The immediate gratification bias and online behavioural advertising in a social media context

(1)

The immediate gratification bias and online

behavioural advertising in a social media

context

Interactions between immediate gratification, trust and privacy concerns on

click-through intentions of retargeted ads

(2)

The immediate gratification bias and online

behavioural advertising in a social media

context

Interactions between immediate gratification, trust and privacy concerns on

click-through intentions of retargeted ads

Master Thesis, MSc Marketing, specialization Marketing Management University of Groningen, Faculty of Economics and Business

(3)

Abstract

Online behavioural advertising (OBA) shows advantages through its possibility to reach a targeted audience. People’s responses to OBA can increase ad effectiveness, but paradoxically, it can also lead to decreasing the effectiveness of an ad. Trust in retailers and consumers’ privacy concerns are found to be two important factors which affect the level of effectiveness of OBA. However, these relationships are investigated from the perspective that consumers are fully rational, and the stream of behavioural economics shows that people are not rational decision makers, since their bounded reality leads to biases. In privacy related decisions people are led by the immediate gratification. Therefore, to fully describe the human behaviour in people’s response to OBA, this study investigates the effect of the immediate gratification bias on the relation between the consumers’ perception of trust and their privacy concern, and the effectiveness of OBA. To investigate the effect of the bias, this study uses two online experiments which manipulate participants for trust, privacy concerns and immediate gratification, shows them a retargeted ad on social media and measures their click-through intentions. The results show that a high trusted company has higher click-through intentions than a low trusted company when there is no immediate gratification, however when there is immediate gratification the click-through intentions are the same regardless of the level of trust. Besides, the results show that the click-through intentions of people with high or low privacy concerns are not different whether there is immediate gratification or not. Thus, the trust in online retailers only matters less for the effectiveness of OBA when people want immediate gratification and the immediate gratification bias do not affect the relation between privacy concerns and the effectiveness of OBA.

Keywords: Immediate gratification; Trust; Privacy concerns; Retargeting; Online marketing;

(4)

Preface

The thesis in front of you is “The effect of the immediate gratification bias on online behavioural advertising in a social media context”. This thesis is written for my graduation of the MSc Marketing at the University of Groningen between September 2020 and January 2021. It combines the two fields of study, Marketing and Psychology, that interested me the most during my time as an MSc Marketing student. Therefore, I worked very passionately on this research and see this thesis as the culmination of my time as a student at the University of Groningen.

During the process of written my thesis my supervisor, Judith De Groot, provided me with very helpful feedback, which helped to bring my thesis to the level I wanted. The research was a complex process, which required to combine all the things I learned as a student. However, after extensive research and hard work, I could answer my research question. During this process, I had questions which I always could ask Judith De Groot and my fellow students. Therefore, I would like to thank them, since without their support this research could not be completed.

Finally, I would like to thank my family and friends for their moral support during the writing process and their open-minded feedback to my research ideas. This helped me a lot by finding solutions to questions and it kept me sharp.

I hope you will enjoy reading it. J.M. Kiewiet

(5)

In recent years, the interest in Online Behavioural Advertising (OBA) increased since advertisers see OBA as one of the most important ways to reach a targeted audience (Boerman, Kruikemeier, Zuiderveen, & Borgesius, 2017). According to Kumar and Gupta (2016) OBA is believed as the future of advertising among others e-commerce. OBA refers to an advertising method that uses data such as visited websites, read articles, watched videos, made purchases, provided information of consumer profiles and everything that has been searched with a search engine. All this data will be used to create a personalized ad which is shown to the targeted consumer (Boerman et al.,2017; Dehling, Zhang, & Sunyaev, 2019). In the online industry, OBA is becoming more common since it increases the ad effectiveness (Aguirre, Mahr, Grewal, de Ruyter, & Wetzels, 2015a; Aguirre, Roggeveen, Grewal, & Wetzels, 2015b; Chen, & Stellaert, 2014), less money on advertising is wasted (Gironda, & Korgaonkar, 2018), and marketeers are able to experiment with relatively low costs on advertising (Alreck, & Settle, 2007). Therefore, online retailers such as Amazon and eBay use OBA. More specifically, 44% of global advertising expenses is spent on digital media and herein a large part of these expenses went to OBA in 2018 (Liang, Jiao, & Liu, 2020).

Although the advantages of OBA have been acknowledged by online retailers, it has also been associated with disadvantages (Aguirre et al., 2015a; Dehling et al., 2019; Smit, Van Noort, & Voorveld, 2014). It may increase the feelings of vulnerability and lower adoption rates of consumers. Consumers feel discomfort since they realize that their personal information is collected without their consent (Aguirre et al., 2015a). Also, consumers may perceive OBA as intimidating and creepy (Dehling et al., 2019) or they are worried about misuse of their personal data (Smit, Van Noort, & Voorveld, 2014). This problem of OBA can also be summarized in the personalization paradox, which refers to the two-sided effect of a high level of personalization. On one hand, “greater personalization typically increases service relevance and customer adoption, but paradoxically, it also may increase customers’ sense of vulnerability and lower adoption rates” (Aguirre et al., 2015a, p. 34)

(8)

retailer (Corbit, Thanasankit, & Yi, 2003). Therefore, consumers who trust the retailers are likely to believe that their personal information is safe in the hands of the retailer (Bleier, & Eissenbeis, 2015a). Secondly, the privacy concerns of consumers affect the effectiveness of OBA (Aguirre et al., 2015a; Aguirre et al., 2015b; Boerman et al., 2017; Gironda, & Korgaonkar, 2018; Van Doorn, & Hoekstra, 2013). A higher level of privacy concerns increases scepticism towards advertising and also leads to more avoidance of them (Aguirre et al., 2015b; Boerman et al., 2017), which decreased the effectiveness of OBA.

(9)

According to Acquisti (2004), the immediate gratification bias offers a potential explanation of people’s decision-making process of privacy. It means that online consumers discount privacy violation risks and choose immediate benefits when they are shopping online (Bandara et al., 2020). Since privacy issues are also part of the personalization paradox the immediate gratification bias might have an impact on the effectiveness of OBA. This is in line with the framework of Boerman et al. (2017), which stated that despite the relevance there is a lack of literature that test the effect of personal biases on people’s behaviour in response to OBA. Therefore, this paper will address the effect of the immediate gratification bias on people’s behaviour in response to OBA and function as a boundary condition of OBA.

For e-commerce companies investigating the effect of the immediate gratification bias on people’s behaviour in response to OBA will be useful since the upcoming of online retailing causes a more impatient consumer (Daugherty, Bolumole, & Grawe, 2018). In the past, customers did not mind if they had to wait for the order but nowadays, they want immediate delivery options. Especially the millennials who grow up with mobile ordering, expecting immediate delivery options (Beckwith, 2017). Furthermore, social media plays an important role in communication to the customers for e-commerce companies (Daugherty et al., 2018). The use of social media is growing fast (Clement, 2020). Therefore, marketeers increasingly using social media advertising in combination with highly relevant targeting techniques (Jung, 2017). However, collecting online behavioural data and use this data for targeted social media advertising can also be seen as creepy or intimidating (Dehling et al., 2019). Despite, it is said that OBA has a promising future (Kumar, & Gupta, 2016). So, the increasing importance of social media for e-commerce makes this a relevant research context.

