1
The Effect of Retargeting on Website Evaluation:
How retargeted ads affect internet users’ evaluation of news websites
Zeynoun Albeik - 11397497
University of Amsterdam
Master’s thesis
Graduate School of Communication Persuasive Communication
Supervisor: mw. dr. M.L. Fransen Date of completion: 13/02/2018
2
Abstract
The use of online behavioral advertising is increasing in the advertising landscape. Retargeting is a form of online behavioral advertising in which Internet users are served with personally targeted advertisements . This study aims to investigate the effects of retargeted advertisements on website evaluation with intrusiveness as a mediator. Because it is expected that internet users have different levels of knowledge on advertising, advertising knowledge was analyzed to investigate whether this moderates the effect of retargeting on intrusiveness. 164 participants took part in an online experiment in which they were either assigned to one of the two conditions. All participants were presented with a scenario identical for both conditions. One condition was presented retargeted ads, while the other condition was presented non-retargeted ads. The analyses showed a significant effect of retargeting on intrusiveness. Additionally, a significant effect was found between intrusiveness and website evaluation. A significant mediating effect was found between advertising type and website evaluation through Intrusiveness. Lastly, advertising knowledge appeared to moderate the effect between retargeting and intrusiveness, such that lower levels of advertising knowledge were associated with higher levels of intrusiveness, contradicting expectations. The findings of the present study have important implications for news website owners renting out advertising space, as it appears retargeting negatively affects website evaluation through intrusiveness.
Keywords: Retargeting, online behavioral advertising, personalized advertising, intrusiveness, website evaluation
3
Introduction
As technology is progressing at a rapid pace, the modern-day internet landscape is changing as well. Over the past few years, the internet changed from being primarily used to find information, to a system that is more directed at social interaction, shopping, and consumer created content (Allen, 2013). Since consumers started using the internet has changed, the advertising landscape has changed accordingly. Companies are greatly investing in online advertising, as it is the most prominent form of advertising today. (Liebl, Rameseder & Rust, 2016).
Display advertising (banners on websites) and search advertising (Google and Bing advertisements) are currently the two dominant forms of online advertising (Xu, Duan, & Whinston, 2014). Both work with a bidding system that allows the advertiser to set a max
cost-per-click (CPC), indicating the maximum price to be paid for a certain ad. Increasing the
bid results in a higher chance of the ad being displayed on Google Search (search ads) or on a website that has allowed for display advertising. Thus, an insurance company that is willing to pay ten dollars for a click has higher chance of being displayed than an insurance company that is only willing to pay five dollars. (Google, 2017). In addition to display and search ads, advertisers have the option to target their advertisements based on consumer’s previous search behavior. This concept, called retargeting, aims to recapture the user’s interest of websites they have previously visited. This means that when consumers visit website X and leave without conversion, the same consumer is targeted in their future browsing behavior by ads intended to make the user return to website X (Lambrecht & Tucker, 2013).
While many consumers may not be familiar with the technology and processes behind online advertising, it seems inevitable to avoid it. A consumer who is looking to buy a certain dress online, might be retargeted by being shown the exact same dress as an advertisement on a completely different website the next day. Retargeting is an example of the many forms
4
of online advertising, and is usually presented in the form of displayed banner ads (Bleier & Eissenbeis, 2015).
Retargeting is a form of personalized advertising. The ads are based on previous online activity, such as clickstream data, search entries, and user profiles (De Bock & Van den Poel, 2010). Retargeted ads are specifically designed to fit the interests of the browsing consumer and should thus be more relevant. In turn, this can increase consumer response and consumer satisfaction, and should lead to higher effectiveness of online display ads. (Gironda & Korgaonkar; Athanasiadis & Mitropoulos 2010; Lambrecht & Tucker, 2013). Unfortunately, highly personalized ads have a downside. While, according to theory, retargeted ads may have numerous benefits for both consumers and advertisers, consumer’s privacy is compromised with the creation of retargeted ads. Retargeting is based on monitoring previous online activity, which means users might unexpectedly encounter display ads with the exact product they had looked at before. This causes a sense of
intrusiveness. Retargeted ads can be a double edged sword. On the one hand, personalized,
relevant ads can lead to higher purchase intentions (Barnard, 2014). On the other hand, the positive effect resulting from personalized, relevant ads is hampered when levels of intrusiveness get too high. This effect can in turn lead to lower purchase intentions (van Doorn & Hoekstra, 2013). Many studies have focused on the effect of retargeted ads on purchase intention and conversions (Bleier & Eissenbeis, 2015; van Doorn & Hoekstra, 2013; Lambrecht & Tucker), no studies have assessed the effect of retargeted ads on the advertising medium. As retargeted ads are proven to affect purchase intention through their effect on intrusiveness and irritation (van Doorn & Hoekstra, 2013; Li, Edwards & Lee, 2002), they might potentially raise consequences for the advertising medium accordingly. A broad range of websites are renting out advertising space that allow for the automatic programming of display advertising. The question remains what the effect is of display ads on the attitude of
5
the consumer towards the website that rents out advertising space. At first glance, it is not evident that the negative effects of retargeting can create a spillover effect, meaning that the advertising medium might accordingly be affected in a negative way. As this effect is not obvious, no research to date investigated the spillover effect of retargeting onto the advertising medium. It is therefore both interesting and refreshing to investigate how retargeted ads affect the consumer’s attitude towards the advertising medium. The current study therefore contributes to existing literature on online behavioral advertising by covering the potential spillover effect of retargeting on website evaluation. In addition, the results are expected to be of high interest to online news websites as these are websites that often rent out advertising space.
