How personal is too personal : personalized advertising and the effects of device type mediated by privacy concerns

(1)

How personal is too personal: Personalized

advertising and the effects of device type

mediated by privacy concerns

Lisa Eastall

Student number: 11110422

Master Thesis

Graduate School of Communication

Master’s programme Communication Science

University of Amsterdam

Supervisor: Dr. Sophie Boerman

30 June 2017

Academic Year 2016 - 2017

Word count: 7415

(2)

2

Abstract

Personalized advertising is dominating the field of marketing. Although this marketing strategy has many positive effects like an increased motivation to process a personalized message, consumers may experience privacy concerns when their personal information is disclosed in an advertisement. Also, due to the development in highly personal online media types like mobile phones and desktops, marketers are able to personalize advertisements on a whole new level. But how personal is too personal? This study aims to find out if there is a difference between viewing a personalized ad on a desktop or mobile device and how privacy concerns interfere with the positive effects of this form of advertising. The personalization of the advertisement in this study will be based on a gender-sensitive cue. In this research an online experiment (N = 196) was conducted to find the effect of personalized advertising on consumers’ click-through intention, with privacy concerns as mediator and device type as a moderator. The results showed that there was no effect of the personalized advertisement on click-through intention. Privacy concerns were relatively high throughout all conditions but were not affected by the personalized ad and did not affect click-through intention. Mobile devices were perceived as more personal, but did not cause higher privacy concerns. This research discusses suggestions for future research and practical implications.

(3)

3

Introduction

How personal do we want advertisements to be? Thanks to personalized advertising marketers can now deliver the right content to the right person at the right time to maximize immediate and future sales (Tam, & Ho, 2006). This marketing strategy has received an increasing amount of interest by researchers (e.g. Tucker, 2014; Aguirre, Mahr, Grewal, de Ruyter, & Wetzels, 2015; Bang, & Wojdynski, 2016). This increase is due to two management trends: the growing emphasis on customer value and satisfaction; and the new technology applied to marketing, including data mining methods of consumer research (Kotler, & Armstrong, 2000). With these data mining methods, online data can be used to build detailed consumer profiles in real-time, at a low cost and create targeted advertisements (Cheung, Kwok, Law, & Tsui, 2003).

Moreover, this “right content” mentioned by Tam and Ho (2006) is made up of personal information of the consumer also known as personalization cues.When seeing these personalization cues, the message is processed in the context of self and therefore more cognitive activity occurs (Dijkstra, 2008). This marketing strategy therefore operates in line with the Elaboration Likelihood Model (ELM). The ELM states that when a message is more personally relevant cognitive activity increases and people engage in central processing (Petty, & Cacioppo, 1979), leading to positive effects like increased attention and even positive behavioral outcomes (Bang, & Wojdynski, 2016; Rimer, & Kreuter, 2006; Tam, & Ho, 2005).

A type of positive behavioral outcome in an online setting is click-through intention (see Aguirre et al., 2015; De Keyzer, Dens, & Pelsmacker, 2015). This is a measurable web based instrument, which reveals if the advertisement is positively perceived and if the consumer desires to interact with the advertisement in a positive way. This is an important evaluation tool that has been studied before in the context of personalized advertising (e.g. Aguirre et al., 2015; De Keyzer et al., 2015; Bright, & Daugherty, 2012). Thus, in line with the ELM and previous findings, this study expects that personalized advertisements has a positive effect on click-through intention.

A less positive development in the field of personalized advertising is the heightened privacy concerns of consumers (e.g., Awad, & Krishnan, 2006; Xu, Luo, Carroll, & Rosson,

2011; Bang, & Wojdynski, 2016). When personalized advertisements become too personal, consumers may feel that their freedom is being threatened (Bang, & Wojdynski, 2016). This is in line with the Reactance Theory, which states that when people feel their autonomy is

(5)

5

threatened they become motivated to act in a manner to restore this (reactance behavior) and feel psychological tension (i.e. privacy concerns; Brehm, & Brehm, 1981). Therefore, this study expects that when the ad is perceived as too personal, psychological tension will arise in the form of privacy concerns, which in turn will lead to a lower click-through intention as a reactance behavior.

Nevertheless, personalized advertising has become easier for marketers, thanks to new online media types like mobile phones and laptops. The individual and interactive characteristics of new online media let companies reach specific consumer groups or even individuals virtually anywhere, anytime, and based on the physical location of the mobile user (Chen , & Hsieh, 2012). But what kind of effect do the characteristics media types have on consumers in the context of personalized advertising? As we have learnt from McLuhan and Fiore (1967), it is important to understand that the medium used conveys a message in its self. Therefore, this research will investigate the different characteristics of these online devices in the context of personalized advertising.

In this day and age, consumers have a very intimate relationship with their mobile phone (Bauer, Reichardt, Bames, & Neumann, 2005). This could be explained by their portable size and the private information they hold. Another explanation for this intimate relationship could be the Theory of Personalization of Appearance (Bauer et al., 2005). This theory states that when consumers customize their products, this leads to emotional effects like emotional bonding (Bauer et al., 2005; Schultz et al., 1989; Mugge, Schoormans, & Schifferstein, 2009). Thus, this study aims to discover if consumers perceive their mobile phones to be more personal than their desktop devices and how this affects personalized advertising.

However, this uber-personal medium has a downside. Although mobile phones seem to be the ultimate medium for one-to-one marketing (Tsang, Ho, & Liang, 2004), this technology raises privacy issues due to its capability to collect, store and disclose personal information (Gratton, 2002 in Cleff, 2007). Accordingly, other studies found that consumers were more concerned about their privacy on mobile phones than on laptops (Chin, Porter Felt, Sekar, & Wagner, 2012). Therefore, in line with previous research, the Reactance theory and the assumption that mobile phones are perceived as more personal than laptops, this research expects that privacy concerns will be higher on mobile than on desktop devices.

This research aims to contribute to the knowledge about the different effects that channel types have in regard to personalized advertising and give recommendations to marketers about how to effectively personally target their online audiences. Thus the main

(6)

6

research question of this study is: how does personalized advertising affect click-through

intention on mobile versus desktop devices and is this mediated by privacy concerns?

Theoretical Framework

Personalized advertising and click-through intention

Personalization is a customer-orientated marketing strategy that aims to deliver the right content to the right person at the right time to maximize immediate and future sales (Tam, & Ho, 2006). This is achieved by creating communication based on information about the recipient, so they can identify themselves with the message (Kalyanaraman, Olivier, Magee, 2010; in Maslowska, Putte, & Smit, 2011).