(10)

certain conditions retargeting is effective to let people click on the advertisement. However, Hamby and Ilyuk (2019) argue that it can also raise the feelings of intrusiveness, thereby decreasing the effectiveness of retargeting on consumer perceptions about retargeting. Indeed, Bleier and Eissenbeis (2015a, p.391) confirmed the mixed outcomes in literature in their statement that “retargeting is a complex and largely unexplored method of ad personalization”. Therefore, to further understand the conditions under which retargeting might be effective, the present study will examine the impact of the immediate gratification bias on people’s response to retargeting.

2. LITERATURE REVIEW

2.1RETARGETING AS A TYPE OF ONLINE BEHAVIOURAL ADVERTISING (OBA)

OBA is a special form of targeted advertising (Smit et al., 2014). However, in literature there are different definitions. It can be defined as “the practice of monitoring people’s online behaviour and using the collected information to show people individually targeted advertisements” (Boerman et al., 2017, p.364). Dehling et al. (2019) define OBA as a pervasive technology, where online ads are tailored to consumers interests and tastes based on their online behaviour with the main goal to increase sales and profit. Liang, Jiao, and Liu (2020, p. 1143) focus more on the information collection component of OBA by defining it as “collecting data based on tracking consumers' online behaviour to show individually targeted ads”. That is, their definition shows the main goal of OBA is to use personal information from online behaviour and create an ad that is tailored to a specific person, which makes it more personally relevant. Therefore, OBA will be defined in this article as the online advertising method whereby consumers’ browsing behaviour is tracked and collected to show them a personalized advertisement.

(11)

comparison to non-personalised ads on Facebook. However, highly personalized ads (based on interest in a subject, gender, age, and their location) decreased click-through rates. Also, Van Doorn and Hoekstra (2013) found that highly personalized ads added with personal identification and transaction information negatively affect but this was on the purchase intention of consumers. The performance of online advertising is commonly measured by the click-through rate of an ad (Yan, Liu, Wang, Zhang, Jinag, & Chen, 2009). Furthermore, advertisers pay usually per click since their aim is to reach as many consumers willing to pay as possible with their advertisements (Dehling et al., 2019). Therefore, the click-through rate will be used to measure the effectiveness of OBA.

Before consumers will click on an ad it is important that an ad is relevant to them since consumers pay more attention to a relevant ad (Jung, 2017). The relevance of an ad for consumers is determined by its perceived personalization since favourable ad responses only appear when people perceive an ad as personalized (Li, 2016). The level of perceived personalization is especially for OBA important because OBA differentiate itself from other online advertising due to their aim for personal relevance (Boerman et al., 2017). When the ad is perceived as relevant by the consumers, it promotes greater attention (Li, 2016). In other words, consumers notice the ads better since the ads are relevant to them. According to Kim and Huh (2017) more attention to the ad causes that consumers more positively evaluate them, and, in the end, a positive evaluation increases the clicks on the advertisements. So, perceived personalization is positively related to the effectiveness of OBA.

(12)

message (Jung, 2017). Ultimately, this loss of control can lead to ad avoidance (Jung, 2017). So, people are less likely to click on an ad and therefore perceived personalization is also negatively related to the effectiveness of OBA. Herewith, OBA can be seen as a double-edged sword since the perceived personalization is both positively as negatively related to the effectiveness of OBA.

(13)

2019). Privacy concerns are one of the underlying factors of the negative responses through retargeted ads (Van Doorn, & Hoekstra, 2013). Besides, Bleier and Eisenbeiss (2015b) discovered the danger of overpersonalization. The danger of overpersonalization means that retargeted ads with specific products are effective when consumers just visited the online store but when the time passes since the last visit to the product the effectiveness will quickly decrease. Thus, for retargeting there is the same double-edged sword as for OBA since the effect of retargeting can both be positive and negative as well.

2.2THE RELATIONSHIP BETWEEN TRUST AND THE EFFECTIVENESS OF OBA

Trust is a key component in building a relationship with online customers. This is especially in e-commerce important (Bleier, & Eisenbeiss, 2015a; Wu, Huang, & Hsu, 2013). Therefore, the concept of e-trust is often used in e-commerce and marketing (Choi, & Mai, 2018). Consumers interact in virtual environments and cannot see, touch or demonstrate the product. As a result, shopping online is perceived as risky (Sanje, & Senol, 2012). E-trust can be defined as an attitude which reflects the consumers’ opinion regarding their confidence in making online purchases. Herein, the consumers’ perception of how their data is protected and the e-tailer’s commitment to honest behaviour and integrity in the foreseeable future, is important (Faraoni, Rialti, & Zollo, 2019). In contrast, Bhattacherjee (2002, p. 213) defines trust more in general as “the willingness of a party [trustor] to be vulnerable to the actions of another party [trustee] based on the expectation that the other [trustee] will perform a particular action important to the trustor, irrespective of the ability to monitor or control that other party [trustee]”. In the online context, for instance on social media, this means that the consumer is the trustor, and the company is the trustee. Furthermore, in marketing literature trust has been defined as “willingness to rely on an exchange partner in whom one has confidence” (Bleier, & Eisenbeiss, 2015a, p.396). Herein, confidence is based on the trustor’s belief in the trustee’s reliability and integrity (Morgan, & Hunt, 1994). The definition of e-trust of Faraoni et al. (2019) is in general similar to the definitions of Bhattacherjee (2002) and Bleier and Eissenbeis (2015a). However, the context of the research is e-commerce, and therefore the definition of e-trust will be used in this study.

(14)

if consumers trust online retailers and have confidence in their reliability and integrity, they are more likely to make a purchase and disclose their information (Wang, & Emurian, 2005). Without trust personal information will not be shared (Wu et al., 2014).

In literature, trust is found as an important factor in the effectiveness of OBA (Boerman et al., 2017). A more trusted retailer can use consumers’ information to create an ad that reflects consumers’ interests in a complete way. These ads which are highly personalized increased click-through rates, but these increased click-through rates only occurred when the retailer was trusted (Bleier, & Eissenbeis, 2015a). The reason that trust increased the click-through rates is that people accept their vulnerability feelings. Aguirre, et al. (2015a, p. 44) showed that trust-building strategies can counteract these feelings “if the personalized advertisements appear in a credible context or incorporates information icons that signal trustworthiness”. Based on these results, this study assumes that when a consumers’ trust in a retailer is stronger, personalized advertisements will more likely result in a stronger intention to click-through these ads compared to non-personalized advertisements (Bleier, & Eisenbeiss, 2015a; Boerman, et al., 2017).

2.3THE RELATIONSHIP BETWEEN PRIVACY CONCERNS AND THE EFFECTIVENESS OF OBA

Privacy concerns play an important role in the effectiveness of OBA, especially in relation to the negative effect of OBA on the effectiveness of retargeting. Consumers have become more concerned about OBA and mainly the impact of it on their privacy (Boerman et al., 2017). Privacy concerns can be seen as the degree of worrying about the violation of people’s right to prevent personal information from sharing it with other parties (Baek, & Morimoto, 2012). So, in the online context this means that consumers have a lack of control about how their online behaviour and information is collected. That is why privacy concerns in the context of OBA will be defined in this study as “the degree to which a consumer is worried about the potential misuse of his or her online activity history data collected by advertisers” (Kim, & Huh, 2017, p.96).

