship between RE and loyalty intentions is positively effected by joy. In using PROCESS model 1, RE is the predictor variable (X), loyalty intentions the outcome variable (Y) and joy the moderating variable (M). Table 4 shows the outcome of the PROCESS analyses. The interaction effect doesn’t show a significant p-value. Therefore, no support was found for hypotheses 3. The coefficient for the product is -.21 and statistically not different from zero (p=. 07).

Table 4: Linear model of predictors of loyalty intentions (H3)

B SE b t p Constant 7.3** .05 132.15 p < .001 Joy .55** .10 5.29 p < .001 Relationship Equity .67** .07 9.40 p < .001 Joy x RE -.21 .12 -1.82 P= .07

(30)

H4 The strength of the relationship between VE and Loyalty Intentions is positively effected by surprise. H5 The strength of the relationship between BE and Loyalty Intentions is positively effected by surprise. H6 The strength of the relationship between RE and Loyalty Intentions is positively effected by surprise.

In hypotheses 4 (Figure 4) is expected that the strength of the relationship between VE and loyalty intentions is positively effected by surprise. In using PROCESS model 1, VE is the predictor variable (X), loyalty intentions the outcome variable (Y), surprise the moderating variable (M). Table 5 shows the outcome of the PROCESS analyses. The interaction effect shows a significant p-value. The coefficient for the product is -.23 and statistically different from zero (p<. 05). However, no support was found for hypotheses 4, since surprise in negatively effecting the relationship between VE and loyalty intentions.

b SE b t p

Constant 7.32** .05 149.29 p < .001

Surprise .36** .07 5.11 p < .001

Value Equity .88** .07 12.78 p < .001

Surprise x VE -.23* .10 -2.45 P < .05

*. Significant at the 0.05 level. **. Significant at the 0.001 level.

In hypotheses 5 (Figure 4) is expected that the strength of the relationship between BE and loyalty intentions is positively effected by surprise. In using PROCESS model 1, BE is the predictor variable (X), loyalty intentions the outcome variable (Y) and surprise the moderating variable (M). Table 6 shows the outcome of the PROCESS analyses. The interaction effect doesn’t show a significant p-value. Therefore, no support was found for

(31)

hypotheses 5. The coefficient for the product is -.12 and statistically not different from zero (p=. 08).

b SE b t p

Constant 7.30** .05 136.01 p < .001

Surprise .30** .07 4.09 p < .001

Brand Equity .56** .06 9.10 p < .001

Surprise x BE -.12 .07 -1.75 P=.08

In hypotheses 6 (Figure 4) is expected that the strength of the relationship between RE and loyalty intentions to be positively effected by surprise. In using PROCESS model 1, RE is the predictor variable (X), loyalty intentions the outcome variable (Y), surprise the moderating variable (M). Table 7 shows the outcome of the PROCESS analyses. The interaction effect doesn’t show a significant p-value. Therefore, no support was found for hypotheses 6. The coefficient for the product is -.13 and statistically not different from zero (p=. 10).

b SE b t p Constant 7.3** .05 132.27 p < .001 Surprise .12 .07 1.63 p = .10 Relationship Equity .73** .07 10.60 p < .001 Surprise x RE -.13 .08 -1.63 P= .10

(32)

H7 The strength of the relationship between VE and Loyalty Intentions is positively effected by humor. H8 The strength of the relationship between BE and Loyalty Intentions is positively effected by humor. H9 The strength of the relationship between RE and Loyalty Intentions is positively effected by humor.

In hypotheses 7 (figure 5) is expected that the strength of the relationship between VE and loyalty intentions is positively effected by humor. In using PROCESS model 1, VE is the predictor variable (X), loyalty intentions the outcome variable (Y), humor the moderating variable (M). Table 8 shows the outcome of the PROCESS analyses. The interaction effect shows a significant p-value. The coefficient for the product is -.22 and statistically different from zero (p<. 05). However, since humor is negatively effecting the relationship between VE and loyalty intentions, no support was found for hypotheses 7.

b SE b t p Constant 7.32** .05 144.84 p < .001 Humor (centred) .44** .07 6.13 p < .001 Value Equity (centred) .86** .07 13.01 p < .001 Humor x VE -.22* .08 -2.57 P < .05

In hypotheses 8 (figure 5) is expected that the strength of the relationship between BE and loyalty intentions is effected by humor. In using PROCESS model 1, BE is the predictor variable (X), loyalty intentions the outcome variable (Y) and humor the moderating variable (M). Table 10 shows the outcome of the PROCESS analyses. The interaction effect shows a significant p-value. The coefficient for the product is -.16 and statistically different from zero (p< .05). However, since humor is negatively effecting the relationship between BE and loyalty intentions, no support was found for hypotheses 8.

