The gamification on brand engagement
All fun and games or a perverted marketing
trick
Bas van der Lip10815996 basvdlip@gmail.com
Page | 2
DECLARATION OF ORIGINALITY
I hereby certify that I am the sole author of this thesis and that no part of this thesis
has been published or submitted for publication.
I certify that, to the best of my knowledge, my thesis does not infringe upon anyone’s
copyright nor violate any proprietary rights and that any ideas, techniques, quotations, or
any other material from the work of other people included in my thesis, published or
otherwise, are fully acknowledged in accordance with the standard referencing practices.
Furthermore, to the extent that I have included copyrighted material that surpasses the
bounds of fair dealing within the meaning of the Canada Copyright Act, I certify that I have
obtained a written permission from the copyright owner(s) to include such material(s) in my
thesis and have included copies of such copyright clearances to my appendix.
I declare that this is a true copy of my thesis, including any final revisions, as
approved by my thesis committee and the Graduate Studies office, and that this thesis has
not been submitted for a higher degree to any other University or Institution.
Page | 3
Contents
Abstract ... 4 Introduction ... 4 Literature review ... 6 Gamification ... 6 Brand engagement ... 7 Game design ... 8 Demographic differences ... 10 Conceptual model ... 15 Methodology ... 15 Data collection ... 16 Variables ... 17 Data analyses ... 21 Results ... 21 Correlations ... 21 Mediation ... 23 Moderation ... 25 Discussion ... 28 Limitation ... 33 Conclusion ... 33 Bibliography ... 34 Appendix ... 38Page | 4
Abstract
This research investigates the effect of gamification on brand engagement. Gamification has already been proven useful for getting people engaged on a subject and has great potential for marketing, but evidence for using it effectively as a marketing tool is still scarce. Analyses were done based on data obtained from the user of the ‘Albert Heijn Moestuintje’. Next to gamification the effects of age and gender were studied. And also there was investigated after the potential explaining variables achievement stimulation and social stimulation. The result showed that gamification has not a significant effect on brand engagement. Thus for gamification to be effective it should be applied in another way.
Introduction
The concept of gamification has been in society for a long time. The idea of using a game for tricking the mind to let a boring task seem fun is nothing new. However, it is only since 2008 that the term ‘gamification’ exists (Deterding, Dixon, Khaled and Nacke, 2011). Deterding et al. (2011) describe gamification as “the use of game design elements in non-game contexts”.
Although the ideology of gamification is popular (Kim, 2015), there is not a lot of literature on the topic at the moment (Conaway and Garay, 2014). Also Huotari and Hamari (2016) agree that the popularity of gamification enhances, but research connecting it with marketing is scarce. The connection between gamification and marketing is already made by Hofacker, De Ruyter, Lurie, Manchanda and Donaldson (2016). However, they claim it is not implemented right and therefore far from reaching its true potential. Gamification has influence on customers’ emotional state and because of that, it could well be used to create brand engagement (Kankanhalli, Taher, Cavusoglu and Kim, 2012).
Despite the research being limited, gamification has a greater impact on modern society every day. Even so, Jesse Schell, game developer, believes that the ultimate goal of gamification is to be implemented so far that nothing seems as an obligation anymore. There will only be things you want to do, because games should be fun. In his hypothetical future there is no living without games as a fair part of society and there is no coming back from. According to the predictions of
Gartner.com, gamification is only at the peak of its expectations and still has a long way to develop. According to Kankanhalli et al. (2012) gamification will make its way as a marketing tool to get customers more brand engaged. To have a big group of brand engaged customers is important. It creates an ingroup, a social group to which a person identifies itself as being a member, and that makes customers feel more connected with the brand (Escalas and Bettman, 2005). According to
Page | 5 Sprott, Czellar and Spangenberg (2009) brand engaged customers have more knowledge and
preference for a brand and also more attention and loyalty to a brand. For these reasons, finding an effective way to get customers engaged with the brand is important.
As mentioned before, gamification is becoming an increasingly bigger part of our society. It has already proven its usefulness as a learning tool. People become engaged and therefore more motivated (Muntean, 2011). However, it is believed that it could also be used as more than a learning tool (Huotari and Hamari, 2016). The researches of Muntean (2011) and Barata, Gama, Jorge and Gonçalves (2013) describe that gamification is an effective concept to get people more engaged with a subject. According to Zhang (2008) and Jung, Schneider and Valacich (2010) this is because a well-designed computer interface can affect the emotional state of the user.
Emotional connection is one of the fundamental aspects of brand engagement (Hollebeek, 2011). Brand engagement is the connection between a brand and its customer on a cognitive, emotional and behavioral level. Having brand engaged customers results in customers that are loyal to the brand and harder to dissatisfy (Sprott, Czellar and Spangenberg, 2009).
However, because of the lacking knowledge there has yet to be done a lot of research on the topic. Specifically connecting gamification to the marketing strategies is not saturated in the research field (Hamari and Järvinen, 2011; Sigala, 2015; Stenros and Sotamaa, 2009). Based on prior literature it could be assumed that gamification would work as a tool for brand engagement. Though, empirical evidence is scarce and researchers have not yet reached a consensus about the optimal game design framework. Hamari and Järvinen (2011) investigated which game design gets the strongest customer relationship. They compare different game designs and their value offering. But they do not connect the game to general engagement with a brand.
Despite disagreements on an optimal method, researchers on the topic of game design agree that gamification is a valuable tool in society and is far from reaching its potential and how it can be implemented as a useful marketing tool (Deterding, 2012).
In this research the focus is on the use of gamification to increase brand engagement. Differences in levels of engagement will be compared between customers of the same brand that use gamification and customers that do not and thus determined if it has an effect. The brand used for this research is ‘Albert Heijn’, a Dutch supermarket, with the gamification part ‘Albert Heijn Moestuintje’. The ‘moestuintje’ is a plant out of a kitchen garden. The customer gets a ‘moestuintje’ if they spend 15 euro on groceries. The customer gets the seeds and ground and needs to grow the plant himself.
According to the existing literature gamification has an impact on emotions, which are necessary to create brand engagement. Yet, a specific correlation between the two is missing.
Page | 6 Therefore, the research question of this thesis will be: ‘Has gamification an impact on a customer’s level of brand engagement?’
Literature review
Gamification
Juul (2010) describes a game as “a rule-based formal system with a variable and quantifiable
outcome, where different outcomes are assigned different values, the player exerts effort in order to influence the outcome, the player feels attached to the outcome, and the consequences of the activity are optional and negotiable” (p. 8). It gets people engaged. Their full attention is focused on a
subject. Because it is so captivating, it will not come across as boring. This makes it useful for interfering with boring tasks or goals. However, games are not the same as gamification (Kapp, 2012). Gamification is usually a part of a game. It contains elements of a game whereby there is no need to play the entire game. More on this in the section ‘game design’.