(15)

(2017) explains why privacy concerns have a negative effect on the effectiveness of OBA. That is, higher privacy concerns increase the avoidance of ads since consumers who know and worry that advertisers collect their personal information and use this for advertising. This explanation is in line with Baek and Morimoto (2012), which found that a higher level of privacy concerns for personalized advertisements create ad scepticism and ad avoidance. The avoidance results in people not clicking on the ads (Jung, 2017). In other words, consumers who avoiding ads are paying less attention to ads, which results in fewer clicks on ads. Finally, Aguirre et al. (2015a) found that privacy concerns affect click-through rates because when consumers find out that their information is collected without their consent the click-through rates sharply decline. So, privacy concerns of consumers have a negative impact on the effectiveness of OBA.

2.4THE INTERACTION-EFFECT BETWEEN IMMEDIATE GRATIFICATION,TRUST AND PRIVACY

CONCERNS ON THE EFFECTIVENESS OF OBA

In the studies which discussed relationships between the consumer perception of trust and their privacy concerns, and the effectiveness of OBA, privacy played an important role (Aguirre et al., 2015a; Aguirre et al., 2015b; Boerman et al., 2017; Bleier, & Eisenbeiss, 2015a; Gironda, & Korgaonkar, 2018; Stevenson, & Pasek, 2015; Van Doorn, & Hoekstra, 2013; Wu et al., 2013). OBA, specifically retargeting, causes a dilemma for the consumers, which is also named the personalization paradox (Aguirre et al., 2015a). High personalized ads based on consumers’ online behaviour can create relevant ads but paradoxically, it can lead to feelings of vulnerability and decreased freedom of online behaviour (Aguirre et al., 2015a; Hamby, & Ilyuk, 2019). In other words, consumers have to make a decision between the benefits or losses of OBA, and this has consequences for their privacy.

(16)

decision making is influenced by mental biases, heuristics, affect and limited cognitive resources (Bandara et al., 2020). The assumption of such limited decision-making processes is in line with Acquisti and Grossklags (2005), who argued that people cannot be rational in privacy-related decisions. People do not have this capacity to process all the information if all the information is already available (Acquisti, 2004). Furthermore, decisions are not always made consciously and analytically (Bandara et al., 2020). In other words, people are not able to rationally make a weight or trade-off between the benefits and losses of their privacy choices. That is why people rely on heuristics or mental shortcuts when they have knowledge limitations, information asymmetry and/or time constraints. These heuristics able people to make decisions quickly but it leads to biases also (Bandara, et al., 2020). This study assumes that the individual decision-making process of OBA is affected by personal biases.

(17)

Figure. 1 Conceptual model

2.5HYPOTHESES DEVELOPMENT

The research question that this study will address is: “What is the effect of the immediate gratification bias on the relation between the consumer perception of trust and their privacy concern, and the effectiveness of OBA for retargeted ads on social media?”. Figure 1. contains the conceptual model of this study.

(18)

Hypothesis 1: The stronger the immediate gratification, the less likely one’s trust will influence the click-through rate.

Hypothesis 2: The stronger the immediate gratification, the less likely one’s privacy concerns will influence the click-through rate.

3.METHODOLOGY

3.1EXPERIMENTAL RESEARCH DESIGN

Two online scenario-based experiments were conducted to test the hypotheses. Experimental designs are not common in the literature stream of OBA, since research in this stream is mainly conducted by surveys or interviews (Chen et al., 2019; Dehling, et al.,2019; Gironda, & Korgaonkar, 2018; Kaan, & Varnaali, 2019; Kim, & Huh, 2017; Li, & Huang, 2016; Smit et al., 2014; Van Doorn, & Hoekstra, 2013). However, this study used online experiments hereby enabling to draw causal inferences (Malthora, 2010).

The first experiment focused on the interaction-effect between trust and immediate gratification on the intention to click-through a personalized ad (Hypothesis 1). It entailed a two by two between-subject experimental design. The independent manipulation variables included trust (high versus low trust) and immediate gratification (strong versus weak immediate gratification). After being exposed to one of the four experimental conditions, the participants answered questions in relation to the dependent variable, click-through intentions.

The second experiment focused on testing the interaction-effect between privacy concerns and immediate gratification on the click-through intentions of personalized ads (Hypothesis 2). The independent manipulation variables included privacy concerns (high versus low concerns) and immediate gratification (strong versus weak immediate gratification). Like in the first experiment, the dependent variable was click-through intentions.

(19)

3.2SAMPLE AND SAMPLING STRATEGY

The research was conducted in the Netherlands. Hence, a Dutch-speaking population was targeted and only people who fluently speak Dutch could participate in the study. Furthermore, people who did not own a social media account were excluded from the study. To easily reach individuals with both preconditions, a Dutch survey invitation was sent out on social media using a convenient sampling strategy. Convenience sampling was relevant in this research because it enabled to quickly reach the targeted population of this study via social media. An a-priori sample size analysis (Soper, 2020) revealed that for a small anticipated effect (f2₌

0.10), with a minimal desired statistical power of 0.80, using a probability level of 0.05, and including seven predictors (trust/privacy concerns, immediate gratification, and the interaction term trust/privacy* gratification, and, the confounding variables of age, frequency of social media use, product relevance and trust/privacy concerns), a minimum sample per experiment of 150 participants was needed (Cohen, 1988). For this study was chosen to conduct first Experiment 1 and then Experiment 2 since this method enabled the study to check whether the manipulation worked and if some changes had to be made. The final sample size for Experiment 1 included 193 participants, which meant that the sample size was sufficient for the aim of the present study.

After analyzing Experiment 1 it was found that age was not a variable that created significant covariance, so that decreased the needed minimum sample size for Experiment 2 into a minimum of 142 (Cohen, 1988). The final sample size of Experiment 2 included 144 participants. Therefore, the sample size was sufficient for the aim of the present study.

3.2.1SAMPLE EXPERIMENT 1

The sample of Experiment 1 was conducted on a group of Dutch social media users who participated in the whole experiment (n=190). This sample consisted of 97 men (51.10%), 92 women (48.40%) and 1 person choose not to disclose (0.50%). The average age in the sample was 29 (Mage =29.26, SD =12.80). Furthermore, the median of the average time spent on social

(20)

3.2.2SAMPLE EXPERIMENT 2

The sample of Experiment 2 was conducted on a group of Dutch social media users who participated in the whole experiment (n= 143). This sample consisted of 82 men (57,30%) and 61 women (42.70%). The average age in the sample was 29 (Mage= 29.77, SD = 13.28).

Furthermore, the median of the average time spent on social media was between 1 and 2 hours a day. Thus, the sample was relatively young (CBS, 2020b) and spent similar time on social media (NOS, 2020) compared to the general Dutch population, which made it a convenience sample.