b SE b t p Constant 7.30** .05 132.90 p < .001 Humor (centred) .36** .08 4.51 p < .001 Brand Equity (centred) .56** .06 9.87 p < .001 Humor x BE -.16* .06 -2.48 P < .05

(33)

In hypotheses 9 (figure 5) is expected that the strength of the relationship between RE and loyalty intentions is positively effected by surprise. In using PROCESS model 1, RE is the predictor variable (X), loyalty intentions the outcome variable (Y) and humor the moderating variable (M). Table 10 shows the outcome of the PROCESS analyses. The interaction effect doesn’t show a significant p-value. Therefore, no support was found for hypotheses 9. The coefficient for the product is -.13 and statistically not different from zero (p=. 0527).

b SE b t p Constant 7.30** .06 125.82 p < .001 Humor (centred) .24* .09 2.76 p < .05 Relationship Equity (centred) .70** .06 11.38 p < .001 Humor x RE -.13 .06 -1.94 P= .0527

H10 The strength of the relationship between Customer Equity and Loyalty intentions is positively effected by Customer Delight.

H11 The respondents’ sex plays a major role with regard to loyalty intentions.

H12 The respondents’ education level plays a major role with regard to loyalty intentions.

In hypotheses 10 (figure 6) is expected that the strength of the relationship between CE and loyalty intentions is positively effected by customer delight. In using PROCESS model 1, CE is the predictor variable (X), loyalty intentions the outcome variable (Y) and customer delight the moderating variable (M). The variables gender, age, education level, family situation and employment situation covariate. Table 11 shows the outcome of the PROCESS analyses. The

(34)

interaction effect shows a significant p-value. The coefficient for the product is -.31 and statistically different from zero (p<. 05). However, no support can be found for hypotheses 10, since customer delight is negatively effecting the relationship between customer equity and loyalty intentions. The variables gender (b=. 34, p<. 05) and family situation (b=-.13, p<. 05) influence loyalty intentions. Age (b=. 03, p=. 44), education level (b=. 09, p=. 17) and employment situation (b=. 08, p=. 20) do not have a significant effect on loyalty intentions. Therefore, support was found for hypotheses 11 and no support was found for hypotheses 12.

B SE b t p Constant 6.70** .31 21.65 p <.001 Customer Delight (centred) .37* .09 3.77 p <.05 Customer Equity 1.01** .07 14.29 p <.001 CD x CE -.31* .09 -3.36 p < .05 Gender .34* .10 3.29 p < .05 Age .03 .04 .78 p=.44 Education level .09 .07 1.36 p=.17 Family situation -.13* .05 -2.54 p<.05 Employment situation .08 .06 1.26 p= .20 R2=. 35

5.4.Conditional effect: moderation for specific groups

In this paragraph will be investigated if significant interaction relationships for specific groups can be discovered that are positively related to loyalty intentions.

Figure 7: H13

H13 The strength of the relationship between VE and Loyalty Intentions for people under 30 years of age is positively effected by humor.

(35)

For the first specific moderation test, only the respondents under 30 years of age were used, which resulted in a sample size of n=106. In PROCESS model 1, VE is the predictor variable (X), loyalty intentions the outcome variable (Y) and humor the moderating variable (M). Table 19 shows the outcome of the PROCESS analyses. The b-values for humor and VE alone represent the regression of the outcome when the other predictor is its mean value. The interaction effect shows a significant p-value. The coefficient for the product is .5352 and statistically different from zero (p<. 05). This means that the relationship between VE and loyalty intentions is positively moderated by humor.

Table 19: Linear model of predictors of loyalty intentions

B SE b t p Constant 7.31** .13 62.39 P<. 001 Humor (centred) .11 .15 .70 P=. 48 Value Equity 1.02** .23 4.47 P<. 001 Humor x VE .5352* .16 3.32 P<. 05 R2=. 33

The interaction effect of BE and humor on loyalty intentions for the respondents under 30 years was not significant (b=. 063, p=. 72). As well as a non-significant effect for the interaction effect of RE and humor on loyalty intentions (b=.33, p=.12). The interaction effect between surprise and VE on loyalty intentions for this age group was not significant (b=-.03, p=.90), just as for surprise and BE (b=.09, p=.54) and surprise and RE (b=.12, p=.55).

For the second specific moderation test, only the respondents with a high education level are used, which resulted in a sample size of n=189 for VE, n=181 for BE and n=169 for RE. No significant relationship between VE, BE, RE and joy, surprise and humor could be found.