Because retargeted ads by definition include personal information, a certain extent of intrusiveness is inherent to a retargeted ad. Accordingly, it will be investigated if the effect of retargeting on website evaluation is mediated by the extent to which website users perceive a sense of intrusiveness. Furthermore, previous literature leads us to believe that persuasion knowledge can influence the effect of retargeting on intrusiveness (Tutaj & van Reijmersdal, 2012; Baek & Marimoto, 2012). In addition, one can reason that internet users possess different levels of knowledge on (online) advertising, which might affect user’s evaluation towards the website. Therefore, the extent to which a consumer is knowledgeable on concepts of online advertising is included as a moderator. Thus, the following research question was formulated:
RQ: What is the effect of retargeting on website evaluation and is this effect mediated by intrusiveness and moderated by advertising knowledge?
Retargeting and intrusiveness
The purpose of retargeted advertising is to tailor the ad specifically to the consumer’s needs in such a way that the content is highly relevant to the consumer’s needs. By fitting the
6
content to the needs of the consumer, the consumer is served the right ad at the right time (van Doorn & Hoekstra, 2013). At first blush, tailoring ads to the consumers’ needs seems to be beneficial for both the consumer and the advertiser. The advertiser can expect higher click through rates and the consumer is served relevant ads only. However, presenting consumers with personalized or retargeted ads automatically implies that a certain extent of data has been previously acquired. Thus, while the ad may be more relevant to the consumer when compared to generic, non-targeted ads, the consumer might perceive the ad as privacy invasive (van Doorn & Hoekstra, 2013; White, Zahay, Thorbjørnsen & Shavitt, 2008; Goldfarb, 2014). The personal data required to set up retargeted ads can therefore create a sense of intrusiveness among consumers, as this information might be considered private. When an ad is perceived as being too intrusive by the consumer, this can significantly lower the potentially positive effect of retargeting on purchase intention or even lead to negative effects (Li, et al., 2002).
Li et al. (2002) refer to intrusiveness as “a psychological reaction to ads that interferes with a consumer's ongoing cognitive processes”. In other words, consumers browsing a website might experience a sense of intrusiveness when they feel ads interrupt the processing of relevant information on the website. The study by Li et al. (2002) showed that irritation was related to the extent to which an ad was perceived to be intrusive, which is in turn unlikely to evoke positive attitudes and more likely to result in ad avoidance.
From an advertising perspective, intrusiveness consists of two components: The first component is the ‘fit’, meaning the extent to which the ad fits the needs of the consumers. The second component is personalization, which refers to the extent to which the ad incorporates personal information about the user (van Doorn & Hoekstra, 2013; Li et al. 2002). Retargeting is focused on displaying ads from websites that the user has previously visited and aims to make the user return to that same website. Retargeting is therefore
7
primarily focused on establishing a high fit with the consumer’s needs. As retargeting by definition requires the acquisition of personal browsing behavior, one could argue that retargeting will naturally lead to inducing a sense of intrusiveness among website users. Therefore, the following hypothesis will be tested:
H1: Participants who are exposed to retargeted ads will experience a higher sense of intrusiveness compared to participants who are exposed to non-retargeted ads
Retargeting and advertising knowledge
Although it is expected that retargeted ads lead to higher levels of intrusiveness, this effect can naturally be influenced by various factors. A person who is knowledgeable on the concepts of online behavioral advertising might respond differently to retargeted ads compared to a person who does not know much about advertising at all.