3.3PROCEDURE OF THE EXPERIMENTS

Both experiments started by asking participants whether they had a social media account or not, how frequently they used social media and questions about their age, gender, and education. When participants indicated that they did not have a social media account, they were thanked and excluded from further participation (Experiment 1: n=3; Experiment 2: n=1). Then the participants were randomly assigned to one of the four experimental conditions. Both experiments consisted of five parts and started with an introduction to the scenario they were in. In Experiment 1 the participants were looking for a new a pair of smart shoes on the website of the online retailer called ‘ShoeHunter’ and every shoe had the right size and fit for the participant. In Experiment 2 participants were looking for a new health insurance on an online comparison website called ‘HealthAssurance4U’, since their old health insurance did not fit with the participant’s current life situation. Every health insurer had several underlaying insurances and supplementary schemes. In the experiments were used fictious names instead of real company names to exclude participants’ pre-assumptions of existing companies (Bleier, & Eissenbeis, 2015a).

Next to the scenario introduction, participants were manipulated for immediate gratification (Experiment 1 and 2), trust (Experiment 1) and privacy concerns (Experiment 2). For the manipulation of immediate gratification, participants received a message with an incentive that they needed the product soon. More specifically, in Experiment 1 participants got a message that they needed the smart shoe for the wedding of their best friend soon. In Experiment 2 participants were told that it was 30th_{of December and before the 1}st_{of January the health}

(21)

(high trust) or four negative (low trust) online reviews of other consumers about the company’s trust. These reviews were based on the trustworthiness of the retailer in Experiment 1. For the manipulation of privacy concerns, the participants either saw four positive (low privacy concerns) or four negative (high privacy concerns) online reviews of other consumers about the privacy policy and the data sharing to third parties of the comparison website in Experiment 2. In the second part of both experiments, participants had to select an item they liked the most, which was added to their favourites. For Experiment 1 participants had to select the smart shoe they liked the most. For Experiment 2 participants had to select the health insurer that suited them best. Selecting these items and place them in their list of favourites, contributes to the realism of the experiment (Bleier, & Eissenbeis, 2015a).

Part three showed the participants a cover story that they left the websites without buying the smart shoes (Experiment 1) or without comparing the health assurances (Experiment 2). In part four participants are told that they are scrolling on a social media platform the next day. Here, participants saw a retargeted ad about the specific product what was added to their list of favourites. More specifically, participants saw the smart shoes (Experiment 1) or the health assurance (Experiment 2) they selected.

Finally, the participants filled out a survey in both experiments, wherein the click-through intentions on a retargeted ad and the product relevance of the particular product in the experiment were measured. Also, the participants answered questions about their privacy concerns and the trust in online companies in both experiments. Finally, in Experiment 2 participants also answered questions about their immediate gratification.

3.4MATERIALS

(22)

Trust is manipulated in high trust in the retailer vs. low trust in the retailer through reviews about the retailer (Appendix 2). Herein, the participants are shown positive or negative reviews related to trust, which let them think that the retailer in the experiment can be trusted or not. This manipulation of trust is adapted from Bleier and Eissenbeis (2015a). They manipulated trust by giving participants specific background information that described the prior experiences and interactions with the retailer in a positive or a negative manner. In this study the specific background was changed to online reviews of the website since the manipulation through reviews has impact on the consumers’ opinion of trust. Reviews play an important role in consumer decision making and they reveal the trustworthiness of the e-tailer (Utz, Kerkhof, & Van den Bos, 2012; Ozturcan, & Gursoy, 2014). That is, in the context of this study the confidence in making online purchases will be influenced since the reviews will say something about the integrity and honesty of the e-tailer towards data protection of consumers.

Experiment 1 chose the smart shoes as the product for the retargeted ads (Appendix 3). A pair of shoes is a product that everyone needs, and it is part of the most sold products online in the Netherlands (Ecommerce Foundation, 2018). Furthermore, online shoe selling in the Netherlands increased the last years (Emerce Fashion, 2019). Therefore, this product left to the imagination of the participant and fitted within the context of the study.

Table 1. Conditions of Experiment 1

Experiment 1 Immediate gratification No immediate gratification

High trust in retailer

- Positive reviews about the retailer

- Emphasis that participants need the product soon

- Positive reviews about the retailer

- No emphasis on immediacy of product

Low trust in retailer

- Negative trust reviews about the retailer

- Emphasis that participants need the product soon

- Negative trust reviews about the retailer

- No emphasis on immediacy of product

(23)

manipulated the same as in Experiment 1 with either a strong immediate gratification versus no immediate gratification or patient (Appendix 4). The only difference with Experiment 1 was that the immediate gratification group received not a message that they needed the advertised product (smart shoes) soon for a wedding but that participants had to imagine that it was 30th

of December and before the 1st_{of January they had to change their health insurance. So,}

changing the participants’ health insurance was urgent.

Furthermore, privacy concerns were manipulated in high privacy concerns versus low privacy concerns through reviews about the privacy violations of the e-tailer or the e-tailers’ good data protection behaviour (Appendix 5). Consumers’ privacy concerns are triggered when they become aware that firms collect and use their personal data without permission (Lainer, & Saini, 2008; Aguirre et al., 2015a). However, reviews about companies’ good data protection behaviour lead to low privacy concerns since openness and asking for collecting and personal data of consumers decreases privacy concerns (Lainer, & Saini, 2008). In other words, the positive reviews about the data protection resulted in low privacy concerns and the negative reviews about privacy violation triggered high privacy concerns.

(24)

Table 2. Conditions of Experiment 2

3.5MEASURES

The constructs of the two experiments were measured on a Likert scale since this scale enabled the study to measure attitudes and intentions. The Likert scale showed statements to the participants and they had to indicate whether they agree or disagree (Malthora, 2010). So, measuring participants trust in online retailers, their privacy concerns, their immediate gratification and their intentions to click-through on retargeted ads in combination with manipulating participants immediate gratification, abled the study to explain the impact of the immediate gratification bias on people’s response to retargeting. Appendix 7 (Experiment 1) and Appendix 8 (Experiment 2) provide an overview of the questions during the experiments.

3.5.1MEASURES EXPERIMENT 1

The dependent variable in Experiment 1, click-through intentions on a retargeted ad was measured in Experiment 1 with four items adapted from Bleier and Eissenbeis (2015a), and Gironda and Korgaonkar (2018), since both studies measure the construct similar to this study. The four items are: “I intend to click-through the ad to acquire further information”, “I plan to click on the ad to acquire further information”, “It is likely that I will click on this ad”, and

Low privacy concerns

- Positive reviews about the comparing website

- Emphasis that participants need to change their health insurance soon

- Positive reviews about the comparing website

- No emphasis on immediacy to change the health insurance

High privacy concerns

- Negative reviews about the comparing website

- Emphasis that participants need to change their health insurance soon

- Negative reviews about the comparing website

(25)

“Probably I will click on this ad”. The participants responded to these questions on a Likert scale from 1 (“Strongly disagree”) to 7 (“Strongly agree”). Alpha reliability was .958.

Furthermore, to check whether the manipulation of trust in Experiment 1 was successful, the trust in retailers was measured on a ten items scale adapted from Bhattacherjee (2002) since their study was focussed on developing and validate an instrument to measure the trust of consumers in online firms. However, the questions of Bhattacherjee are based on a specific retailer and this experiment wanted to measure the trust in general of consumers in online firms and therefore the ten items were adapted in the following questions: “Online retailers have the skills and expertise to perform transactions in an expected manner”, “Online retailers have access to the information to handle transactions appropriately”, “Online retailers have the ability to meet most customers’ needs”, “Online retailers are fair in their conduct of customer transactions”, “Online retailers are fair in their use of private user data collected during a transaction”, “Online retailers are fair in their customer service policies following a transaction”, “Online retailers are open and receptive to customer needs”, “Online retailers keeping their customers’ best interest in mind during most transactions”, “Online retailers make good-faith efforts to address most customer’ concerns”, and “Overall, online retailers are trustworthy”. All these items were measured on a seven-point rating scale from 1 (“Strongly disagree”) to 7 (“Strongly agree”). Alpha reliability was .899.