Page | 7 with technology, applications for children to learn things have become lucrative (Muntean, 2011). For example, ‘Dragonbox Algebra’: a mathematics application that teaches children to solve riddles in equitation form. The goal is still to find the ‘x’, but instead of a formula it becomes a game. Because of this, children do not even realize that they are learning mathematics.
Brand engagement
Brand engagement focuses on the emotional commitment of a customer to a brand. According to Van Doorn, Lemon, Mittal, Nass, Pick, Pirner and Verhoef (2010), it is not just the transactions a customer makes, but how the customer feels towards the brand. Hollebeek (2011) describes in her research the importance of brand engagement and how it can be accomplished. She claims that brand engagement can be reached on a cognitive, emotional and behavioral level. Gamification responds to the emotional level of brand engagement. If customers are engaged on these levels, it leads to trust, commitment and customer satisfaction. These factors determine the quality of the relationship. If the relationship is of high quality, it translates into brand loyalty.
According to Sprott et al. (2009) brand engagement leads to a favorable attitude toward the brand. For example, this causes that customers are less reluctant if prices increase or when they have to wait longer on a new product. People tend to react better on negative actions or changes in the brand. It also makes customers less likely to switch brands, even though there can be better options. Sprott et al. (2009) claim that brand engagement can be measured by how customers view
themselves as individual. If part of this view is an important brand in their life, then the engagement is very strong. The stronger the engagement, the more it will be shown in a customer’s behavior (Van Doorn et al. 2010). Gamification engages people with subjects, and because of that it could be used as a marketing tool. Therefore the main hypothesis of this research will be:
Hypothesis 1 = H1: Gamification used as a marketing tool will increase the level of brand engagement of the customer.
Algesheimer, Dholakia and Herrmann (2005) describe brand engagement via a brand community. Members of the community perceive similarities with other members and dissimilarities with non-members stronger. This identification with other non-members develops an emotional connection with the group and therefore the brand. This makes customers more agreeable to the norms of the group (Bhattacharya, Rao and Glynn, 1995). The formation and controlling of a community are also factors that could be influenced by gamification. The social part of a game design, as in competing to your
Page | 8 friends or other users, can create a community feeling (Ducheneaut and Moore, 2004). Concluded out of the research of Algesheimer et al. (2005), gamification can contribute to brand engagement through creating a community feeling. Therefore the mediator ‘social stimulation’ will be tested, with the following hypotheses:
Hypothesis 2a = H1: Being part of a social community makes people more engaged with a brand. Hypothesis 2b = H1: The impact of gamification on brand engagement is mediated via social feeling.
Game design
As described in section 1 of the introduction, gamification is the application of game elements in a non-game environment. This is done to give emotional and social benefits to a certain situation, often to make it more fun. Without the fun element it does not add value according to Barata, Gama, Jorge and Gonçalves (2013) and therefore does not improve the brand engagement. A well thought out game design is crucial for a game to be interesting and thus add any value that gets people engaged (Hofacker, De Ruyter, Lurie, Manchanda and Donaldson, 2016). However, there is no scientific consensus on the optimal game design to get people engaged with a brand (Mora, Alberto, Riera, 2015).
A game has influence on a player because they feel attached to the outcome (Juul, 2003). This attachment is formed because a player puts effort in the game. The effort can make a difference in the outcome. A wanted outcome gives a higher value, whereas an unwanted outcome gives a lower value. Being able to have an impact on the outcome guided by a rule-based formal system is what makes it fun for a player. Huotari and Hamari (2012) claim that the quality of the game strongly depends on the game experience. According to them, the game experience should contain
challenging and suspenseful experiences for the players.
Huotari and Hamari (2012), but also Zhang (2008) and Jung, Schneider and Valacich (2010), use the term ‘affordances’. The affordance is what the environment provides (Gibson, 1979, p. 127). The above mentioned researchers claim that motivational affordances are important in a game. It is a stimulus that affects the players’ psychological state, which makes the players more attached to the game. Being attached to the game means that players will use the game elements voluntarily instead of seeing it as something they have to interact with. According to Hamari and other
before-mentioned researchers, it all comes down to the psychological states and emotions where the affordances have influence over. Being able to affect the players on an emotional level creates value. Therefore, with Albert Heijn Moestuintje as gamification, Albert Heijn should be able to create value for the customers.
Page | 9
Achievement stimulation
Part of creating value can be reached by giving the customer the possibility to achieve something. Ryan and Deci (2000) discuss the self-determination theory. It describes the motivation that people have when they make certain choices. For this research it means that the motivation behind growing a ‘moestuintje’ will be partly explained. There are three components that drive the decision-making: competence, relatedness and autonomy. The focus in this research is on the component
‘competence’. According to Ryan and Deci people have competence needs. Positive feedback on their actions motivates them to do more of those tasks. This is part of the intrinsic motivation. Intrinsic motivation is the drive to seek new, more difficult tasks and to challenge oneself.
Accomplishing new challenges and receiving positive feedback fulfills the need for competence. For the ‘moestuintjes’, the challenge is to let them grow well and to create a fulfilling feeling of the result. However, Ryan and Deci claim, the perceived difficulty of the challenge will play part in the feeling of achievement a person can get off it. For example, someone that has grown lots of ‘kitchen gardens?’ will not feel very achieving after growing a ‘moestuintje’. However, someone that grows a plant for the first time will see it as quite an accomplishment.
Next to intrinsic motivation, there is extrinsic motivation. Extrinsic motivation is motivation from external sources, a reward for the effort one puts in a task. In the case of a ‘moestuintje’ it would be the fruit one can enjoy if it grows well. In this case the extrinsic motivation is not strong of value, so it can be seen more as an intrinsic motivation because people can see the creating a fruit as an achievement. It would be an extrinsic motivation if someone would sell their plants. However, the extrinsic reward can diminish the effect of the intrinsic rewards, because the task now feels like something that has to be done instead of a voluntary task (Ryan and Deci, 2000).
In the research of Zhao, Jia and Maes (2018) differences were measured in response time for various levels of achievement motivation. Groups of high achievement motivation and low
achievement motivation were used to compare the level of the congruency sequence effects. This technique of measurement has been used by Duthoo, Abrahamse, Braem, Boehler and Notebaert (2014) to observe the association between achievement motivation and response time to a
congruence sequence. To observe the difference in congruency sequence effects the ‘Eriksen flanker task’ was used (Duthoo et al. 2014). Differences in reaction times are measured to various sequences of images that request a response based on congruency. High achievement motivation groups responded faster when the same sort of images followed each other up, meaning that there was a
Page | 10 positive association between achievement motivation and congruency sequence effects. According to Zhao et al. (2018) it could be explained by the positive emotion felt when having fast responses to earlier images. That indicates that once a participant had success before, it was more eager to get success again. For the ‘moestuintjes’, this signifies that if a customer feels the positive emotion of achievement when a plant is grown to its full potential, the customer will want to feel that positive emotion again and is likely to try to grow another ‘moestuintje’.