Hence, the definition of personalization: “making something identifiable as belonging to a particular person” (“Personalize,” n.d.).This can be done by using personal information such as the consumers’ name, past buying history, demographics, psychographics, locations and lifestyle interests (Baek, & Morimoto, 2012). According to Bauer and Spiekermann (2011), the best foundation for successful personalized advertising is a profile-driven personalization based on socio-demographics. This means that advertising systems need to adapt to the situational context of a consumer (Bauer, & Spiekermann, 2011) and use more unique private information to personalize their message by creating individual consumer profiles (Tucker, 2014). Moreover, according to Howard and Kerin (2004) when an advertisement contained a consumer's first name, the consumer had a higher purchase intention for the product recommended by the advertisement.

However, in Aguirre et al. (2015) mixed effects were found when the personalization was based on demographics (i.e. age, gender and location). This study found that when advertisements are perceived as highly personal, click-through intention decreases: this is known as the personalization paradox (Aguirre et al., 2015).De Keyzer et al. (2015) also used gender as personalization cue. This study showed that gender induced a positive effect on the consumers’ response (in the form of attitude toward the brand and click-intention), but only when the consumer perceive the advertisement to be personally relevant. Due to these varied results this study is interested in using gender as a personalization cue.

It is apparent that there has been an increasing amount of research and utilization noticed over the past few years (e.g. Aguirre et al., 2015, Bang, & Wojdynski, 2016; Li, 2016). A major motivation for this global interest is directly related to the increasing doubts

(7)

7

about the effectiveness of traditional advertising methods (Jin and Villegas, 2007; in Yu, & Cude, 2009). These doubts have been noticed due to the increasing advertising clutter (Rotfeld, 2006; in Yu, & Cude, 2009) and consumers’ avoidance of advertising (Kim, & Pasadeos, 2007; in Yu, & Cude, 2009).

Another reason for this rise in interest towards personalized advertising or one-to-one marketing is due to two management trends: the increased emphasis on customer value and satisfaction; and the new technology applied to marketing, including data mining methods of consumer research (Kotler, & Armstrong, 2010). These data mining methods use online data to build detailed consumer profiles in real-time, at a low cost and to target advertisements (Cheung et al., 2003).

In practice marketers routinely use personalization, both offline and online. For example, when they refer to a customer by name or modify the service offering to accommodate customers’ needs (Shen, & Dwayne Ball, 2009). Practitioners realize that personalized services increase their business advantage by raising customer satisfaction (Rust, & Chung, 2006) and loyalty (Ansari, & Mela, 2003). In turn, this leads to improvements in the company’s profitability and higher content evaluations (Ansari, & Mela, 2003; Zhang, & Wedel, 2009).

The success of personalized advertising can be explained by people’s cognitive sensitivity to personal cues (Dijkstra, 2008). Moreover, an increase in cognitive activity occurs when seeing personal information in an advertised message which leads to self-referencing. Self-referencing means the content is processed in the context of self, making the message personally relevant (Dijkstra, 2008). In other words, personal information used in personalized advertising does not change the content of the message, but it is a signal which emphasizes that the message is addressed to the individual, aspiring to make the message more meaningful and hopefully more persuasive (Dijkstra, 2005; Dijkstra, 2008).

By implying that the content was made especially for you, the motivation to process a personalized message increases (Hawkins, Kreuter, Resnicow, Fishbein, & Dijkstra, 2008). This is in line with the Elaboration Likelihood Model (ELM). The ELM describes that there are two routes to attitude formation: the central and the peripheral route (Petty, & Cacioppo, 1979). When consumers are motivated and able to process the message they engage in the central route. In the context of personalized advertising, this means that when a message is perceived as more relevant (or personal), the cognitive ability and motivation to process the message increase (Petty, & Cacioppo, 1979). Central processing means they process the

(8)

8

message deeply and focus on its quality of arguments. This leads to lasting changes in attitude and more resistance to counterattacks (Petty, & Cacioppo, 1979).

In line with this theory, previous research has found personalized advertising to lead to persuasion effects like increased attention and even behavior changes (Bang, & Wojdynski, 2016; Rimer, & Kreuter, 2006; Tam, & Ho, 2005). In Bang and Wojdynski’s (2016) eye-tracking study, they measured how participants perceived personalized advertisements. This study showed that personalized advertisements attract more and longer attention that non-personalized ads. Therefore, personalization has a strong attention-grabbing effect (Bang, & Wojdynski, 2016). Moreover, Tam and Ho (2005) found that online personalization encourages consumers to pay attention to and engage with the content, and also behave in the desired way (e.g. purchase the product).

The objective of this research is to evaluate the effectiveness of personalized advertising based on gender, by measuring if there is a positive interaction with the advertisement. As in previous studies (e.g. Aguirre et al., 2015; De Keyzer et al., 2015; Bright, & Daugherty, 2012), this will be measured in the form of click-through intention. This is an online measuring instrument that shows if advertisements are positively perceived and if consumers desire to interact with the advertisement.

In line with the ELM and previous research (e.g. Aguirre et al., 2015; De Keyzer et al., 2015) it can be assumed that ads that are perceived as personally relevant lead to central processing and therefore to a positive intention to click on the advertisement. Thus, the following hypothesis has been formulated:

H1: The exposure to personalized advertisement (based on gender) causes a higher click-through intention than the exposure to a non-personalized advertisement.

Personalized advertising and privacy concerns

Unfortunately for marketers, personalized advertising does not always work as it is intended to. An explanation for the sometimes negative effects of personalized advertising, could be that consumers are becoming more worried about their privacy (e.g. Awad, & Krishnan, 2006; Baek, & Morimoto, 2012; Bang, & Wojdynski, 2016). According to Kluwer (2017), 82% of consumers identify privacy concerns as an unfavorable factor that impacted their purchase intention.

What exactly are privacy concerns? Privacy concerns can be defined as the degree to which a consumer is worried about the potential invasion of the right to prevent the disclosure of

(9)

9

personal information to others (Westin, 1968). This mental state arises from the feeling that they are not in control of their personal information and that this information is vulnerable (Dinev, & Hart, 2006). Because personalized advertising is the art of collecting and disclosing personal information, it makes sense to expect that privacy concerns might be triggered when consumers are confronted with a personalized advertisement. When consumers feel that they have lost control of their privacy (Nowak, & Phelps, 1997), it may result in resistance to sharing their personal information (Rubini, 2001). Also, they obtain a lower willingness to be profiled online for personalized advertising (Awad, & Krishnan, 2006).

This reaction can be explained by the Self-Determination Theory (SDT). This theory states that the innate needs for autonomy, relatedness and competence are central to human motivations (Deci, & Ryan, 2000). In the case of privacy concerns people feel the need for autonomy is being threatened. This need relates to the feeling of maintaining behavioral control. When this control is threatened, people become resistant to whatever is threatening their autonomy. This kind of resistance has been identified as the Reactance Theory (Brehm, & Brehm, 1981). The Reactance Theory states that when people feel their freedom is threatened, they become motivated to act in a manner to restore this. Reactance not only produces behavioral attempts to restore freedom, but also psychological tension (i.e. privacy concerns; Brehm, & Brehm, 1981).