Retargeting is a form of online advertising in which the advertiser’s goal is to make the consumer return to the advertiser’s website and complete a conversion goal (Lambrecht & Tucker, 2013). In this process, the consumer is attempted to be persuaded by the advertiser to perform a certain behavior, which results in the consumer being given the choice to either accept or reject the persuasion attempt. The process of how one agent attempts to persuade a certain target was actively studied by Friestad and Wright (1994) under the term persuasion knowledge. The theory of persuasion knowledge refers to the process of how consumers are subjective to marketer’s persuasion attempts (Campbell & Kirmani, 2000; Friestad & Wright, 1994). The persuasion knowledge model (PKM) considers the interaction between advertiser and consumer as a game in which both the advertiser (agent) and the consumer (target) try to accomplish a goal. The model is created from the perspective that the target has intuitive theories about this interaction, and uses this to cope with persuasion attempts (Campbell & Kirmani, 2000).
8
A study by Tutaj and van Reijmersdal (2012) showed that the activation of persuasion knowledge varies for different advertising formats. Comparing banner ads (prominent ads) to sponsored content (subtle ads), banner ads were found to be easier to recognize as advertising because the intent and the source are more apparent. More importantly, participants reported a higher understanding of persuasive intent for the prominent ads. In addition, the study showed that the extent to which consumers understood the advertiser’s intent was positively correlated with the level of irritation experienced. Consumers that are irritated by an ad are inclined to avoid it as well (Li et al., 2002; Zufryden, Pedrick, & Sankaralingam, 1993). The occurrence of consumer’s irritation in ads is closely related to the concept of intrusiveness.
Regarding the type of advertisement the current study is investigating, two elements are important. Firstly, retargeted ads are normally presented in the form of banner ads (Bleier & Eissenbeis, 2015). Tutaj and van Reijmersdal (2012) found that banner ads are more prominent compared to sponsored content in the sense that the source and intent of the ad are more apparent. Along these lines, consumers have a higher understanding of the persuasive intent of prominent ads compared to less prominent ads. The level of understanding of persuasive intent is positively related to the level of irritation consumers experience when processing ads (Tutaj & van Reijmersdal, 2012). This will in turn lead to a higher perceived sense of intrusiveness (Li et al., 2002). When we combine these findings with the fact that retargeted ads can lead to higher perceived intrusiveness from a privacy invading perspective (Baek & Marimoto, 2012), it is expected that the extent to which someone understands persuasive intent moderates the effect of retargeted ads on intrusiveness.
As mentioned before, persuasion knowledge is based on intuitive theories the consumer has about marketer’s persuasion attempts. Throughout the literature, a broad range of concepts is used to explain persuasion knowledge. The most used conceptualizations are advertising literacy, understanding the persuasive or selling intent, recognition of commercial
9
content, advertising skepticism, and advertising avoidance (Bijmolt, Claassen, & Brus, 1998; Obermiller, Spangenberg, and MacLachlan, 2005; Lawlor & Prothero, 2008; Rozendaal, Buijzen, and Valkenburg 2010). For the current study, the relevant aspect of persuasion knowledge specifically and solely pertains to knowledge on (behavioral) advertising. Advertising knowledge will therefore be assigned as a moderator in the current study. Based on previous findings, it is expected that consumers with high knowledge of advertising and cookies will experience higher levels of intrusiveness compared to consumers with little knowledge on the subject. Therefore, the following hypothesis was formulated:
H2: The effect of retargeting on intrusiveness will be stronger for participants that have a high level of advertising knowledge, compared to participants with a low level of advertising knowledge.
Website evaluation
Retargeted ads aim to persuade people to return to a previous website. With this call for action aimed at the consumer, the consumer is left with a choice to either give in to the desired behavior, or to reject this behavior. The decision to accept or resist the behavior, is related to the concept of persuasion resistance. Knowles and Linn (2004) describe four elements of resistance: reactance, scrutiny, distrust, and inertia. For the current study, the concept of reactance can be useful in explaining the psychological process website users face when confronted with retargeted ads, as it is related to the motivational and affective aspects of resistance to persuasion (Knowles and Linn, 2004). Psychological reactance theory (Brehm, 1966) explains why people are motivated to resist persuasion attempts. According to the theory, when people feel that they are limited to behave freely, they are likely to experience reactance. This may in turn lead to the motivation to change their behavior and attitudes in such a way that a sense of autonomy and freedom is reassured (Baek &
10
Morimoto, 2012; Brehm, 1966; Brehm & Brehm, 1981). Li et al. (2002) investigated the theory of psychological reactance by exposing participants to unwanted pop-up ads. They found that the unwanted exposure leads to advertising avoidance. Along these lines, ads that contain personal information can induce psychological reactance as web users might perceive this as a limitation of free behavior (White et al. 2008).