Weiner (1985) argues that the motivation received by accomplishing a task is positively correlated to the causality someone believes to have in the outcome and the effort that is put in. The effort, in combination with an outcome that can be influenced, affects the self-image someone has. People strive to have a positive self-image. However, failing a task that could have been
accomplished has negative consequences for the self-image. For that reason a person is more likely to engage in a task when they are more certain that it will have a positive outcome. However, Atkinson (1957) claims the difficulty and outcome uncertainty have an influence on the positive feelings that are gained if the task is completed. A more challenging task will give greater reward. The book of Heckhausen and Heckhausen (2008) clarifies that the self-image already starts with young children. Their facial expressions changed with the different outcomes of building a tower of building blocks. There appeared a smile for achieving success and a sad face for failure. The principle is the same for adults. When they succeed in growing a ‘moestuintje’ they will smile and if not with facial expressions have a positive feeling about it, because it enhances their self-image. The strength of the effect on self-image is dependent on the value that the person has given to growing a ‘moestuintje’. The claims of Heckhausen and Heckhausen (2008) agree with the self-determination theory. Their need for competence is shown and the effect on self-image it has.
Based on the theory in this section it is expected that achievement stimulation has an effect on the level of engagement people have. Also it is expected that achievement stimulation can be reached through gamification. Therefore the following hypotheses will be tested:
Hypothesis 3a = H1: A sense of achievement makes people more engaged to a subject.
Hypothesis 3b = H1: The impact of gamification on brand engagement is mediated via a sense of achievement.
Demographic differences
Page | 11 Hoffman (1972) concludes in his research that men are more focused on working to accomplish and have more achievement needs in comparison to women. The researches of Koivisto and Harmani (2014) comply with the result of Hoffman (1972). Additionally, they find that women are more focused on interpersonal factors. Women would be motivated more by affiliation needs than men (Hoffman, 1972). Carlson (1972) too, claims that women are more concerned to be in line with others, whereas men tend to act more on themselves. The focus of women is, because of that, more on interpersonal relationships than reaching an objective. For this research it signifies that women would be more motivated by the mediator ‘community feeling’. Because of the focus on
interpersonal relationships, women would be more likely to thrive well in a community and have a greater urge to share their achievements. That would indicate that it is not the achievement that drives the motivation, but the community in which a woman can share her achievements.
The research of Miller (1986) is in line with these expectations is. According to Miller, women consider pleasing others more important than men. This is also a factor that suggests that women would be more driven by a community feeling than men. Women would be more susceptible to helping others instead of only focusing on their own achievement. Pease and Pease (2003) agree with Miller (1986). According to them, men find things as prestige, power and ownership important. They define themselves by what they have accomplished and what kind of job they have. Women on the other hand, value meeting interesting people and being of service to others. They care about the quality of their relationships. The findings of Yee (2006) confirm the assumptions of Pease and Pease (2003) and Miller (1986) with his research. Women value relationships. Measured on three social aspects, women appear to focus more on relationship as subcomponent of the social part. However, on the overall social aspect, men and women scored evenly. This means that within the social part, specifically the relationships are valued. The achieving component was very obvious nonetheless. Men scored higher on all achieving subcomponents, signifying that men value accomplishing a task more than women.
Next to that, Becker (1986) suggests that women are easier influenced by others, whereas men tend to follow their own opinion. According to Minton, Kagan and Levine (1971), based on their parenting behavior, women are more compliant than men. This could result in higher scores for women on the social compartment. Men would be driven to make their own success, while women can be persuaded in growing a ‘moestuintje’ by others. For this research it signifies that men should be more motivated through the mediator ‘sense of achievement’ and women would be more motivated through the mediator ‘community feeling’. They will more often collect a ‘moestuintje’ together with someone or for someone. This would also signify that they are more understanding of others collecting ‘moestuintjes’ and will talk about them more often.
Page | 12 have a clearer guideline in the technology branch, because they value the ease of use of something. For the ‘moestuintje’ this could indicate that women are more prone to give up if a ‘moestuintje’ is unclear how to use or fails. Therefore they would be less likely to feel the same affection for it as men would.
Oly Ndubisi (2006) researches the difference in customer loyalty between men and women and came to a similar conclusion. Women tend to be more loyal than men when they have trust in a brand. However, it is only so when women have trust in a brand. At an earlier stage, where the customer is not yet very familiar with a brand, men score higher on loyalty. When trust is gained over time, women tend to score significantly higher on loyalty (Oly Ndubisi, 2006). Since Albert Heijn is a long existing brand and almost every person living in the Netherlands has been in touch with it, this research assumes that all participants have trust in the brand. This would mean that women would be more loyal customers to Albert Heijn than men and should have a higher level of brand
engagement, especially the social aspect.
From theory it can be concluded that the different genders become engaged via different factors. However, it is expected that the engaging factor for women will be stronger, because the achievements that can be reached with a ‘moestuintje’ are limited. The social component can be expanded to great extent and therefore will have a bigger impact on the level of brand engagement. Based on that, the following hypotheses have emerged:
Hypothesis 4a = H1: When familiar with a brand, women are more brand engaged than men Hypothesis 4b = H1: The effect of gamification on brand engagement will be stronger on women.
Age
The other demographic difference is ‘age’. According to Koivisto and Hamari (2014) gamification is more likely to have an effect on younger people. This is due to the digital divide of the generations. Older people did not grow up with technology and have more ‘computer anxiety’; hence they find it less useful. Younger users value the usefulness of technology more because they understand it better. It is possible that the older users will miss sense of identification with a community and therefore be less prone to engagement.
Page | 13 In the research of Venkatesh, Morris, Davis and Davis (2003) it becomes clear that younger workers have a greater need for extrinsic rewards. So it would not necessarily be the understanding of technology that distinguishes the younger people from the older people, but a greater drive to win. Especially on the achievement based aspects, young people would be more motivated to win and obtain an extrinsic reward. Steinberg, Albert, Cauffman, Banich, Graham and Woolard (2008) confirm the findings of Venkatesh et al. (2003). They too claim that younger people are more reward seeking than older people. Being rewarded can be their sole motivational driver. Casey, Jones and Hare (2008) even believe that the increase in reward seeking behavior during pubertal maturation is biologically programmed. It would be a stimulus for mating. This could indicate that growing a ‘moestuintje’ successfully is a strong motivator for the younger users and will therefore be quicker to repeat the process and obtain that same rewarding perception.