Finally, the effect of privacy concerns was in Experiment 1 measured as a confounding variable. Privacy concerns of participants were measured on a four items scale adapted from Gironda and Korgaonkar (2018) since these items fit with the e-commerce context of the study. The four items are “I am concerned that retargeted ads are collecting too much information about me”, “I am concerned that the information collected about me for retargeted ads could be misused”, “I am concerned about collecting of my information by retargeting advertisers because of what others might do with it”, and “All things considered, I believe that my privacy is seriously threatened by retargeted advertising”. These items are on a seven-point rating scale ranged from 1 (“Strongly disagree”) to 7 (“Strongly agree”). Alpha reliability was .912.

3.5.2MEASURES EXPERIMENT 2

(26)

Furthermore, to check whether the manipulation of the privacy concerns in Experiment 2 was successful, privacy concerns of participants were measured on the same four items scale adapted from Gironda and Korgaonkar (2018) as in Experiment 1. However, the items were focused on the specific online firm in Experiment 2. The four items are “I am concerned that retargeted ads from HealthInsurance4You are collecting too much information about me”, “I am concerned that the information collected about me for retargeted ads from HealthInsurance4You could be misused”, “I am concerned about collecting of my information by HealthInsurance4You’s retargeted ads because of what others might do with it”, and “All things considered, I believe that my privacy is seriously threatened by retargeted ads of HealthInsurance4You”. These items were measured on a seven-point rating scale ranged from 1 (“Strongly disagree”) to 7 (“Strongly agree”). Alpha reliability was .931.

(27)

Finally, to check whether the manipulation of immediate gratification was successful during the experiment, the impatience of participants was measured on a three items scale. The three items are “As a person in this scenario I feel the urgency to choose a health insurance”, “During this scenario I felt impatient to make a choice about a health insurance”, and “During this scenario I felt rushed to choose a health insurance”. These items were measured on a seven-point scale ranged from 1 (“Strongly disagree”) to 7 (“Strongly agree”). Alpha reliability was .868.

3.6MANIPULATION CHECKS

Post experimental manipulation checks found that trust in the online retailer of Experiment 1, averaged across the ten items of Bhattacherjee (2002) on a seven-point scale from 1 (“Strongly disagree”) to 7 (“Strongly agree”), was not significantly higher in the high trust manipulated scenarios (M=4.6, SD=0.93) compared to the low trust manipulated scenarios (M=4.5,

SD=0.83), t(188)=0.51 and p=.305. Therefore, the manipulation of trust was unsuccessful.

The manipulation of privacy concerns in Experiment 2 was assessed with the averaged across the four items of Gironda and Korgaonkar (2018) on a seven-point scale from 1 (“Strongly disagree”) to 7 (“Strongly agree”). The privacy concerns of participants manipulated for low privacy concerns (M=3.74, SD=1.60) did not significantly differ from participants manipulated for high concerns (M=4.61, SD=1.47), t(141)= -3.41, p=.162. Therefore, the manipulation of privacy concerns was unsuccessful.

Lastly, the manipulation of immediate gratification in Experiment 2 was assessed with the averaged across three items about impatience during the experiment on a seven-point scale from 1 (“Strongly disagree”) to 7 (“Strongly agree”). The impatience of participants manipulated for immediate gratification (M= 3.92, SD= 1.71) did significantly differ from participants not manipulated for immediate gratification (M= 3.23, SD= 1.32), t(141)= 2,71, p=.046. Therefore, the manipulation of immediate gratification was successful.

3.7PLAN OF ANALYSIS

(28)

First of all, the data in Qualtrics gave an overview of the outcomes of each experiment and its survey. This overview allowed to inspect the used items of the survey and to do data cleaning. This data cleaning included a consistency check of the values of the data and finding missing responses. Depending on how much values were missing in the response there was chosen to exclude the outcomes or substitute it for a neutral value, which was the mean of all the answers (Malthora, 2010).

When the items were inspected and the data was cleaned, the descriptive statistics were shown since this gave an overview from the participants in the dataset. In the context of this study, the descriptive statistics showed the male/female ratio, the age distribution of the participants and the frequency of social media use.

Next, the constructs of the experiments were measured by multiple questions and through a correlation and a reliability analysis the items were checked if they could form one variable. In the correlation analysis there was checked if there was a correlation between the items, which meant that the p-value was lower than 0.05 to be significant. For the reliability analysis the Cronbach’s Alpha was used since this test determined whether the different items could be combined into one. The Cronbach’s Alpha should be at least 0.60 to combine the items in one variable. When the check to combine items into one variable was done, a new sum variable was constructed.

To test hypotheses 1 and 2 and find the effect of the interaction-effect a two-way ANCOVA-analysis was executed to compare the means of the four experimental conditions to each other and control for confounding variables. Ruling out variance was necessary since it was expected that both experiments were affected by confounding variables, the age of participants, the frequency of social media use and the relevance of the products in the experiment.

Firstly, the age of participants creates covariance since the use of social media by age is different (CBS, 2020a). Secondly, the frequency of social media use is not only per age different but also per person (De Best, 2020). That is why the frequency of social media use is a control variable in this study. Finally, the relevance of the products in the experiments was taken as a confounding variable since the value of the product will be different for each person.

(29)

experimental variables, since the manipulation checks showed that the manipulation of trust and privacy concerns were unsuccessful.

4.RESULTS

4.1RESULTS EXPERIMENT 1

To check Hypothesis 1, we observed the means and standard error (SE) of the various experimental conditions, since the descriptive information was not corrected for the covariates in this study. The estimated marginal means are summarized in combination with the number of participants per experimental condition in Table 3. Herein, the click-through intentions are shown per experimental condition and the click-through intentions were corrected for the covariates in this study, privacy concerns, the frequency of social media use, trust in retailers and the relevance of the offered product in the experiment. In line with hypothesis 1, the estimated marginal means information reveals that the data is acting in the expected way. Participants in the low trust condition had a higher click-through intention when they had immediate gratification (M=4.19, SE=0.23) compared to participants in the low trust condition with no immediate gratification (M=3.98, SE=0.20). Participants in the high trust condition had a higher click-through intention when they had no immediate gratification (M=4.51, SE=0.22) compared to participants in the high trust condition with immediate gratification (M=3.98,

SE=0.20). Therefore, it is likely that is, the stronger the immediate gratification, the less likely

one’s trust will influence the click-through rate.

Table 3. Average click-through intentions per experimental condition

Mean SE N Mean SE N High trust in retailer 3.93 0.22 48 4.51 0.22 47 Low trust in retailer 4.19 0.23 41 3.98 0.20 54

(30)

dependent variable was the click-through intentions. Privacy concerns, participants’ trust in online retailers, the frequency of social media use and the relevance of the product were used as covariates to control for individual differences. The ANCOVA showed that the interaction-effect was borderline significant (F(1, 182) = 3.18, p = .076, d High trust= 0.11, and d Low trust=0.00). The analysis provided therefore some initial support for Hypothesis 1.