Additionally, retargeted, intrusive ads can be experienced as annoying by the consumer, which can lead to reactance and undesired (rejecting) behavior (van Doorn & Hoekstra, 2013; Ying, Korneliussen & Grønhaug, 2009). Van Doorn and Hoekstra (2013) investigated the relationship between ad personalization and intrusiveness, and how this affects purchase intention. They found that ads that are highly fitting to the consumer result in higher purchase intentions. However, the more the ad is perceived as intrusive, the weaker the effect on purchase intention becomes. Thus, as mentioned before, retargeting can be a double edged sword: A highly relevant ad can result in higher purchase intentions but can simultaneously have a negative effect on purchase intentions through intrusiveness. According to a survey on privacy concerns by KPMG, less than 10% of internet users feel they have control over how firms use their data (2016). When web users feel that an ad is intrusive, they can experience this as an invasion of their privacy (Baek & Marimoto, 2012). Privacy concerns can have a negative effect on perceived information control, trust, and purchase intention (Milne & Boza, 1999).
As the aforementioned studies did not investigate reactance in relation to how this affects consumer’s attitude towards the medium, the current study aims to investigate how intrusiveness and reactance affects Internet user’s evaluation of the medium. No research to date has studied the effects of retargeting in relation to website evaluation. However, a study by Speck and Elliot (1997) investigated the effect of perceived ad clutter on attitude towards the medium. It was found that ad clutter negatively affects consumer’s attitudes towards the
11
media used. The hypotheses used by Speck and Elliot (1997) were based on a study by Ha (1996), who stated that ad clutter leads to intrusiveness. Additionally, Ha (1996) found that ad clutter leads to negative effects towards both advertising and advertising media. Based on these findings, it is expected that intrusiveness will lead to a less well-evaluated advertising medium. To assess this, the following hypotheses will be tested:
H3: Intrusiveness is negatively related to evaluation towards the medium
On the one hand, multiple studies have shown that retargeted ads can be beneficial concerning purchase intention, conversion rates, and ad effectiveness. On the other hand, the double-edged sword appears to be inevitable. Retargeted ads that fit the consumer’s needs inevitably tag along a sense of intrusiveness among consumers or internet users. As previously hypothesized, exposure to retargeted ads will induce a higher sense of intrusiveness compared to non-retargeted ads (H1). In addition, it was hypothesized that higher experienced levels of intrusiveness will negatively affect website evaluation. Taking these expectations together, it is expected that intrusiveness will mediate the effect of retargeting on website evaluation. Therefore, the following hypothesis was formulated:
H4: Participants who are exposed to retargeted ads (compared to participants who are exposed to non-retargeted ads) will experience a higher sense of intrusiveness, which will in turn lead to a less well-evaluated website
Conceptual model
Figure 1 displays the mediated moderation model used in the study. It shows the main effect of Advertising Type on Intrusiveness will be measured (H1), as well as the effect of Intrusiveness on Medium Evaluation (H3). Additionally, it will be investigated how
12 H1 H4 4 H2 H3
Advertising Knowledge moderates the effect of Advertising Type on Intrusiveness (H2). Lastly, the mediation effect of retargeting on medium evaluation through intrusiveness will be assessed (H4).
Figure 1: Mediated moderation model used for the study
Method
Participants and design
To answer the hypothesized model, an experiment was conducted among participants aged 18 to 80 years old. A total of 162 subjects participated in the experiment, of which 122 were included for further analysis. 19 participants were excluded because they did not complete the experiment and 21 participants were excluded because they did not provide the correct answer for the manipulation check. Of the 122 remaining participants 58.2% were female (N = 71) and 41.8% were male (N = 51). The age of the participants ranged from 18 to 66 years old (M = 24.93, SD = 6.13) Participants were randomly assigned to one of the two conditions (retargeted vs non-retargeted) of the between-subjects design. The moderator Advertising Knowledge was included as a continuous variable.