However, extrinsic rewards do not have to be the only catalyst for gamification. Another stimulant is building relationships and becoming part of a social network. Rhodes (1983) declares that the older workers in her research are more resolute in the social part of the game. As workers became older, their need for affiliation was increased in the work environment. But this same craving for partnership could exist in the rest of their life, or as in this case, gamification.
Morris and Venkatesh (2000) supplement the findings of Rhodes with their research on age differences in the work environment. Older workers tend to feel a greater need to please others. Therefore chances are greater that they would conform to the opinions of the majority. Next to that, Morris and Venkatesh (2000) find in their research that younger workers have a higher need for autonomy. That could suggest that older workers rather have a positive relationship with their colleagues and superiors whereas younger workers have less affiliation with a friendly work environment (Rhodes, 1983).
According to Levy (1988), the need for socialization increases for women when they are aging because of the decline in physical appearance. Once women start to lose a great physical appearance it would become harder for them to maintain a social network, more so than for men. Therefore socialization for older women becomes extra important. This indicates that, especially women, will become more affected by the social part at an older age. The social part of gamification could help them in maintaining relationships or building new ones. This is the same result as Cumming and Henry (1961) found. Elderly are likely to experience a decrease in their social network. Staying connected through a community could be very helpful for them.
Ijselsteijn, Nap, De Kort and Poels (2007) find that the successful performance on the various tasks and subjective preference are positively correlated. This signifies that for the elderly, things that are easier to use have a greater preference. Thus an upcoming era in digital games is not helpful for them. Understanding new technology takes time, causing a less preferable attitude toward it.
Page | 14 According to Ijsselsteijn et al. (2007) this would be an opportunity not used to its full potential, because playing games is a great option for elderly to socialize outside their network. For them it is harder to get in contact with new people than younger people. Socializing through gaming is therefore a great opportunity to connect with others.
However, Chung, Park, Wang, Fulk and McLaughlin (2010) claim that age is not a moderating factor on use of online communities. The factor that is detaining older people from the use of communities, is the ease of use of those online communities. Building relationships based on ‘moestuintjes’ does not have to be necessarily online. This decreases difficulties that older people experience, because the online aspect, their vulnerability, falls due. Thus the effect of this factor would not be as strong as when an online community is the main socialization network. Based on the theory, it is expected that achieving motivation of younger people is stronger than the desire for a social network. Therefore the following hypothesis will be tested:
Hypothesis 5 = H1: The effect of gamification on brand engagement will be stronger on younger customers.
Page | 15
Conceptual model
Brand engagement is important for a brand. Based upon the current knowledge, gamification supposedly has the potential to make a great impact on the brand engagement, making customers feel more connected to the brand. If it can be implemented well, it could be of high value for a company’s marketing strategy. Therefore the following main hypothesis will be tested: ‘Customers that take part in the gamification of a company are more engaged with the brand and will therefore have greater loyalty to it’.
Methodology
Through a deductive approach, this research observes the difference in brand engagement of customers that use ‘Albert Heijn Moestuintje’ versus customers that do not. It is important to clearly establish the level of brand engagement to test the hypothesis. Therefore, the level of brand
engagement will be defined through the use of surveys. Surveys are being used because they collect Moderator: Age Moderator: Gender Independent variable: Use of gamification Dependent variable: Level of brand engagement Mediator variable: Community feeling Mediator variable: Sense of achievement
Page | 16 the necessary data and are fit to make a great sample size. First, some demographic data will be asked to categorize the participants so the effect of the moderators can be measured. Then the participants will give their opinion on the brand in form of quality, price and overall. After that, it will be measured how much of their groceries the participants get at Albert Heijn in comparison to other
supermarkets.
A sample of Albert Heijn customers will be used to generalize the research on brand engagement in any gamification aspects of a brand. The ‘gamified’ part of Albert Heijn is used to measure the effect of gamification on brand engagement. The reliability of the research is partly dependent on the response rate of the survey. According to Armstrong and Overton (1977) it is possible that people that respond are different from people that do not respond. For this research it could mean that mainly customers that are already more engaged will respond, hence the reliability decreases.
Data collection
Participants will be accessed through online platforms and approached in public locations. Via these methods it is expected to get the most participants, therefore a big population with a lot of variety can be reached (Wright, 2005). A part of the participants will be accessed through general social media platforms. This is because there needs to be a group of people that do not use the Albert Heijn ‘moestuintje’ to compare with. If all the online responses come from a specialized ‘moestuintjes’ platform, chances are that the data has a low external validity. The other participants will be accessed on a platform designed for Albert Heijn moestuintje users, the Facebook group of Albert Heijn moestuintjes. This way there will be sufficient participants for both groups. Answering the questions will be based on the Likert scale, a rating scale from 1 to 5, so the questions are easy to answer for the participants. It will also be useful to categorize participants in their level of
engagement, since this is measured with the same scale.
In their research, Huotari and Hamari (2017) claim that gameful experience is subjective. So everyone experiences it different. To control potential demographic differences, ‘age’ and ‘gender’ are applied as moderating variables. This is because, as described in the literature review, it is possible that these factors have an effect on the outcome. The collected data will be analyzed using descriptive statistics in SPSS, so the hypothesis can be rejected or assumed to be true on the basis of a significant p-value.
The sample size is expected to be relatively small. The willingness of people to fill in surveys is not great in this era and the data needed to be collected in a relatively short period of time. Because of a small sample size the external validity can be compromised. Next to that the sampling is
Page | 17 also somewhat convened, because the survey was also placed in a Facebook group for
‘moestuintjes’. The presence of participants on such platform could indicate that they are already more engaged than the neutral participant, but not necessarily through gamification.
Variables
There are six main variables applicable. The first variable is the level of engagement the customer has with the brand, this is the dependent variable. This will be measured with the Likert scale, so there will be more clearly defined levels of engagement and different levels can be compared. There will be five different levels of engagement. 1=strongly disengaged, 2=disengaged, 3=neutral 4=engaged and 5=strongly engaged. Distinction between the levels of engagement has been made asking the questions in Likert scale form as well.
Brand engagement is not a tangible asset, so it cannot be measured exactly. It is, however, within reach to analyze behaviors of people. Based on the researches of Sprott et al. (2009) and Hollebeek and Chen (2014) certain behavioral threats can be determined as a base line for brand engagement. One of the behavioral threats is the attitude towards the brand, according to Sprott et al. (2009). Their research shows a significant relation between a positive brand attitude and a positive attitude towards new products of that brand. This would mean that people that are more brand engaged will probably tend to have a more positive attitude about the products of that brand. To measure the attitude toward the brand, the participants are asked if they think Albert Heijn is overall the best supermarket for them. The attitude toward Albert Heijn is measured, but Albert Heijn is also compared to other supermarkets.