An example of reactance behavior with regards to advertising was discovered by Baek and Morimoto (2012). They found that consumers dealt with their privacy concerns by using various advertisement avoidance tools like online ad blockers, filtering e-mail and subscribing to do-not-call programs (Baek, & Morimoto, 2012). Another example of reactance behavior is lower click-through intention. This became evident in Aguirre et al. (2015). Here consumers felt more vulnerable when they realize their information has been collected without their permission. This resulted in less click-through intention than when overt data collection strategies were used. Evidently, reactance behavior makes it harder for advertisers to reach their target audience and leads to undesired behavior. Therefore, this is an important concept for marketers to understand so they can act accordingly.

All in all, based on the Reactance Theory, this research expects that after seeing a personalized advertisement, consumers will experience a threat in their need for autonomy due to the personal information disclosed and therefore privacy concerns will be raised. Thus, the following hypothesis has been formulated:

(10)

10

H2a: Exposure to personalized advertising based on gender causes higher privacy concerns than exposure to non-personalized advertising.

It is important to note that, privacy concerns arise when the advertisement is perceived as too personal (Bang, & Wojdynski, 2016). Advertisers get too personal when they extend beyond friendly recognition and suggest an inappropriate level of familiarity with consumer’s preferences and behaviors (Bang, & Wojdynski, 2016). Privacy concerns might then offset or inhibit the positive effects of personalized advertisement (Phelps, Dsouza, & Nowak, 2001; Van Doorn, & Hoekstra, 2013 in Bang, & Wojdynski, 2016; Maslowska et al., 2011).

In line with Baek and Morimoto (2012), Maslowska et al. (2011) and the Reactance Theory this research expects that individuals with higher privacy concerns will engage in reactance behavior and therefore have a lower click-through intention. Based on this knowledge, the following hypothesis will be tested:

H2b: Higher privacy concerns cause a lower click-through intention.

Channel factors

In personalized (and general) advertising, it is important to understand why certain medium types are used and which effect these have on consumers. Each media type has unique characteristics which convey different marketing messages to consumers (Dijkstra, Buijtels, & Van Raaij, 2005). As McLuhan famously said “The medium is the message” (1967, p. 107), meaning that the medium itself conveys a particular message and that different types of media lead to different effects on consumers. As previously mentioned this study is interested in discovering the different effects of mobile and desktop devices in the context of personalized advertising.

Nowadays, most mobile phone users have a very personal (almost intimate) relationship with their phone (Bauer et al., 2005). This could be due to the different characteristics and functions of mobile phones. Unlike laptops, mobile phones are much smaller (pocket-size) making them ultra mobile and easy to take everywhere, plus they have a rich set of sensors like location and motion making them feel even more personal (e.g. a weather application; Kumar, Kim , & Helmy, 2013). Furthermore, consumers often use their phones as an alarm clock and keep it under their pillow when they sleep (Bauer et al., 2005). Also, mobile phones have been shown to have an impact on the sense of belonging due to

(11)

11

their mobility (Srivastava, 2005). Specifically, no matter where we are, we can stay connected with our friends and family.

A theoretical explanation for the intimate relationship with mobile phones could be the Theory of Personalization of Appearance. This theory states that user, system, and contextual dispositions lead to the personalization of appearance, which has cognitive, social, and emotional effects on the user (Blom, & Monk, 2003). Therefore, because consumers can express their individuality by customizing their phone model (brand, color, size etc.), constantly carry it around with them and because mobile phones contain highly personal information (like contacts, messages and important dates), consumers experience personal bonding with their phones (Bauer et al., 2005; Mugge, Schoormans, & Schifferstein, 2009).

According to Schultz et al. (1989), people are able to feel an emotional bond with a product. People then feel they are deeply attached to their possessions, resulting in positive emotions like happiness, love and pride. When feeling these emotions people tend to care for their products by repairing it frequently and procrastinate replacement (Schultz et al., 1989).

In line with the studies mentioned above and the Theory of Personalization of Appearance, it can be argued that mobile phones are experienced as very personal. Thus, the sub-research question is raised: are mobile devices perceived as more personal than desktop

devices?

Device type and privacy concerns

Thanks to the individual and interactive characteristics of new online media, marketers can now easily personalize advertisement content (Bauer et al., 2005; Scharl, Dickinger, Murphy, 2005; Leppäniemi, & Karjaluoto, 2005). According to Leppäniemi, & Karjaluoto (2005), personalization is easier to achieve in a mobile environment due to the highly personal nature and settings of a mobile phone, its location awareness and time sensitiveness.

However, this highly personal potential of the mobile channel has a downside. This advertising channel can evoke the ‘big brother’ feeling that someone is tracking your movements as well as buying behavior and then utilizing it in mobile advertising campaigns. From users’ point of view, invasion of privacy and general security concerns relating to wireless medium have been identified as one of the main obstacles to the success of online advertising (Leppäniemi, & Karjaluoto, 2005). Other studies have found that consumers generally have negative attitudes toward mobile advertising unless they have specifically consented to it (Tsang et al., 2014). Also, personalized mobile advertising, unless carefully monitored, may be extremely intrusive and lead to privacy issues (Cleff, 2007).

(12)

12

Privacy concerns have also been revealed in regard to desktop devices. According to Günther and Spiekermann (2005), consumers are concerned that their personal information can be accessed and tracked constantly. Also they are worried their information is disseminated and used in ways unknown to them (Günther, & Spiekermann, 2005).

When comparing the two devices, it can be argued that privacy concerns will be higher on mobile devices than on laptops, due to its highly personal characteristics. This assumption was tested in a study by Chin et al. (2012). In this study people were interviewed and asked about their privacy concerns on mobile and laptops. The results showed that participants were more concerned about their privacy on their phones than on their laptops. Furthermore, they were less willing to make shopping purchase, provide their Social Security Number or check their bank accounts on their mobiles (Chin et al., 2012). This was because many people were worried that they would lose their phone due to misplacement or theft, damage their phones or that a systematic error would occur and they would lose all their data (Chin et al., 2012).

Apparently, consumers experience their mobile phones to be more vulnerable than their laptops. Therefore, in line with these previous studies, the Theory of Personalization of Appearance and the Reactance Theory, this research expects that privacy concerns will be higher when seeing a personalized advertisement on a mobile device. Thus, the following hypothesis has been formulated:

H3: The exposure to personalized advertising seen on a mobile device causes higher privacy concerns than non-personalized advertising or a personalized advertisement on a desktop device (a) leading to a lower click-through intention (b).

(13)

13

In Figure 1 you can see the moderation mediated model that is proposed in this research.

Figure 1 Proposed moderation mediated model, in which especially a personalized

advertisement (vs. a non-personalized advertisement) seen on a mobile device causes privacy concerns (H3a), which consequently affects consumers' click-through intention (H3b).