In addition to the ANCOVA and to further validate the results of Hypothesis 1, the Process Regression of Hayes was conducted. The continuous variables of trust were included to check whether the direction of results was the same to the ANCOVA, since the manipulation of trust failed. Herein, the dependent variable is the click-through intentions of participants. The independent variable is the participants’ trust in retailers (i.e., continuous rather than the dummy-coded experimental variable) and the moderator is manipulated for immediate gratification. The relevance of the product, manipulated for trust and the frequency of social media use are the confounding variables. The results showed that it was not (borderline) significant, F(6, 183) = 6.11, p = .138. Therefore, using the continuous variable rather than the manipulation variable of trust resulted in rejecting Hypothesis 1.

4.2CONCLUSION AND COMMENTS IN RELATION TO EXPERIMENT 1

(31)

4.3RESULTS EXPERIMENT 2

To check Hypothesis 2, we observed the means and standard error (SE) of the various experimental conditions, since the descriptive information was not corrected for the covariates in this study. The estimated marginal means are summarized in combination with the number of participants per experimental condition in Table 4. Herein, the click-through intentions are shown per experimental condition and the click-through intentions were corrected for the covariates in this study, privacy concerns of participants, the frequency of social media use, trust in heath comparison websites and the relevance of the offered product in the experiment. In line with Hypothesis 2, the estimated marginal means information reveals that the data is acting in the expected way. Participants in the low privacy concerns condition had a higher click-through intention when they had immediate gratification (M=3.84, SE=0.29) compared to participants in the low privacy concerns condition with no immediate gratification (M=3.39,

SE=0.22). Participants in a high privacy concerns condition had higher click-through intentions

when they had immediate gratification (M=3.39, SE=0.26) compared to participants in a high privacy concerns condition with no immediate gratification (M=3.07, SE=0.24). However, the differences are small and therefore it is not likely that the stronger the immediate gratification, the less likely one’s privacy concerns will influence the click-through rate.

Table 4. Average click-through intentions per experimental condition

Mean SE N Mean SE N Low privacy concerns 3.84 0.29 25 3.39 0.22 46 High privacy concerns 3.39 0.26 32 3.07 0.24 40

(32)

ANCOVA showed that the interaction-effect was non-significant (F(1, 135) = 0.07, p= .791, d

High privacy concerns= -0.16, and d Low privacy concerns=-0.36). Therefore, the analysis showed that

Hypothesis 2 can be rejected.

Furthermore, the Process Regression of Hayes was conducted to further validate the results of the ANCOVA. Herein, the dependent variable is the click-through intentions of participants. The independent variable is the continuous variable of the participants’ privacy concerns and the moderator is the continuous variable of the participants’ immediate gratification. The product relevance, manipulated for privacy concerns and the frequency of social media use are the covariates. The results showed that it was not significant, F(5, 137) = 8.95, p =.632. Therefore, using the continuous variable rather than the manipulation variable of privacy concerns resulted in rejecting Hypothesis 2.

Finally, a Process Regression of Hayes was conducted to validate the results of Experiment 1, since trust was measured more specific during the experiment. Herein, the dependent variable was the click-through intentions of participants. The independent variable was the continuous variable of participants’ trust in online retailers and the immediate gratification of participants was the moderator. Furthermore, the covariates were participants privacy concerns, the frequency of social media use and the relevance of the product. The Process Regression of Hayes showed in Experiment 2 no moderating effect of immediate gratification both with immediate gratification as a continuous variable (F(6, 136)=7,86, t = -0.41, and p=.175) and a dummy code (F(6, 136)=6,08, t = -1.36, and p=.686). Therefore, the results of Experiment 2 could not validate the results of Experiment 1.

5.DISCUSSION

(33)

conducted to investigate the immediate gratification bias as a boundary condition in people’s response to retargeted ads on social media.

(34)

A possible explanation of the non-significance of the manipulation of trust is that trust was measured as the general trust in online retailers instead of measuring the trust in the specific retailer in the experiment. Another explanation could be that participants were not able to sufficiently imagine their research scenario. Therefore, they made their choices based on their own imagination instead of the given scenario, whereas the decision state of the buyer is important. Retargeted ads are more effective if the buyer is browsing goal-oriented and has narrowed down their preferences (Bleier, & Eissenbeis, 2015b; Lambrecht, & Tucker, 2013). So, when the participant did not imagine that he or she needs the shoe then the intended effect could not be met. Finally, the manipulation itself could be too less emphasized during the experiment or was not carefully read by the participants. On that account, participants missed (parts of) the manipulation in the experiment, and therefore they were not totally directed into the intended experimental condition.

(35)

5.1THEORETICAL CONTRIBUTIONS

This research contributed to the literature in two ways. First of all, the design of the study was not common in the OBA literature stream, since most research in the stream of OBA was conducted by surveys or interviews (Chen et al., 2019; Dehling, et al.,2019; Gironda, & Korgaonkar, 2018; Kaan, & Varnaali, 2019; Kim, & Huh, 2017; Li, & Huang, 2016; Smit et al., 2014; Van Doorn, & Hoekstra, 2013). By doing research based on surveys or interviews, a certain relation was explained. However, this research used two online experiments to investigate the causality of immediate gratification and trust on the click-through intentions (Experiment 1) and the causality of immediate gratification and privacy concerns on the click-through intentions (Experiment 2). The study found that there was only an interaction-effect between immediate gratification and trust on the effectiveness of OBA. Therefore, the study contributed to the OBA literature stream by drawing causal inferences instead of describing a relation.

Secondly, this study contributed to the literature stream of OBA by exploring the impact of the immediate gratification bias on people’s response to retargeted ads. Hereby, a new perspective was added to OBA, since the effectiveness of OBA was focused on the “traditional” economic theories such as the privacy calculus or the social exchange theory (Boerman et al., 2017). The assumption was that consumer decisions were made fully rational, informed and unbounded in processing all the information (Acquisiti, 2004). However, the results of this study showed that people are not completely rational in their decisions but that they were affected by personal biases. More specifically, the immediate gratification bias affected people’s clicking behaviour on OBA. Therefore, this study created a start to take into account the assumptions of the behavioural economic stream by investigating the effectiveness of OBA.

5.2PRACTICAL CONTRIBUTIONS

(36)

advantage for example online companies who focus on impulsive buying like Wish and AliExpress. These ecommerce platforms offer divers and low-priced assortment with all kinds of impulsive products called ‘must-haves’. Showing these ‘must-have’ products in an ad create immediacy to get the product and therefore the trust of the retailer is less important.

However, not all companies focus on impulsive buying, since consumers are more likely to rely on immediacy for some rather than on other products. Also, some companies are perceived as more trustworthy compared to others e.g., car sale companies are in general not really seen as trustworthy while Coolblue is regarded as very trustworthy in general (Consumentenbond, 2019). Therefore, these companies are recommended to focus on developing and improving the trustworthiness of their retargeting strategies. This could be done by creating transparency about how and when the company collected the personal information for the retargeted ads (Aguirre et al., 2015a; Bleier, & Eissenbeis, 2015a). Also, companies might benefit from only placing retargeted ads on trustworthy websites (Aguirre et al., 2015a).