The next behavioral threat is ‘price insensitivity’. Sprott et al. (2009) conclude based on their results that customers that are more engaged are less sensitive to price changes. This means that more engaged customers could be partly defined by their willingness to still buy the product of a brand if the price went up. For that reason the participant is asked if he is stilling willing to go to Albert Heijn for the groceries when the prices of Albert Heijn increase. If the participant is still willing to go to Albert Heijn, he is not price sensitive and therefore more engaged with the brand. If the participant will immediately go to a competitor of Albert Heijn it means that they are price sensitive when it comes to Albert Heijn, and are therefore less engaged with the brand.
Another behavioral threat is according to Sprott et al. (2009) the brand possession. A person that possesses a lot of products from a single brand is more likely to see that brand as part of who he is. It creates self-extension through the brand. Groceries are however not a product that are in someone’s possession for a long time, because it is edible. The share of groceries coming from supermarket can be measured however. For that reason the participants are asked how much of their total groceries comes from Albert Heijn.
Page | 18 The last behavioral threat is the perceived quality. Hollebeek and Chen (2014) argue that it reflects the perceived utilitarian value it has to a customer, in other words the functionality of the brand. An engaged customer will rank the quality of the brand higher because he finds the brand very useful. Hence the participant will be asked about the quality of the service and products of Albert Heijn.
The second, independent variable, is the distinction between customers that use the gamification and customers that do not. With 0=no use and 1=use it is also a dummy variable. A correlation between variables 1 and 2 should confirm the hypothesis. The gamification used in this research is the ‘moestuintje’ of Albert Heijn. People that have ever received a ‘moestuintje’ are users of the gamification of Albert Heijn.
Moderators
The third variable is ‘gender’, a moderator. It is also a dummy variable, with 0=women and 1=men. It is expected that gamification will have a bigger impact on men due to them being more competitive. This would also mean that men that used a ‘moestuintje’ should score higher on one of the later explained mediator variable ‘sense of achievement’. Women, however, should score higher on the other mediator variable ‘community feeling’ since they value relationships more. This also means that the effects of the different genders should partly even each other out, because the sexes have different parts of the gamification that gets them engaged.
The fourth variable is a moderator too, ‘age’. Like the other moderator, this variable has suggestions that it could strengthen the correlation on both sides. On one hand it is expected that gamification has a bigger impact on younger people. They are more focused on achieving. As Venkatesh et al. (2003) claimed, younger people have a greater need for extrinsic motivation and a stronger drive to win in comparison to older people. Because of their drive for achievement they would be more engaged in the game design and therefore easier subjective to gamification. Next to that it is expected that younger people will make more use of gamification overall, because it involves technology in many cases. For this research however this is not the case. Older people would however be more thriving for intrinsic motivation. They have a greater need for relationships. Therefore they would be engaged through a social feeling that can be created through the
gamification.
The data will be categorized in three groups: young, middle-aged and older people. The groups get the following values in SPSS: 1 = young, 2 = middle-aged, 3= older. The group with young
Page | 19 people is from age 0 to 25. The middle-aged group is from the age 26 to 40 and the group with older people is everyone above the age of 40. The different groups all represent a different phase of life: youth, new adults and old adults. Especially the outer groups have a different look on the possible affecting variables. The expected differences are between younger and older people, but by dividing the data in three groups a more clear effect, if any, can be shown because the boundary of young and old people is quite vague. Age is after all an ever changing concept. By creating a middle group there will be a clear difference between young and old people and it could display an increasing or decreasing line through all groups. The boundaries on the groups are made based on more or less equal group sizes so the sizes will still be big enough for a possible significant result.
Mediators
The fifth variable is a mediator, ‘social stimulation’. A part of the increased brand engagement could be due to a community feeling. It is not as much the gamification part that would increase the
engagement, but the feeling that is created by being part of the group that plays the game. The social factor could make people feel that they belong to a brand. To measure this, people are divided based on their likeliness to socially interact with other people about the brand. This variable is divided into levels of social stimulation. Levels differ from 1 to 5 based on the Likert scale. To determine the level of social stimulation four questions in the survey were dedicated to this subject, all in form of the Likert scale. The scores were added up and divided by four to get a number that determined in which category a person can be divided.
The level of social stimulation is based on how connected people feel with Albert Heijn. According to Bhattacharya et al. (1995) it is important for customers to be able to identify themselves with a group, and in this case a brand, Albert Heijn. Due to the emotional connection that is formed the customer will be more agreeable to the norms of the group. The question on how connected a customer feels with the brand shows to what level the customer can identify itself with the group.
Another point of measurement is the level of collective collecting. The question measures to what extend the customer collect the ‘moestuintjes’ for itself or for others. This also displays level of engagement with a social community that focuses on the gamification of Albert Heijn. According to Algesheimer et al. (2005), this community feeling should cause a higher level of brand engagement. The next point of measurement is also to measure to what extent the customer can identify itself with the community. It is asked in to what level the participant can identify with others obsession for the ‘moestuintjes’. The last question to determine the level of social stimulation is to what extend the customer talks to others about the ‘moestuintjes’. Being occupied with
Page | 20 social need to share information about the activity, which means that the customer likes to be part of the social community.
The sixth and last variable is a mediator too, ‘achieving stimulation’. It is possible that a part of the effect of gamification can be explained through the achieving stimulation. The achieving factor gets people engaged with the game and therefore they start to feel more in touch with the brand that designed the game. For this mediator variable there were also four questions in the survey based on the Likert scale. Here too, the scores of the questions were added up and divided by four to determine the level of achieving stimulation which are also 1 to 5.
To establish the level of achieving stimulation, the feelings, goals and actions of the customer are questioned. The first question is about how the customer feels if the ‘moestuintje’ grows well. If the ‘moestuintje’ grows well, the task can be considered as accomplished. According to Ryan and Dece (2000) this should give a positive feeling and an improvement to the self-image.
Also the feeling of failure is measured. In the following question the participant is asked about the feeling when a ‘moestuintje’ does not grow well. In this case it can be considered a failure and it would be destructive for the self-image. As Heckhausen and Heckhausen (2008) explain with their example of children constructing towers with building blocks, failing a task leads to a negative feeling. Because failing the task is bad for the self-image, the participant is less likely to try it again, because the outcome certainty for a good outcome is low.
Another point that is measured is the goal of collecting all the types of ‘moestuintjes’ (five different types). That the customer wants to collect more ‘moestuintjes’, suggests that he
experienced a positive feeling from earlier collected ‘moestuintjes’. It also represents a more difficult challenge. Growing five ‘moestuintjes’ well is harder than growing one. Also because of the chance to not grow well each ‘moestuintje’ has, the outcome of growing all five ‘moestuintjes’ well becomes less certain. According to Atkinson (1957) a harder task with more outcome uncertainty gives a greater positive feeling when achieved.