Method

Design, participants and procedure

To test the hypotheses, an online experiment was conducted with a 2 (personalized vs. not personalized ad) x 2 (device: mobile vs. desktop) between-subjects design. In total, 196 participants (57.1% female; M age = 28.29 years, SD = 10.92) completed the questionnaire. For this experiment a convenient sample was used. These participants were recruited via Facebook and Whatsapp, and were randomly assigned to one of the four conditions (non-personalized mobile N = 54, non-(non-personalized desktop N = 42, (non-personalized mobile N = 62, personalized desktop N = 38). The participants either saw one of the two personalized advertisement (based on their gender) or the neutral advertisement. The first question therefore was: “what is your gender?”. Each participant only saw one advertisement.

Participation took about 4 minutes and there was no merit used to stimulate the participants. At the beginning of the experiment, participants were told that the subject of the research project was attitude towards online advertising and they were asked to imagine they were looking for a beach holiday. Participants then viewed one of the three beach holiday advertisements for at least five seconds due to a timer and subsequently filled out a

Device: -Mobile -Desktop

Click through intention Advertisement: -Personalized -Non-personalized Privacy concerns H1 H3a H2a H2b H3b

(14)

14

questionnaire. At the end of the questionnaire, participants were debriefed and thanked. Ethical approval for this study was granted by the review board of the research institute and data was collected in May 2017.

Stimulus material and manipulation

Personalized advertisement

The stimulus material consisted of personalized and non-personalized advertisements promoting beach holidays. The personalized advertisement was based on gender and therefore there were two ads (one for male and one for female participants). The text included in the ad for men was: “Dreaming of white sandy beaches? Click here to see your perfect holiday!” and for women: “In need of some peace and quiet? We’ve got the ideal holiday for you!”. The pictures used in these ads were personalized based on gender. Men saw a man and women saw a woman sitting on a beach chair. The third advertisement was not personalized. It showed two empty sun beds and the text stated: “Click here for beach holidays”. See Appendix A for the stimulus materials.

Device type

Participants were randomly asked to either use their mobile phone (59.2%) or their personal computer (40.8%) to complete the experiment. Furthermore, their actual device use was registered automatically.

Pretest

The stimulus material was pre-tested under a convenient sample of 29 participants (63% female; M age = 28.33 years, SD = 13.38). Six advertisements were tested under both males and females to test which advertisements they found the most personal to them. It turned out that the male group responded more positively to male models than the female group (M = 3.80, SD = 1.56; t(21) = 2.46, p = .023). Men felt the most personal ad was one with a male model (M = 5.50, SD = 1.60), this however did not differ significantly from the ad with the female model (M = 4.13, SD = 2.42; t(6) = 1.26, p = .25) but did with the ad of the empty beds (M = 2.06, SD = 1.73; t(17) = 3.44, p = .003). The most personal ad to women was the ad with a female model in it (M = 5.60, SD = 1.12). This was significantly more personal than the ad with the male model (M = 3.80, 1.57) and the empty sun beds (M = 2.23, SD = 1.83;

t(12) = -3.81, p = .002). As mentioned, both groups thought that the least personal

(15)

15

1.83). This advertisement was therefore used as a neutral condition. See Appendix B for the pre-test.

Measures

Perceived personalization

To measure to which extent the participants experienced the advertisement to be personalized, a four item seven-point scale (1 = Strongly disagree, 7 = Strongly agree) was used. The items used were: “I recognize myself in this situation.”, “The advertisement was directed personally to me.”, “The advertisement took my personal interests into account” and “I feel I can identify with the advertisement”. These items were inspired by the studies of Aguirre et al. (2015) and Dijkstra (2005) and formed a reliable scale (Eigenvalue = 3.03, explained variance = 75.62%,

α = .89, M = 3.70, SD = 0.66).

Mobile versus desktop

To measure if participants perceived their mobile phones as more personal than desktops, a three item seven-point scale (1 = Very impersonal, 7 = Very personal) was used. The statement was: how personal do you find the following devices? The items included were: mobile, laptop and tablet. Tablet was added as a control variable. This scale was also reliable (Eigenvalue = 1.77, explained variance = 59.00%, α = .65, M = 5.49, SD = 1.29).

Privacy concerns

To measure privacy concerns a seven-point Likert scale (1 = Strongly agree, 7 = Strongly

disagree) was used with five items. An example of an item is “In general, I believe that

personal data is misused often.”. An overview of these items can be found in Appendix C. These items were derived from the research of Baek and Morimoto (2012). A reliability analysis showed that these four items formed a reliable scale (Eigenvalue = 3.25, explained variance = 64.97%, α = .86, M = 5.3, SD = 0.62).

Click-through intention

To measure click-through intention participants were first told to image they had recently been searching for beach holidays. Subsequently, a single item was used to asses click-through intention: “Would you like to click on the advertisement to get further information?” measured by seven-point Likert scale (1 = No, definitely not, 7 = Yes, definitely). This scale is derived from Aguirre et al. (2015; M = 3.81, SD =1.91).

(16)

16 Control variables

Participants were asked if they have ever been on a beach holiday (Yes or No), how often they search for beach holidays online (1 = Never, 6 = Once a day; M = 2.42, SD = 1.05) and their attitude towards it, on a four item scale (Fun, Relaxing, Worthwhile and Enjoyable) using a seven-point Likert scale (1 = Strongly disagree, 7 = Strongly agree; Eigenvalue = 3.29, explained variance = 82.34%, α = .93, M = 5.73, SD = 0.65). Furthermore, participants were asked if they own a mobile phone and laptop computer (99.5% said yes to owning a mobile phone and 96.4% own a laptop). Other control variables were age, gender and education level. See Appendix C for the entire experiment questionnaire.

Results

Manipulation check

To check if the manipulation of the personalized advertisements was successful a one-way ANOVA was conducted. This showed that there was a significant difference in perceived personalization between conditions, F(3) = 3.82, p = .011. A post-hoc Bonferroni test indicated that the participants in the personalized desktop condition differed significantly from the non-personalized desktop condition (Mdifference = 0.78, p = .037). Furthermore, the

non-personalized mobile condition differed significantly with the non-personalized desktop condition (Mdifference = 0.78, p = .037). See table 1 for an overview of the perceived

personalization scores.

Randomization check

The four conditions did not significantly differ with regards to attitude towards beach holidays (F(3, 192) = 1.42, p = .238), if they own a mobile (χ² (3) = 2.17, p = .537) and desktop (χ² (3) = 2.56, p = .465), how often they search for beach holidays (χ² (15) = 16.40, p = .356) education level (χ² (21) = 23.05, p = .342) and age (F(3, 192) = .74, p = .528). They did significantly differ with regards to gender (χ² (3) = 10.56, p = .014). There were more females in the personalized desktop condition (65.8%) and the not personalized mobile condition (70.4%), and more men in the not personalized desktop condition (59.5%). That is why gender will be included as a covariate in all analyses.