Medium evaluation (Attitude & Credibility)
Advertising type Intrusiveness
Advertising Knowledge
13 Procedure
The experiment was conducted online and did not require supervision. Convenience sampling was used to collect participants. Invitations were sent to participate through private messages on social media. Participants were required to read an informed consent and were asked whether they agree with participating before starting the experiment. Participants were then randomly presented one of two possible scenarios in which they were asked to look for a specific product. Hereafter, participants were presented the news website NRC, containing either retargeted or non-retargeted ads. Next, participants were asked to fill in the survey to evaluate the NRC website. Additionally, participants were asked to fill in demographical information such as age, gender, and level of education.
Retargeting
Participants were presented a scenario and were asked to imagine themselves being the person in that scenario. The person in the scenario would repeatedly have Internet connection problems. After discussing this problem with the subject’s roommates it was decided something had to be done about the connection problems. Thus, the person starts looking for a new Internet provider online (Ziggo) and was therefore browsing the Ziggo website. The next day, this person would browse the NRC website (appendix I shows the complete scenario used for the experiment). After reading the scenario, participants in both conditions were presented an image of the NRC homepage. However, participants in one condition would be retargeted, meaning the homepage contained banners displaying Ziggo advertisements. In the second, non-retargeted condition, participants were presented with banners displaying Huawei ads. Appendix II contains the manipulation materials used for the NRC homepage.
14
A pre-test was conducted (N = 24, 13 males, 11 females, Mage = 25.55) to find out which online Dutch news website was considered to be most neutral, meaning it did not evoke any strong positive or negative opinions beforehand. Both attitude and credibility were tested. Three news websites were tested; NRC was considered most appropriate for the current study (Attitude: M = 3.9; SD = .70; Credibility: M = 3.68; SD = .76 ) compared to Nu.nl (Attitude: M = 3.21; SD = 1.02; Credibility: M = 3.03; SD = .81 ) and NOS (Attitude:
M = 3.90; SD = .81; Credibility: M = 3.56; SD = .85 ). While the mean score for each of the
constructs is important to measure neutrality, the standard deviation serves as an important indicator here, as a low standard deviation indicates a low variance. The results of the pre-test showed that opinions towards Nu.nl varied widely while NRC showed a lower variance and standard deviation.
Website evaluation
The website was evaluated using two factors; attitude toward the website and perceived credibility of the medium. The scale for attitude was based on Machleit and Wilson’s (1988) scale to measure attitude. The scale for credibility was based on Samsup’s (2005) media credibility scale. Attitude towards the website was measured with three 5-point Likert scales. All scales were introduced by ‘I think NRC is’ and anchored by ‘disagree – agree’. The three items measuring attitude were favorable, good, and positive. The reliability of ‘attitude’ comprising three items was good: α = .86 (M = 3.63; SD = .66). Credibility towards the medium was measured with five 5-point Likert scales. The five items measuring credibility were believable, trustworthy, accurate, unbiased, and complete. The reliability of ‘credibility’ comprising five items was good: α = .87 (M = 3.52; SD = .70).
15 Intrusiveness (mediator)
Intrusiveness was measured with nine 5-point Likert scales, anchored by ‘ strongly disagree – strongly agree’, based on Edwards, Li, and Lee, (2002), and Mooradian, (1996): ‘The advertiser knows a lot about me’; ‘This offer gives me an uneasy feeling’; ‘This offer gives me an unsafe feeling’; ‘I think this offer is alarming’; ‘I think this offer is disturbing’; ‘I think this offer is obtrusive’; ‘I think this offer is irritating’; ‘I think this offer is annoying’; ‘I think this offer is uncomfortable’. The reliability of ‘intrusiveness’ comprising nine items was good: α = .90 (M = 3.31; SD = .81).
Advertising Knowledge (moderator)
Lastly, online advertising knowledge was measured in the final phase of the survey. The scales are based on Smit, Van Noort, and Voorveld (2013) item for knowledge on online behavioral advertising and knowledge about cookies. Online behavioral advertising was measured with eight true/false statements of which three were false. Statements that were answered correctly were coded ‘1’ and statements that were answered incorrectly were coded ‘0’. In addition, knowledge about cookies was measured with eight true/false statements of which five were false. Statements that were answered correctly were coded ‘1’ and statements that were answered incorrectly were coded ‘0’. For both knowledge of online behavioral advertising and knowledge about cookies, the total score of the correct answers (ranging between 0 and 16) indicated the value for online advertising knowledge (range: 9-15, M = 11.53).
Knowledge statements
None of the participants answered all statements correctly for both the Online Behavioral Advertising (OBA) knowledge statements and the cookie knowledge statements.