The last point of measurement is the willingness to collect a new ‘moestuintje’ when an earlier ‘moestuintje’ grows well and can be considered as an accomplishment. The positive feeling it gives makes the customer consider collecting a new ‘moestuintje’, because that same positive feeling could be reached again.
Page | 21
Data analyses
All analyses were done with the statistics program SPSS (version 24). First the reliability of the data was established through the Cronbach’s Alpha test. After some slight adjustments to make the variables reliable a correlation analysis was done. This gave an insight in which variables were correlated to each other and whether or not significant. To test the mediator variables, multiple linear regressions were done to determine if there was a significant correlation between the
independent, dependent and mediator variable. To test for the moderator variables, there were also multiple linear regressions used. An interaction variable was created to test the correlation under different values from the moderating variable.
Results
First, the Cronbach’s Alpha test was used to test the reliability of the data gathered for the dependent variable, brand engagement, and the mediator variables ‘community feeling’ and ‘achievement’. The values are shown in the table between parentheses. Originally the variable ‘community feeling’ had a lower value of 0,69. Since this was not close to the necessary value of 0,80 one of the questions and its responses was left out. It concerns the question ‘I feel connected with Albert Heijn’ were participants needed to choose the level agreeing to the statement. The answers to this question were not in line with the other three questions, thus it is left out. A possible
explanation for different responses will be discussed in the segment ‘discussion’. The new value of 0,79 was more acceptable and was further used in the analyses. With a value of 0,81, the dependent variable, brand engagement was reliable enough without necessary interferences. The other
mediator variable, achievement, was also reliable with a value of 0,85.
Correlations
In table 1 descriptive statistics are displayed. Given are the mean of each variable and the
corresponding standard deviation. The other values indicate the correlation between the variables. With 34,1 only the variable ‘age’ is noticeably higher than the rest. This is because age is answered on a normal scale whereas the other variables were measured in Likert scale or as a dummy variable. The variable ‘gamification’ and ‘gender’ have a lower value because these are the dummy variables. The same is the case for the standard deviation. The dummy variables have only recorded values between 0 and 1 and the variables measured with the Likert scale had values between 1 and 5.
The remainder values are the correlations between the variables. The correlation between gamification and brand engagement is only 0,1 and is not significant, meaning that there is no statistically substantiate proof to confirm the first hypothesis. However, there are some other
Page | 22 variables that are significantly correlated. There is a negative correlation of 0,3 between the variables ‘gamification’ and ‘gender’. Holding all the other variables constant, this indicates that women collected ‘moestuintjes’ more often than men. An explanation for this could be that women more often do big loads of groceries (for the family) and are thus more often eligible for a ‘moestuintje’.
Another significant correlation is the one between the variables ‘community feeling’ and ‘gamification’. This indicates that people who used gamification felt also more connected to the community around the ‘moestuintje’ or with other people collecting ‘moestuintjes’. It stands to reason that people collecting ‘moestuintjes’ will be more involved with other people collecting ‘moestuintjes’ than people who only buy groceries.
There is also a significant correlation between the variables ‘gamification’ and ‘achievement’. However, this result is negligible, because participants that did not use gamification have not filled in the achievement based questions and thus automatically scored 0.
There is a significant negative correlations of 0,3 between the variables ‘gender’ and ‘achievement’. This is the opposite of what was expected, a positive correlation. This indicates that women are more engaged via achievement than men. Possible explanations follow in the segment ‘discussion’.
Next to that there are also correlations between the dependent variables ‘brand
engagement’ and the mediator variables ‘community feeling’ and ‘achievement’. This means that the mediator variables have an influence on the dependent variables. If there is a mediation effect it cannot yet be concluded.
Ultimately there is a correlation between the two mediator variables. People that scored high on ‘community feeling’ also scored high on ‘achievement’. This could be an indicator that gamification gets people engaged with a subject, in this case the ‘moestuintje’.
Table 1. Descriptive statistics
M SD 1 2 3 4 5 6 1. Gamification 0,6 0,5 2. Age 34,1 11,9 0,1 3. Gendera 0,4 0,5 -0,3** -0.1 4. Brand Engagementb 3,5 0,9 0,1 0,0 -0,1 (0,81) 5. Community feelingb 2,9 1,2 0,7** 0,1 -0,1 0,1* (0,79) 6. Achievementb 3,8 1,0 0,3** 0,1 -0,3** 0,2* 0,6** (0,85)
Page | 23
** Correlation is significant at the 0,01 level (2-tailed). * Correlation is significant at the 0,05 level (2-tailed).
a 0 = female, 1 = male. b 1 = strongly disengaged, 2 = disengaged, 3 = neutral, 4 = engaged, 5 = strongly engaged.
Reliabilities are displayed on the diagonal between the parentheses.
Mediation
Judd and Kenny (1981) recommend estimating a series of regression models to find mediation. Thus to control for mediation, multiple linear regressions were done. For the first mediating variable (achievement) a linear regression is done between the independent variable (gamification) and the mediator variable to see if there was a significant correlation. Also a linear regression is done between the mediator variable and the dependent variable (brand engagement), and a linear regression between the independent and the dependent variable. Finally, a linear regression between the dependent variable and both the independent and mediating variable is done.
Table 2 shows that there is a significant correlation between ‘achievement’ and ‘brand engagement’ and also between ‘achievement’ and ‘gamification’. However, there is no significant correlation between ‘gamification’ and ‘brand engagement’. To provide evidence that confirms the hypothesis there needs to be a significant correlation between all three variables, otherwise it cannot be proved that ‘achievement’ is the explaining variable behind the correlation. However, the correlation between ‘gamification’ and ‘brand engagement’ is also absent.
In the first equation there is a significant correlation between ‘achievement’ and ‘brand engagement’, meaning that the mediator variable has an influence on the dependent variable. The second equation shows that there is also a significant correlation between ‘gamification’ and ‘achievement’. So the independent variable has an influence on the mediator variable. In the third equation is no significant correlation however. Thence ‘gamification has no influence on ‘brand engagement’. Because a significant correlation between the independent and dependent variable is missing, it can be concluded that the mediator, although affecting the dependent variable, is not the explanation behind the not significant correlation between ‘gamification’ and ‘brand engagement’. In the fourth equation both the independent and mediator variable are tested on their correlation to the dependent variable. In this equation the mediator variable is controlled. For perfect mediation the independent variable has no longer effect on the dependent variable, because the mediator is controlled and is responsible for the correlation (Baron and Kenny, 1986). The correlation between ‘gamification’ and ‘brand engagement’ is lower in the fourth equation. However, since no mediation has been established based on the first three equations, the results in the fourth equation cannot provide evidence for mediation.
Page | 24 Also for the other mediating variable, community feeling a series of regression models is applied, as suggested by Judd and Kenny (1981). Between the independent variable ‘gamification’, the mediating variable ‘community feeling’ and the dependent variable ‘brand engagement’ linear regressions were done.