(17)

17 Testing the hypotheses

To test hypotheses 1 and 2a a MANCOVA was conducted with click-through intentions and privacy concerns as dependent variable, personalization as fixed factor and gender as a covariate. Hypothesis one expects that the exposure to the personalized advertisement causes a higher click-through intention than the exposure to the non-personalized advertisement. And hypothesis 2a expects that the exposure to the personalized advertisement has a positive effect on privacy concerns. This analysis showed that after seeing the personalized advertisement, click-through intention (M = 3.94, SD = 1.90) was not significantly different than after seeing the non-personalized advertisement (M = 3.68, SD = 1.91), F(1) = 3.47, p = .307. There was also no effect found on privacy concerns after seeing the personalized advertisement (M = 5.27, SD = 1.23) versus the non-personalized advertisement (M = 5.39, SD = 1.18), F(1) = 1.23, p = .355. Therefore, hypothesis 1 and 2a are not supported.

Table 1 An overview of click-through, privacy concerns and personalization scores per

condition

Personalized Not personalized

Mobile Desktop Total Mobile Desktop Total

Perceived personalization 3.33(1.28)ac 3.78(1.55)a 3.51(1.40) 3.63(1.30)abc 2.85(1.41)ab 3.28(1.40) Click-through intention 3.79(1.88)a 4.18(1.93)a 3.94(1.90) 4.02 (1.84)a 3.24(1.94)a 3.68(1.91) Privacy concerns 5.14(1.31)a 5.36(1.08)a 5.27(1.23) 5.15(1.19)a 5.69(1.10)a 5.39(1.18)

Note: means with standard deviation between brackets. Where there is a letter they do not significantly differ: a= no significant difference between means, b = significant difference and c = significant difference.

To answer the research question about if mobile devices are perceived as more personal than desktop devices, a paired t-test was utilized. This showed that mobile devices (M = 6.18, SD = 1.10) were perceived as more personal than desktop devices (M = 5.51, SD = 1.25), t(195) = 6.85, p < .000.

Furthermore, this study wanted to test if privacy concerns have a negative effect on click-through intention (H2b). And it was expected that the advertisement seen on a mobile device causes higher privacy concerns than on a desktop device (H3a), leading to a lower

(18)

click-18

through intention on mobile (H3b). To test hypothesis 2b, 3a and 3b Model 7 of Hayes’ (2013) PROCESS macro in SPSS was used. The dependent variable in this analysis was click-through intention, with personalization as the independent variable, privacy concerns as mediator and device type as moderator (plus gender as covariate). See Figure 2 for the b-coefficient scores (and se) tested by Model 7.

Figure 2 B-coefficient scores of Model 7 of Hayes (2013) of personalized advertising on

click-through intention with privacy concerns as mediator and device type as moderator.

Note: se scores are between brackets.

The results show that privacy concerns have no direct effect on click-through intention (b = 0.05, se = 0.11, p = .659; see figure 2). Hypothesis 2b is therefore not supported. Furthermore, there was no interaction effect between device type and personalized advertisement (with privacy concerns as the dependent variable) found in the analysis (b = 0.47, se = 0.35, p = .180). Nor was there an indirect effect = 0.00, boot SE = 0.02, BCBCI [-.075, .035] of privacy concerns on click-through intention. Hypothesis 3a is therefore not supported. Furthermore, the analysis showed that there was no significant index of moderated mediation (Index = 0.023, boot SE = 0.06, BCBCI [-.069, .195]). Hypothesis 3b is also not supported.

Device: -Mobile -Desktop Privacy Concerns Click-through intention Advertisement: -Personalized -Not-personalized -0.27 (0.26), -0.27 (0.26) 0.05 (0.11) -0.53 (0.53) 0.47 (0.35)

(19)

19 Desktop

Since the personalization only worked on desktop (see manipulation check) and there was no interaction effect found of device type and personalization, this section shows the results of all the analyses but only for participants in the desktop group (N = 80).

To test if the personalized advertisement shown on a desktop device had an effect on click-through intention (H1) and on privacy concerns (H2a) a MANCOVA was conducted. This showed that there was no effect on click-through intention when seeing the personalized advertisement (M = 4.18, SD = 1.93) versus the non-personalized advertisement (M = 3.24,

SD = 1.94), F(1) = 9.44, p = .108. This analysis revealed that there was no effect on privacy

concerns when seeing the personalized advertisement (M = 5.36, SD = 1.08) versus the non-personalized advertisement (M = 5.69, SD = 1.10), F(1) = 3.54, p = .085.

Additionally, this study wanted to test if privacy concerns have a negative effect on through intention (H2b) on desktop devices. And if privacy concerns are raised and click-through intention is lowered, when seeing a personalized advertisement on a desktop device. In other words, this study wanted to test if there is a mediation effect of privacy concerns (H3). To test hypothesis 2b and 3 Model 4 of Hayes’ (2013) PROCESS macro in SPSS was used. The dependent variable in this analysis was click-through intention, with personalization as the independent variable and privacy concerns as mediator (plus gender as covariate). See figure 3 for the B-coefficient scores tested by Model 4.

The results of the analysis show that privacy concerns have no direct effect on click-through intention (b = -0.14, se = 0.20, p = .482). Hypothesis 2b is therefore not supported. Moreover, there was no total effect found of personalization on clickthrough intention (b = -0.71, se = 0.44, p = .108). There was no direct effect found of personalization on click-through intention controlling for privacy concerns (b = -0.65, se = 0.45, p = .151). There was no indirect effect = -0.06, boot SE = 0.10, BCBCI [-.394, .068] of privacy concerns on click-through intention. Hypothesis 3 is therefore not supported.

Conclusion

Due to new data mining technologies advertisers can be more personal than ever. But when are they too personal? The aim of this study was to investigate the effect of personalized advertising on click-through intention. In addition, the moderating role of device type (mobile or desktop) and mediating role of privacy concerns were taken into account. In contrary to

(20)

20

what was expected, the personalized advertisement did not lead to higher click-through intention than the generic advertisement. Furthermore, there was no mediating effect found of privacy concerns on click-through intention. Meaning that personalized advertisements did not increase privacy concerns, nor did higher privacy concerns lead to a lower click-through intention. Participants did feel that mobile phones are more personal than desktop devices. However, there was no moderating effect found of device type on privacy concerns; participants did not experience higher privacy concerns on mobile devices. This did not affect click-through intention. Therefore, no moderated mediation effect was found. Because the manipulation only worked on desktop devices, this condition was further tested as a subsample. However, in the desktop group again no effects were found of the personalized advertisement on privacy concerns or on click-through intention. Furthermore, privacy concerns did not mediate between the personalized advertisement and click through intention.

Discussion

Findings, limitations and suggestions for future studies Personalization and click-through intention

A reason that there were no effects found of personalized advertisement on click-through intention, could be that the personalized advertisement was not perceived as personal. Overall, the average score of perceived personalization in the personalized condition was under neutral and therefore relatively low (M = 3.51, SD = 1.40). Also, there was no difference found when comparing this to the non-personalized condition (M = 3.28, SD = 1.40). This means that gender was not perceived as an effective personalization cue.