5.3LIMITATIONS AND DIRECTIONS FOR FUTURE RESEARCH

The study had some limitations, which should be discussed and would provide new directions for future research. Firstly, in order to increase the generalizability of study and its findings, this study should be done on other social media platforms than Instagram such as Facebook, Twitter, Snapchat and LinkedIn. Findings could differ per platform since the characteristics of users per social media platform are different. The age of the average user is herein an important characteristic since younger people in general are more impatient and impulsive through the technological improvements they are used to (Beckwith, 2017). Therefore, the interaction-effect between immediate gratification, trust and privacy concerns on the interaction-effectiveness OBA could be stronger for a platform as Snapchat, since their average of the users is between the 15-19 (Bekkema, 2020). However, this study fitted to the group of the participants, since most of the participants were aged between 20 and 30 years old and this group is represented the most on Instagram in the Netherlands (Statista, 2020). So further research should be done by testing other social media platforms.

(37)

personalization is very narrow and deep. Other combinations of personalization depths and breaths could give different findings as Bleier and Eissenbeis (2015a) already show in their research about the importance of trust for personalized advertising. For instance, less depth in the personalized ad would give a weaker interaction-effect between the immediate gratification, trust and privacy concerns on the effectiveness of OBA, since the ad was less tailored to the consumers’ interests and preferences based on their past browsing behaviour (Bleier, & Eissenbeis, 2015a). So, future research should be done by testing other combinations of the depth and breaths of personalization.

(38)

LITERATURE

Abelhamid, M., Gaia, J., & Sanders, G. L. (2017). Putting the focus back on the patient: How privacy concerns affect personal health information sharing intentions. Journal of Medical

Internet Research, 19(9), 1-12.

Acquisti, A. (2004). Privacy in electronic commerce and the economics of immediate

gratification. In Proceedings of the 5th ACM Conference on Electronic Commerce, New York. Acquisti, A., Grossklags, J. (2005). Privacy and rationality in individual decision making.

IEEE Secur. Priv., 3(1), 26–33.

Aguirre, E., Mahr, D., Grewal, D., de Ruyter, K., & Wetzels, M. (2015a). Unravelling the personalization paradox: The effect of information collection and trust-building strategies on online advertisement effectiveness. Journal of Retailing, 91(1), 34-49.

Aguirre, E., Roggeveen, A. L., Grewal, D., & Wetzels, M. (2015b). The personalization-privacy paradox: implications for new media. Journal of Consumer Marketing, 33(2), 98-110. Alreck, P. L., & Settle, R. B. (2007). Consumer reactions to online behavioural tracking and targeting. Journal of Database Marketing and Consumer Strategy Management, 15, 11-23. An, S., Kang, H., & Jin, H. (2017). Self-regulation for online behavioral advertising (OBA): Analysis of OBA notices. Journal of Promotion Management, 24(2), 270-291.

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human

Decision Process, 50, 179-211.

Baek, T. H., & Morimoto, M. (2012). Stay away from me. Journal of Advertising, 41(1), 59-76.

Bandara, R., Fernando, M., & Akter, S. (2020). Explicating the privacy paradox: A qualitative inquiry of online shoppingconsumers. Journal of Retailing and Customer Services, 52, 2-9. Beckwith, S. (2017). The amazon effect: no longer a phantom menace. Logistics

(39)

Bekkema, S. (2020, 27 January). TikTok & Facebook (!) grootste verassingen in socialmedia-onderzoek 2020. Retrieved on 5 January from

https://www.frankwatching.com/archive/2020/01/27/social-media-onderzoek-2020/

Bhattacherjee, A. (2002). Individual trust in online firms: Scale development and initial test.

Journal of management information systems, 19(1), 211-241.

Boerman, S. C., Kruikemeier, S., & Zuiderveen Borgesius, F. J. (2017). Online behavioral advertising: A conceptual framework and research agenda. Journal of Advertising, 46(3), 363-376.

Bleier, A., & Eissenbeis, A. (2015a). The importance of trust for personalized advertising.

Journal of Retailing, 91(3), 390-409.

Bleier, A., & Eissenbeis, A. (2015b). Personalized Online Advertising Effectiveness: The Interplay of What, When, and Where. Marketing Science, 34(5), 669-688.

Braun, M., & Moe, W. (2013). Online display advertising: modelling the effects of multiple creatives and individual impression history. Marketing Science, 32(5), 753-767.

CBS (2020a, 20 January). More elderly active on social media. Retrieved on 29 October from https://www.cbs.nl/en-gb/news/2020/04/more-elderly-active-on-social-media

CBS (2020b, 16 December). Bevolkingspiramide. Retrieved on 28 December from https://www.cbs.nl/nl-nl/visualisaties/bevolkingspiramide

Chang, K., Chen, W., & Tan, B. (2012). Advertising effectiveness in social networking sites: social ties, expertise, and product type. IEEE Transactions on Engineering Management,

59(4), 634-643.

Chen, J., & Stallaert, J. (2014). An economic analysis of online advertising using behavioral targeting. MIS Quarterly, 38(2), 429-450.

(40)

Choi, Y., & Mai, D. Q. (2018). The Sustainable Role of the E-Trust in the B2C E-Commerce of Vietnam. Sustainability, 10(2), 1-18.

Clement, J. (2020, 15 July). Number of social media network users worldwide from 2017 to 2025. Retrieved on 18 September from https://www.statista.com/statistics/278414/number-of-worldwide-social-network-users/

Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences (2nd_{). New Jersey,}

USA: Lawrence Earlbaum Associates.

Consumentenbond (2019, 21 June). De beste webwinkels volgens online-shoppers. Retrieved on 5 January from https://www.consumentenbond.nl/online-kopen/de-beste-en-slechtste-webwinkels

Corbit, B. J., Thanasankit, T., & Yi, H. (2003). Trust and e-commerce: a study of consumer perceptions. Electronic Commerce Research and Applications, 2(3), 203-215.

Costello, C. (2019, May 13). Marketing Metrics: Consumers 70% More Likely to Convert with Retargeting. Retrieved from https://kenshoo.com/blog/retargeting-stats/

Daugherty, P. J., Bolumole, Y., & Grawe, S. J. (2018). The new age of customer impatience an agenda for reawakening logistics customer service research. International Journal of

Physical Distribution & Logistics Management, 49(1), 4-32.

De Best, R. (2020, 4 March). How many times a day do you actively use socia media? Retrieved on 29 October from https://www.statista.com/statistics/632736/daily-social-media-usage-in-the-netherlands-by-frequency-and-age-group/

Dehling, T., Zhang, Y., & Sunyaev, A. (2019). Consumer perceptions of online behavioral advertising. IEEE 21st Conference on Business Informatics, 1, 345-354.

Ecommerce Foundation (2018). Ecommerce Report The Netherlands 2018. Retrieved on 16 November on

(41)

Emerce Fashion (2019, 3 September). Online aandeel van schoenenaankopen aar 28 procent. Retrieved on 16 November on

https://www.emerce.nl/nieuws/online-aandeel-schoenenaankopen-28-procent

Faraoni, M., Rialti, R., & Zollo, L. (2019). Exploring loyalty Antecedents in B2C

e-commerce. Empirical results from Italian grocery retailers. British Food Journal, 121(2), 574-589.