In the first equation of the second part of table 2, it can be seen that there is no significant correlation between the mediator variable ‘community feeling’ and the dependent variable ‘brand engagement’. This indicates that the mediator variable has no influence on the dependent variable. Because influence of the mediator variable on the dependent variable cannot be proven, mediation can also not be proven. The mediator variable should have affected the dependent variable for mediation after all. The second equation measures effect of ‘gamification’ on ‘community feeling’. As can be seen in the table, there is a significant correlation between the independent variable and the mediator variable. Thus the use of gamification has an influence on the level of community feeling. This is however not enough to prove that ‘community feeling’ is the explanatory variable. Especially when it is taken in account that ‘community feeling’ has no influence on ‘brand engagement’ .The third equation is the same as the third equation in the first part of the table. It concerns the influence of ‘gamification’ on ‘brand engagement’ again. It was already established that there was no
significant correlation between the independent and the dependent variable. It is determined that mediation does not appear via ‘community feeling’. In the fourth equation it can be seen that the coefficient is even higher than the coefficient in the third equation. When mediation occurs it needed to be lower. This indicates that if the mediator variable is controlled, the independent variable has the same or more effect (not significant) on the dependent variable. To prove that the ‘community feeling’ explains a correlation between ‘gamification’ and ‘brand engagement’, the influence of gamification should be low when the mediator variable is controlled. This is because the controlled mediator variable should be responsible for the influence on the dependent variable. Because the independent variable and the mediator variable have the same effect on the dependent variable, the independent variable should not differ anymore.
It can also be noticed that, except for the correlation between ‘community feeling’ and ‘gamification’, the variables only explain each other for a small part, because of the low values on R-square.
Table 2. Regression results mediation
Page | 25
Dependent variable
Brand engagement Gamification Brand engagement Brand engagement
Coefficient SE B Coefficient SE B Coefficient SE B Coefficient SE B
Constant 8,7*** ,33 1,0 ,97 31,3*** ,11 3,4*** ,84 Achievement 2,0* ,09 ,21 2,9** ,98 ,29 1,9 ,09 ,21 Gamification 1,2 ,14 ,10 0,1 ,87 ,01 R2 ,04 ,08 ,01 ,04 Dependent variable
Brand engagement Gamification Brand engagement Brand engagement
Coefficient SE B Coefficient SE B Coefficient SE B Coefficient SE B
Constant 18,0*** ,19 19,4*** ,10 31,3*** ,11 17,4*** ,20 Community feeling 0,4 ,06 ,04 13,6*** ,13 ,74 -,7 ,09 -,09 Gamification 1,2 ,14 ,10 1,4 ,21 ,16 R2 ,00 ,55 ,01 ,01 Note: N = 154, * p < ,05, ** p < ,01, *** p < ,001 Moderation
In table three the effect of the moderator variables has been estimated. A moderator variable is used to maximize the effectiveness of the independent variable on the dependent variable (Baron and Kenny, 1986). It could occur that the correlation between the independent and dependent variable is stronger at a certain value of the moderator value, meaning that the correlations change based on the value of the moderator variable. In that case, the value of the moderator has effect on the correlation between the independent and dependent variable. In this research, the moderator variables are ‘age’ and ‘gender’. To determine if moderation occurs in this research, first the correlation between the moderator variable and the dependent variable needs to be established. According to Baron and Kenny (1986) the equation between the moderator variable and the dependent variable displays the intensity the moderator could have as a predictor. After that the correlation between the independent variable and the dependent variable is established. The correlation between the independent and dependent variable show the controllability as a
moderator (Baron and Kenny, 1986). In the second model the same equations happen, only now with an interaction variable added. The equation with the interaction variable and the dependent variable needs to be significant to support a hypothesis for moderation. The effects of the first two equations (without the interaction variable) do not have to be significant to test for moderation.
In model 1 of the first part of the table, the equations are shown between the moderator variable ‘gender’ and the dependent variable ‘brand engagement’ and also between the independent
Page | 26 variable ‘gamification’ and ‘brand engagement’. As been established in the previous section of the result, there is no positive correlation between ‘gamification’ and ‘brand engagement’. The only way moderation can occur now is if the correlation becomes significant at a certain value of the
moderator variable. The correlation between ‘gender’ and ‘brand engagement’ is also not significant. This indicates that gender does not have an influence on the level of engagement. It was expected that ‘gender’ would have an effect on ‘brand engagement’. However, a significant correlation between the moderator and the dependent variable is not necessary for moderation. It is important that the correlation is significant as the independent variable comes in the equation too. This happens with the interaction variable. The estimation of the interaction variable can be found in model 2. The correlation between the interaction variable and the dependent variable is not significant; hence moderation cannot be proven for the moderator variable ‘gender’.
The second part of table 2 is dedicated to the other moderator variable ‘age’. The first equation focuses on the correlation between the moderator variable ‘age’ and the dependent variable ‘brand engagement’. There is no significant correlation between these variables. This indicates that age has no effect on the level of brand engagement. This is however not necessary to prove moderation, because the moderation does not have to affect the dependent variable, but the correlation between the independent and the dependent variable. The second equation has been shown multiple times before. It involves the independent variable ‘gamification’ and the dependent variable ‘gamification’. As been established before, there is no significant correlation and
‘gamification’ has no influence on ‘brand engagement’. However, this does not exclude moderation. The correlation only needs to be significant at a certain value of the moderator and differ somewhat logically from the significance at other values. Thence a significant correlation only needs to occur partly. As last there is the equation with the interaction variable and the dependent variable. There is also no significant correlation between the interaction variable and the dependent variable. This indicates that there is also no significant correlation when the moderator variable is categorized and there is no logical relation between the different categories. Therefore moderation cannot be proven for the moderator variable ‘age’.
It is also noticeable that the value of R-square stays the same in both cases. This indicates by adding the interaction variable there is not a bigger part of the relation between the variables explained. If moderation had occurred the R-square value should have increased because a bigger part of the relation could have been explained.
Page | 27 Table 3. Regression result moderation
Brand engagement (DV)
Model 1 Model 2
Coefficient SE Beta Coefficient SE Beta
(Constant) 25,08*** ,14 20,55*** ,16
Gender -,52 ,15 -0,44 ,01 ,22 ,00
Gamification 1,03 ,15 ,087 1,09 ,19 ,12
Gen*Gam -,48 ,30 -,06
R2 ,01 ,01
Coefficient SE Beta Coefficient SE Beta
(Constant) 17,43*** ,20 13,76*** ,26
Age -,22 ,09 -,02 -,53 ,13 -,06
Gamification 1,25 ,14 ,10 1,26 ,14 ,10
Age*Gam ,52 ,18 ,06
R2 ,01 ,01
Note: Dependent variable is brand engagement, N = 154, * p < ,05, ** p < ,01, *** p < ,001 Gen = Gender, Gam = Gamification
Page | 28
Discussion
First the results will be discussed in order of the hypotheses. After that there will be a segment dedicated to unexpected findings.