Furthermore, the manipulation check showed that only participants in the desktop condition perceived a significant difference between the personalized and non-personalized advertisement. This could be due to practical reasons like the screen resolution of a desktop is higher than of a mobile. Moreover, the average recorded time for desktop was 287.97 seconds and for mobile this was 528.58 seconds. An explanation for this is that when people use their mobile phones they are easily distracted, e.g. when receiving messages or calls.

However, the perceived personalization scores in the desktop group were also under average and therefore relatively low (M = 3.78, SD = 1.55). Nevertheless, this group was analyzed separately, but again no effects were found. This could also be due to the ineffectiveness of gender as a personalization cue or due to the limitations of this study which will be discussed hereafter.

(21)

21

The findings of this study cannot confirm or reject the ELM, due to two reasons. The first reason is because the advertisement was not perceived as personal. It is therefore remains unclear if consumers engage in central processing when seeing a personalized advertisement. And the second reason is that even if the advertisement was perceived as personal, this research did not measure how participants processed the message.

A limitation of this research is the stimulus material was manipulated on more than one aspect. The text as well as the photographs differed between all three conditions. This means that even if there would have been an effect, it would be unclear what triggered this effect.

Another limitation with regards to the personalization of the advertisement was that all participants were told that they would see an advertisement based on their recent browsing behavior. This means that all conditions were personalized based on the participant’s fictitious browsing behavior. Because the advertisement was even so not perceived as personal, future researchers could explore if browsing behavior is an effective cue. Another explanation for this ineffectiveness could be that the ad was not based on the participants’ actual browsing behavior.

Furthermore, this study did not explicitly measure the attitude towards gender as a personalization cue. This could be helpful for future research and advertisers. Also the gender cue should be analyzed in a more gender-sensitive context than beach holidays, like clothing. To see if this personalization cue could have positive effects in a more relevant context.

High privacy concerns

The results show that privacy concerns were relatively high throughout the four conditions. Participants in the personalized condition averagely scored 5.27 (SD = 1.23) on privacy concerns, and the non-personalized condition had an average of 5.39 (SD = 1.18). These high scores were, however, not due to the personalized advertisement. That is why these findings are only partially in line with the Reactance Theory (Brehm, & Brehm, 1981). Participants did feel that their autonomy was threatened (due to high privacy concerns) and therefore had a low click-through intention (i.e. reactance behavior).

Privacy concerns seem to have been present regardless if they saw the personalized advertisement or not. Another study showed on a scale from 1 to 5 (1 = Not concerned, 5 =

Highly concerned) the participants scored a 2.75 on location based services and privacy

concerns (Barkuus, & Dey, 2003). More recent studies show higher privacy concerns about consumers’ location data being collected and used (M = 5.19, SD = 1.14) and concerns that

(22)

22

personal information is being distributed to third parties without consent (M = 6.32, SD = 0.99) (Eastin, Brinson, Doorey, & Wilcox, 2015). On the six item scale used by Baek and Morimoto (2012) participants scored between 5.48 and 6.08 on privacy concerns.

It therefore seems that consumers’ privacy concerns are currently high. However, according to the privacy paradox there is a difference between the intention people have to give out their personal information and their actual behavior (Norberg, Horne, & Horne, 2007). People may feel worried about their privacy but not do anything about it. So even though privacy concerns were high, this had no negative effect on click-through intention.

Another limitation concerning privacy concerns is that they were measured as a general concept, while this research was interested the difference experienced on mobile and desktop devices. This research used the channel type as an implicit measure, with no result. Therefore, future research should find a different way to measure privacy concerns for example by explicitly mentioning the device type. In addition, this study measured how high privacy concerns were, but not where they came from. Future studies could take more triggers of privacy concerns into account.

Effect of device characteristics

The outcomes of this research showed that mobiles are perceived as more personal than desktops. It is important to note that this was measured by a small three item scale. These findings are in line with Srivastava (2005) who pointed out that the personalization of a mobile phone can reflect self-identity and fulfill the need for relatedness. Furthermore, this outcome is also in line with the Theory of Personalization of Appearance. However, participants were not asked why they found mobiles to be more personal, only how personal they perceived each device. This could be an interesting subject for future research, so marketers can anticipate on these different characteristics. Two useful measures for future research on personal objects, are the amount of effort invested in and self-expression value of the object (Muggea et al., 2008). These were found to be reliable predictors of why people experienced an emotional bond with a product.

This study expected that mobile phones would threaten the need for autonomy and create privacy concerns more than desktop devices. This was not the case. Even though mobiles were perceived as more personal than desktops, privacy concerns were not higher on mobile phones. A reason for the lack of evidence for this outcome could be that the fulfillment of need of relatedness could outweigh the negative effects on the need of autonomy; or because consumers are aware of privacy issues but feel more in control and less

(23)

23

of an immediate threat due to the interactiveness of the web (Fournier, & Avery, 2011); or simply because the advertisement was not perceived as personalized, and participants therefore did not experience a threat.

Another reason could be that the measures used in this research were not specifically about laptop or desktop but about privacy concerns raised by advertising in general. Therefore, no conclusion can be drawn from these results about the difference in privacy concerns on mobile and desktop. Future research should focus on explicitly measuring privacy concerns on mobile phones and laptops in the context of personalized advertising.

Practical implications

Marketers need to understand that not all personalization cues are perceived as personal and do not always affect the consumer. This research shows that gender is not an effective personalization cue in the context of beach holidays. This could be because beach holidays are a neutral subject (not gender sensitive). However, for products like clothes gender could be an effective personalization cue, because it would then increase relevance for the consumer. Therefore, marketers need to pay great attention to the context and product that is advertised and make sure the consumer perceives the advertisement to be personalized and relevant to them.

Consumers have alarmingly high privacy concerns. However, these concerns do not necessarily result into negative or no behavior. Sometimes, it is important to inform the consumer that their personal information is going to be used in an advertisement by overt collection strategies. However, other times this can also result in resistance from the beginning of the advertising process. This highly depends on sensitivity of the personal information used.

An interesting outcome of this study is that mobile devices are perceived as more personal than desktop devices. This could be due to the higher customization options and more frequent use of mobile devices. Further investigation is necessary to understand what this means for advertisers in the field of personalized advertising. It could be that because mobiles are so personal, it would be better not to over-personalize advertisements to avoid counter effects like reactance behavior.