Franco-Watkins, A. M., Mattson, R. E., & Jackson, M.D. (2016). Journal of Behavioral

Decision Making, 29, 206-217.

Gimpel, H., Kleindienst, D., & Waldmann, D. (2018). The disclosure of private data: measuring the privacy paradox in digital services. Electronic Markets,28, 475-490. Gironda, J. T., & Korgaonkar, P. K. (2018). iSpy? Tailored versus Invasive Ads and Consumers’ Perceptions of Personalized Advertising. Electronic Commerce Research and

Applications, 29, 64-77.

Gupta, N. (2019). Social media targeting: Opinions, clicks and Conversion. Abhigyan, 37(2), 1-9.

Hamby, A., & Ilyuk, V. (2019). A wolf in sheep’s clothing: how humanized, retargeted ads intrude in social contexts. The Journal of Consumer Affairs, 53(3),1167-1191.

Jung, A.R. (2017). The influence of perceived ad relevance on social media advertising: An empirical examination of a mediating role of privacy concern. Computer in Human Behavior,

70, 303-309.

Jones, S. C. (2007). Implications of behavioral decision theory for health marketing.

Marketing Theory, 7(1), 75-91.

Kim, K., & Zauberman, G. (2019). The effect of music tempo on consumer impatience in intertemporal decisions. European Journal of Marketing, 53(3), 504-517.

Kim, H., & Huh, J. (2017). Perceived Relevance and Privacy Concern Regarding Online Behavioral Advertising (OBA) and Their Role in Consumer Responses. Journal of Current

(42)

Kumar, V., & Gupta, S. (2016). Conceptualizing the evolution and future of advertising.

Journal of Advertising, 45(3), 302-317.

Kobsa, A., Cho, H., & Knijnenburg, B. P. (2016). The Effect of Personalization Provider Characteristics on Privacy Attitudes and Behaviors: An Elaboration Likelihood Model Approach. Journal of the Association for Infromation Science and Technology, 67(11), 2587-2606.

Lainer, C. D., & Saini, A. (2008). Understanding consumer privacy: A review and future directions. Academy of Marketing Science Review, 12(2), 1-48.

Lambrecht, A., & Tucker, C. (2013). When does retargeting work? Information specificity in online advertising. Journal of Marketing Research, 50(5), 561-576.

Li, C. (2016). When does web-based personalization really work? The distinction between actual personalization and perceived personalization. Computer in Human Behavior, 54, 25-33.

Liu, P.G., Wisdom, J., Roberto, C. A., Liu, L.J., & Ubel, P. A. (2013). Using Behavioral Economics to Design More Effective Food Policies to Address Obesity. Applied Economic

Perspectives and Policy, 31(1), 6-24.

Liang, R., Jiao, Z, & Liu, Z. (2020). An empirical study on the influencing factors of

customers’ acceptance intention towards online behavioral advertising. Technical Gazette, 27 (4), 1142-1149.

Malkoc, S. A., & Zauberman, G. (2018). Psychological analysis of consumer intemporal decisions. Consumer Psychology Review, 2(1), 97-113.

Malthora, N. K. (2010). Marketing Research (6th_{). New Jersey, USA: Pearson.}

(43)

NOS (2020, 25 January). Nederlanders zitten anderhalf uur per dag op sociale media.

Retrieved on 28 December from https://nos.nl/artikel/2320098-nederlanders-zitten-anderhalf-uur-per-dag-op-sociale-media.html

O’Donoghue, T., & Rabin, R. (2000). The economics of immediate gratification. Journal of

Behavioral Decision Making, 13, 233-250.

Ozturcan, S., & Gursoy, G. (2014). On-line reviews’ impact on trust building. Global

Conference on Business and Finance Proceedings, 9(1), 350-360.

Rijksoverheid (2020). Is een zorgverzekering verplicht? Retrieved on 16 November on https://www.rijksoverheid.nl/onderwerpen/zorgverzekering/vraag-en-antwoord/ben-ik-verplicht-een-zorgverzekering-af-te-sluiten

Sahni, N.S., Narayanan, S., & Kalyanam, K. (2019). An experimental investigation of the effects of retargeted advertising: The role of frequency and timing. Journal of Marketing

Research, 56(3), 401-418.

Sanje, G., & Sanol, I. (2012). The importance of online behavioral advertising for online retailers. International Journal of Business and Social Science, 18(3), 114-121.

Smit, E. G., Van Noort, G., Voorveld, H. A. M. (2014). Understanding online behavioural advertising: User knowledge, privacy concerns and online coping behaviour in Europe.

Computers in human behavior, 32, 15-22.

Simon, H.A. (1982). Models of Bounded Rationality: Empirically Grounded Economic

Reason (3rd_{ed.). MIT Press, Cambridge, MA.}

Stevenson, D., & Pasek, J. (2015). Privacy concern, trust, and desire for content

personalization. In TPRC 43: The 43rd Research Conference on Communication, Information

and Internet Policy.

Soper, D.S. (2020). A-priori Sample Size Calculator for Student t-Tests [Software]. Retrieved on 16 November from https://www.danielsoper.com/statcalc

(44)

https://www.statista.com/statistics/946096/instagram-user-distribution-in-the-netherlands-by-age-group-and-gender/

Sutanto, J., Palma, E., Tan, C., & Phang, C. W. (2013). Addressing the personalization-privacy paradox: An empirical assessment from a field experiment on smartphone users. Mis

Quarterly, 37(4), 1141-1164.

Tucker, C. E. (2012). The economics of advertising and privacy. International Journal of

Industrial Organization, 30(3), 326-329.

Tucker, C. E. (2014). Social networks, personalized advertising, and privacy controls. Journal

of Marketing Research,51(5), 546-562.

Utz, S., Kerkhof, P., & Van den Bos, J. (2012). Consumers rule: How consumer reviews influence perceived trustworthiness of online stores. Electronic Commerce Research and

Applications, 11, 49-58.

Van Doorn, J., & Hoekstra, J. C. (2013). Customization of Online Advertising: The Role of Intrusiveness. Marketing Letters, 24(4), 339– 351.

Varnali, K. (2019). Online behavioral advertising: An integrative review. Journal of

Marketing Communications, 1-22.

Wang, Y. D., & Emurian, H. H. (2005). Trust in e-commerce: considerations of interface design factors. Journal of Electronic Commerce in Organizations, 3(4), 42-60.

Wu, C., Huang, Y., & Hsu, C. (2013). Benevolence trust: a key determinant of user continuance use of online social networks. Inf Syst E-Bus Manage, 12, 189-211. Yan, J., Liu, N., Wang, G., Zhang, W., Jiang, Y., & Chen, Z. (2009). How much can behavioral targeting help online advertising? International World Wide Web Conference

Committee, 261-270.

The immediate gratification bias and online behavioural advertising in a social media context

The immediate gratification bias and online

behavioural advertising in a social media

context

Interactions between immediate gratification, trust and privacy concerns on

click-through intentions of retargeted ads

The immediate gratification bias and online

behavioural advertising in a social media

context

Interactions between immediate gratification, trust and privacy concerns on

click-through intentions of retargeted ads

Table of Contents