Hypothesis 1 = H1: Gamification used as a marketing tool will increase the level of brand engagement of the customer.
It was expected that gamification would get people engaged and that, because of it, gamification could be used as a marketing tool. However, the analyses showed that there is no significant correlation between gamification and the level of brand engagement. The correlation was only 0,1 and not close to significance. Gamification has already proven its effectiveness in practice, so the results are in contrast with that. A possible explanation could be that the customers do get engaged with Albert Heijn’s gamification, but that is does not translate to brand engagement. In that case, engaging people via gamification does not directly increase people’s level of brand engagement.
Page | 29 the brand is not associated strongly enough. People would get excited about the ‘moestuintjes’ and excitement causes brand engagement, only the connection between gamification and brand engagement is not made by the people. In the research of Muntean (2011), children become engaged with a subject through gamification and it works. However, being engaged to a game in order to learn something without realizing it, does not necessarily make the children more engaged to the company that provided the gamification method.
In the segment of Hypothesis 2b the mediator ‘community feeling’ is discussed. There can be observed that a connection can be made between gamification and getting engaged with a subject and subsequently being engaged leading to brand engagement.
Hypothesis 2a = H1: Being part of a social community makes people more engaged with a brand.
By looking at the significant correlation of 0,1 it can be concluded that social stimulation has an effect on the level of brand engagement. However the correlation is low and it is significant with p<0,05. But it is enough to accept this hypothesis. This is in line with the assumptions of Algesheimer et al. (2005). They expected brand engagement via a brand community. The identification with other members of the group develops an emotional connection that is necessary for brand engagement. Also Bhattacharya et al. (1995) expected that being part of a group made customers more agreeable to the norms of the group. A group based on the gamification of a brand is not likely to take a hostile position toward the brand. Therefore the members of the group are likely to be positively engaged to the brand.
Hypothesis 2b = H1: The impact of gamification on brand engagement is mediated via social feeling.
There is no statistical evidence to confirm this hypothesis. In table 2 could be seen that the necessary significant correlations were not present to back up the hypothesis. This is mainly due to the fact that hypothesis 1 is rejected. Without the correlation between gamification and brand engagement, mediation was difficult to prove. However, gamification had an effect on the community feeling. There was a strong, significant correlation of 0,7.
In the segment of the previous hypothesis the correlation between social stimulation and brand engagement is discussed. It was only a correlation of 0,1, but significant. This is in line with the research of Ducheneaut and Moore (2004). They argue that the social part of a game design can create a ‘community feeling’. The ‘moestuintjes’ of Albert Heijn, however not specifically designed as social game, have contributed to a community feeling. Without the ‘moestuintje’ being a
Page | 30 share and swap or collect for someone else. Next to that the positive feelings generated by a well growing ‘moestuintje’ can easily be shared.
This means that only the correlation between gamification and brand engagement is lacking to prove mediation. Because it was unexpected that the main correlation was lacking it would still be interesting to keep this mediator variable in mind for future research.
Hypothesis 3a = H1: A sense of achievement makes people more engaged to a subject.
The correlation between ‘achievement’ and ‘brand engagement’ was 0,2. It is not a strong correlation and the significance was only p<0,05. It can be concluded, however, that achievement stimulation makes people more engaged and it is enough evidence to accept the hypothesis.
Although the correlation is weak, it is in line with the literature. Ryan and Deci (2000) argue that people become stimulated via achievement because of the positive feeling they get from accomplishing a task. People become engaged because they want to experience the feeling again. Next to that it contributes to their self-image in positive manner. According to Weiner (1985) this is important because people strive to have a positive self-image. Even so, failing a ‘moestuintje’ can lead to a decrease in self-image. Therefore it is more likely for a person to engage with a
‘moestuintje’ again if the previous outcome generated positive feelings.
Hypothesis 3b = H1: The impact of gamification on brand engagement is mediated via a sense of achievement.
This hypothesis, just as 2b, cannot be proven due to the fact that there is no significant correlation between gamification and brand engagement. The other correlations that were necessary to prove mediation were significant. Gamification causes achievement stimulation with a significant
correlation of 0,3. Also, as discussed in the segment of the previous hypothesis, there was a significant correlation of 0,2 between achievement stimulation and brand engagement. Again only the correlation between gamification and brand engagement is missing to prove mediation.
Accept for the lacking correlation between gamification and brand engagement was the outcome of this result in line with what was expected. The effect of achievement stimulation, as explained in the section above, was significant and gamification had an influence on the achievement stimulation. This is in line with the research of Juul (2003). A game design makes the player feel attached to the outcome and therefore the player puts effort in the game. It also confirms Weiner’s
Page | 31 (1985) assumptions of wanting to create a positive outcome.
Hypothesis 4a = H1: When familiar with a brand, women are more brand engaged than men
The correlation between gender and brand engagement is negative. This is because for women a value of 0 was used and for men a value of 1. So the correlation is in line with the hypothesis. However, it is very weak (0,1) and not significant. Therefore the hypothesis cannot be proven.
Although the result is not significant, it is in line with the research of Oly Ndubisi (2006). Once women get familiar with a brand they become more engaged than men. All the participants were assumed to be familiar with Albert Heijn, because it exists for a long time and is not a niche market. Everyone comes in touch with a supermarket. However, the correlation was weak (only 0,1). It could be possible that the effect of gender flattens because of the branch Albert Heijn is in. Groceries are not products to show off to other people. People would more easily identify themselves with a brand that has cool products, i.e. the Harley Davidson community. A motorcycle is something that a person wants to show to the world and adapt the corresponding image with. Therefore they will view that brand as an extension of themselves. Something that is less likely to occur with groceries. Thus the effect may be weaker in this case.
Hypothesis 4b = H1: The effect of gamification on brand engagement will be stronger on women.
In this hypothesis a moderator variable is tested. There may not be correlation between gamification and brand engagement, but that is not necessary for moderation. Also between gender and brand engagement was not a significant correlation as described in the previous segment. To prove moderation an interaction variable is decisive. However, the interaction variable did also not have a significant correlation toward brand engagement. Therefore there is not enough evidence to assume this hypothesis would be true.
It was expected that gender would have an effect on the correlation between gamification and brand engagement. The literature is agreeable with each other. But for the different genders there are different factors that should engage them. Women would be more engaged via social stimulation, whereas men would be engaged via achievement stimulation. There is a possibility that these factors would cancel each other out so there would not be a significant correlation.
Hypothesis 5 = H1: The effect of gamification on brand engagement will be stronger on younger customers.