(24)

24

Aguirre, E., Mahr, D., Grewal, D., Ruyter, K. D., & Wetzels, M. (2015). Unraveling the personalization paradox: The effect of information collection and trust-building strategies on online advertisement effectiveness. Journal of Retailing,91(1), 34-49. doi:10.1016/j.jretai.2014.09.005

Ansari, A., & Mela, C. F. (2003). E-Customization. Journal of Marketing Research,40(2), 131-145. doi:10.1509/jmkr.40.2.131.19224

Awad, N. F., , & Krishnan, M. S. (2006). The personalization privacy paradox: an empirical evaluation of information transparency and the willingness to be profiled online for personalization. MIS quarterly, 13-28. doi: 10.2307/25148715

Baek, T. H., & Morimoto, M. (2012). Stay away from me. Journal of Advertising,41(1), 59- 76. doi:10.2753/joa0091-3367410105

Bang, H., & Wojdynski, B. W. (2016). Tracking users visual attention and responses to personalized advertising based on task cognitive demand. Computers in Human

Behavior,55, 867-876. doi:10.1016/j.chb.2015.10.025

Barkhuus, L., & Dey, A. K. (2003, July). Location-based services for mobile telephony: A study of users' privacy concerns. Interact (Vol. 3, pp. 702-712).

Bauer, C., & Spiekermann, S. (2011). Conceptualizing context for pervasive

advertising. Pervasive Advertising Human-Computer Interaction Series,159-183. doi:10.1007/978-0-85729-352-7_8

Bauer, H. H., Reichardt, T., Barnes, S. J., , & Neumann, M. M. (2005). Driving consumer acceptance of mobile marketing: A theoretical framework and empirical study. Journal of electronic commerce research, 6(3), 181.

Brehm, S. S., & Brehm, J. W. (1981). Introduction: Freedom, control, and reactance theory. Psychological Reactance,1-7. doi:10.1016/b978-0-12-129840-1.50005-3

(25)

25

Bright, L. F., & Daugherty, T. (2012). Does customization impact advertising effectiveness? An exploratory study of consumer perceptions of advertising in customized online environments. Journal of Marketing Communications,18(1), 19-37. doi:10.1080/13527266.2011.620767

Blom, J., & Monk, A. (2003). Theory of personalization of appearance: Why users

personalize their pcs and mobile phones. Human-Computer Interaction,18(3), 193-228. doi:10.1207/s15327051hci1803_1

Chen, P., & Hsieh, H. (2012). Personalized mobile advertising: Its key attributes, trends, and social impact. Technological Forecasting and Social Change,79(3), 543-557. doi:10.1016/j.techfore.2011.08.011

Cheung, K., Kwok, J. T., Law, M. H., & Tsui, K. (2003). Mining customer product ratings for personalized marketing. Decision Support Systems,35(2), 231-243. doi:10.1016/s0167-9236(02)00108-2

Chin, E., Felt, A. P., Sekar, V., & Wagner, D. (2012). Measuring user confidence in

smartphone security and privacy. Proceedings of the Eighth Symposium on Usable

Privacy and Security - SOUPS 12. doi:10.1145/2335356.2335358

Cleff, E. B. (2007). Privacy issues in mobile advertising. International Review of Law,

Computers & Technology,21(3), 225-236. doi:10.1080/13600860701701421

De Keyzer, F. D., Dens, N., & Pelsmacker, P. D. (2015). Is this for me? How consumers respond to personalized advertising on social network sites. Journal of Interactive

Advertising,15(2), 124-134. doi:10.1080/15252019.2015.1082450

Deci, E. L., , & Ryan, R. M. (2000). "The What and Why of Goal Pursuits: Human Needs and the Self-Determination of Behavior". Psychological Inquiry,11(4), 269-318. doi:10.1207/s15327965pli1104_02

(26)

26

from smoking cessation. Health Education Research,20(5), 527-539. doi:10.1093/her/cyh014

Dijkstra, A. (2008). The psychology of tailoring‐ingredients in computer‐tailored

persuasion. Social and personality psychology compass, 2(2), 765-784. doi: 10.1111/j.1751-9004.2008.00081.x

Dijkstra, M., Buijtels, H. E., & Raaij, W. V. (2005). Separate and joint effects of medium type on consumer responses: a comparison of television, print, and the Internet. Journal of Business Research,58(3), 377-386. doi:10.1016/s0148-2963(03)00105-x

Dinev, T., & Hart, P. (2006). An extended privacy calculus model for e-commerce

transactions. Information Systems Research,17(1), 61-80. doi:10.1287/isre.1060.0080

Eastin, M. S., Brinson, N. H., Doorey, A., & Wilcox, G. (2016). Living in a big data world: Predicting mobile commerce activity through privacy concerns. Computers in Human

Fournier, S. M., & Avery, J. J. (n.d.). The uninvited brand. SSRN Electronic Journal. doi:10.2139/ssrn.1963056

Günther, O., & Spiekermann, S. (2005). RFID and the perception of

control. Communications of the ACM,48(9), 73. doi:10.1145/1081992.1082023

Hawkins, R. P., Kreuter, M., Resnicow, K., Fishbein, M., & Dijkstra, A. (2008).

Understanding tailoring in communicating about health. Health Education

Research,23(3), 454-466. doi:10.1093/her/cyn004

Howard, D. J., & Kerin, R. A. (2004). The effects of personalized product recommendations on advertisement response rates: The “Try this. It works!” Technique. Journal of

Consumer Psychology,14(3), 271-279. doi:10.1207/s15327663jcp1403_8

(27)

27

Education.

Kluwer, W. (2017, February 01). Consumers are increasingly aware of data privacy issues. Retrieved June 23, 2017, from http://www.effacts.com/en/blog/consumers-are- increasingly-aware-of-data-privacy-issues/

Kumar, U., Kim, J., & Helmy, A. (2013). Comparing wireless network usage. Proceedings of

the 19th annual international conference on Mobile computing & networking - MobiCom 13. doi:10.1145/2500423.2504586

Leppaniemi, M., & Karjaluoto, H. (2005). Factors influencing consumers willingness to accept mobile advertising: A conceptual model. International Journal of Mobile

Communications,3(3), 197. doi:10.1504/ijmc.2005.006580

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

Maslowska, E., van den Putte, B., & Smit, E. G. (2011). The effectiveness of personalized e- mail and the role of personal characteristics. Cyberpsychology, Behavior, and Social

Networking, 14(12), 765-770. DOI: 10.1089/cyber.2011.0050

McLuhan, M., & Fiore, Q. (1967). The medium is the message. New York, 123, 126-128.

Mugge, R., Schoormans, J. P., & Schifferstein, H. N. (2009). Emotional bonding with

personalised products. Journal of Engineering Design,20(5), 467-476. doi:10.1080/09544820802698550

Norberg, P. A., Horne, D. R., & Horne, D. A. (2007). The privacy paradox: Personal

information disclosure intentions versus behaviors. Journal of Consumer Affairs,41(1), 100-126. doi:10.1111/j.1745-6606.2006.00070.x

(28)

28

information: Determining how and when privacy matters. Journal of Direct

Marketing,11(4), 94-108. doi:10.1002/(sici)1522-7138(199723)11:4<94::aid-dir11>3.0.co;2-f

Petty, R. E., & Cacioppo, J. T. (1979). Issue involvement can increase or decrease persuasion by enhancing message-relevant cognitive responses. Journal of Personality and Social

Psychology,37(10), 1915-1926. doi:10.1037//0022-3514.37.10.1915

Phelps, J. E., Dsouza, G., & Nowak, G. J. (2001). Antecedents and consequences of

consumer privacy concerns: An empirical investigation. Journal of Interactive

Marketing,15(4), 2-17. doi:10.1002/dir.1019

Rimer, B. K., & Kreuter, M. W. (2006). Advancing tailored health communication: A

persuasion and message effects perspective. Journal of Communication,56(S1). doi:10.1111/j.1460-2466.2006.00289.x

Rubini, D. (2010). Overcoming the paradox of personalization: Building adoption, loyalty, and trust in digital markets. Design Management Journal (Former Series),12(2), 49- 54. doi:10.1111/j.1948-7169.2001.tb00544.x

Rust, R. T., & Chung, T. S. (2006). Marketing models of service and

relationships. Marketing Science,25(6), 560-580. doi:10.1287/mksc.1050.0139

Scharl, A., Dickinger, A., , & Murphy, J. (2005). Diffusion and success factors of mobile marketing. Electronic commerce research and applications, 4(2), 159-173. doi: 10.1016/j.elerap.2004.10.006

Schultz, S. E., Kleine, R. E., & Kernan, J. B. (1989). ‘These are a few of my favorite things’: Toward an explication of attachment as a consumer behavior construct. Advances in

consumer research, 16(1), 359-366.

(29)

29

the role of technology‐mediated personalization (TMP) in service relationships. Journal of Services Marketing,23(2), 79-91. doi:10.1108/08876040910946341

Srivastava, L. (2005). Mobile phones and the evolution of social behaviour. Behaviour &

Information Technology,24(2), 111-129. doi:10.1080/01449290512331321910

Tam, K. Y., , & Ho, S. Y. (2006). Understanding the impact of web personalization on user information processing and decision outcomes. MIS quarterly, 865-890.

Tsang, M. M., Ho, S. C., , & Liang, T. P. (2004). Consumer attitudes toward mobile

advertising: An empirical study. International journal of electronic commerce, 8(3), 65-78.

Tucker, C. (n.d.). Social networks, personalized advertising, and privacy controls. SSRN

Electronic Journal. doi:10.2139/ssrn.1694319

Westin, A. F. (1968). Privacy and freedom. Washington and Lee Law Review, 25(1), 166.

Xu, H., Luo, X., Carroll, J. M., & Rosson, M. B. (2011). The personalization privacy

paradox: An exploratory study of decision making process for location-aware marketing. Decision Support Systems,51(1), 42-52. doi:10.1016/j.dss.2010.11.017

Yu, J., & Cude, B. (2009). ‘Hello, Mrs. Sarah Jones! We recommend this product!’

Consumers perceptions about personalized advertising: comparisons across advertisements delivered via three different types of media. International Journal of

Consumer Studies,33(4), 503-514. doi:10.1111/j.1470-6431.2009.00784.x

Zhang, J., & Wedel, M. (2009). The Effectiveness of customized promotions in online and offline stores. Journal of Marketing Research,46(2), 190-206. doi:10.1509/jmkr.46.2.190

(30)

30

Stimulus material

Personalized advertisement for women:

Personalized advertisement for men:

(31)

31

(32)

32

Pre-test

This study is about personalized advertising. You will be presented with six beach holiday advertisements and asked to rate each of these based on how personal you find them. This study should take about 1-2 minutes to complete. Your participation in this research is voluntary and anonymous. You have the right to withdraw at any time during the study. Please be assured that your answers will be kept completely confidential and will not be shared with third parties. Participating in the research will not entail your being subjected to any appreciable risk or discomfort.

Do you consent to participate in this research project?  I agree

 I disagree

What is your gender?  Male

 Female

(33)

33

How personal do you find this advertisement?

Answer: Very person al Person al A bit personal Neutral A bit impersonal Impersonal Very impersonal Advertisement 1       

(34)

34

Answer: Very personal Personal A bit personal Neutral A bit impersonal Impersonal Very impersonal Advertisement 2       

(35)

35

(36)

36

(37)

37

(38)

38

This is the end of the survey. Feel free to leave any comments below:

(39)

39

Experiment Personalized Advertising

Dear participant,

I would like to invite you to participate in a research study to be conducted under the auspices of the Graduate School of Communication, a part of the University of Amsterdam.

The subject of the study for which I am requesting your cooperation is online advertising. In the online survey, an advertisement will be displayed, then a few questions will be asked. The goal of this research is to generate insight into people's responses to online advertising. The survey should take around 6 minutes to complete.

As this research is being carried out under the responsibility of the ASCoR, University of Amsterdam, we can guarantee that: 1) Your anonymity will be safeguarded, and that your personal information will not be passed on to third parties under any conditions, unless you first give your express permission for this. 2) You can refuse to participate in the research or cut short your participation without having to give a reason for doing so. 3) Participating in the research will not entail your being subjected to any appreciable risk or discomfort, the researchers will not deliberately mislead you, and you will not be exposed to any explicitly offensive material. 4) No later than five months after the conclusion of the research, we will be able to provide you with a research report that explains the general results of the research.

For more information about the research and the invitation to participate, you are welcome to contact the project leader lisa.eastall@student.uva.nl at any time.

Should you have any complaints or comments about the course of the research and the procedures it involves as a consequence of your participation in this research, you can contact the designated supervisor s.c.boerman@uva.nl.

We hope that we have provided you with sufficient information. We would like to take this opportunity to thank you in advance for your assistance with this research, which we greatly appreciate.

I hereby declare that I have been informed in a clear manner about the nature and method of the research, as described in the email invitation for this study.

(40)

40

I agree, fully and voluntarily, to participate in this research study. With this, I retain the right to withdraw my consent, without having to give a reason for doing so. I am aware that I may halt my participation in the experiment at any time. If my research results are used in scientific publications or are made public in another way, this will be done such a way that my anonymity is completely safeguarded. My personal data will not be passed on to third parties without my express permission. If I wish to receive more information about the research, either now or in future, I can contact lisa.eastall@student.uva.nl. Should I have any complaints about this research, I can contact the designated member of the Ethics Committee representing the ASCoR, at the following address: ASCoR secretariat, Ethics Committee, University of Amsterdam, Postbus 15793, 1001 NG Amsterdam; 020 - 525 3680; ascor-secr-fmg@uva.nl.

I agree to participate I disagree

Only if the previous question was answered with “I agree to participate”, participants saw the following questions:

What is your gender? Male

Female

On the next page you will see an advertisement. Please study this advertisement carefully, look at the text as well as the picture.

(41)

How personal is too personal : personalized advertising and the effects of device type mediated by privacy concerns