16
Table 1 shows the percentages for the participants who answered the OBA knowledge questions correctly. Table 2 shows the percentages for the cookie knowledge statements.
Table 1: OBA Knowledge statements (correct answers are marked ‘*’)
True % False % When I visit a website, I see the same ads as someone else visiting that
website
3.3 96.7*
Companies should only gather and store information about my Internet use (such as search terms, visited sites, online purchases) when I give them permission to do so
50.8 49.2*
The ads that appear on a website differ per visitor 98.4* 1.6 It is punishable for companies to gather and store information about the
Internet use of individuals.
18.0 82.0*
Your browsing history determines which ads you are going to see during your next visit
95.9* 4.1
Companies are allowed to store information about Internet use, provided that it is not traceable to a person
82.8* 17.2
Companies create different user segments based on their Internet behaviour, and they show these groups targeted ads
95.9* 4.1
Online content and services can be offered for free because of online advertising revenues
84.4* 15.6
Table 2: Cookies knowledge statements (correct answers are marked ‘*’)
True % False % Cookies collect browsing history; they save the websites you visited 77.9 22.1* A virus scanner prevents companies from storing information based on search 2.5 97.5*
17
behaviour, visited websites and online purchases
My browsing history is being saved by means of cookies 67.2 32.8* Cookies are used to place ads based on your Internet behaviour 94.3* 5.7 Software can ensure that cookies are automatically removed 76.2* 23.8 Cookies ensure, for instance, that your passwords are being stored 45.1* 54.9 Cookies are person-based; it is possible to relate the stored information to an
individual
51.6 48.4*
If cookies are not regularly removed, your computer will slow down 58.2 41.8*
Manipulation check
Manipulation checks were conducted for both the brand in the scenario (Ziggo) and the brand displayed in the advertisements (Ziggo or Huawei). For the scenario, participants were asked which brand was mentioned with answer options Tele2 (1), Ziggo (2), KPN (3), and Telfort (4). Regarding the advertisements, participants were asked which brand was displayed in the advertisements on the NRC homepage. Answer options were Nokia (1), KPN (2), Apple (3), Ziggo (4), Telfort (5), Samsung (6), Huawei (7), XS4ALL (8), Blackberry (9), and Tele2 (10).
Results
Manipulation check
The manipulation checks for the scenario and the advertisements were successful. In total, 21 participants (13%) were excluded from the study because they did not provide the right answer for one or both of the manipulation checks.
18 Control variables
A Chi-square showed that the variables gender (χ2(1) = .37, p = .541), website visiting frequency (χ2(3) = 4.60, p = .203), and education level (χ2(5) = 2.96, p = .706) were equally distributed between the two conditions. A one-way analysis of variance showed there was no significant relation between advertising type and age (F(1, 120) = .573, p = .451). Based on these results, none of the control variables were included for further analysis.
Hypothesis testing
To test the hypotheses for the present study, PROCESS (Hayes, 2013; model 7) was used. The first hypothesis was tested to see if retargeted ads lead to higher levels of intrusiveness. The analysis showed a significant effect of Advertising Type on Intrusiveness (b = -.21, SE = .08, p =.012, 95%BCBCI [-.38, -.05]). Retargeted ads lead to higher levels of intrusiveness compared to non-retargeted ads, confirming H1.
The second hypothesis was tested to see if Advertising Knowledge moderates the effect of Advertising Type on Intrusiveness. The analysis showed a significant interaction effect between Advertising Type and Advertising Knowledge on Intrusiveness (b = -.21, SE = .08, p =.012, 95%BCBCI [-.38, -.05]). Contrary to the hypothesis, a stronger effect on Intrusiveness occurred for lower levels of Advertising Knowledge (SE = -.21, 95%BCBCI [-.47, .01]) compared to higher levels of Advertising Knowledge (SE = -.14, 95%BCBCI [-.33, .01]). Thus, H2 can be rejected. Table 3 shows the values relating to three different levels of Advertising Knowledge.
19
Table 3: Values relating to three different levels of Advertising Knowledge
Ad Knowledge SE 95%BCBCI
9.87 -.21 -.47, -.01
11.53 -.14 -.33, -.01
13.20 -.08 -.25, .01
The third hypothesis (H3) was assessed to test the effects of Intrusiveness on both Attitude and Credibility Towards the Medium. The analysis showed a significant negative effect of Intrusiveness on Attitude Towards the Medium (b = -.18, SE = .08, p =.027, 95%BCBCI [-.35, -.02]). Accordingly, the analysis showed a significant negative effect of Intrusiveness on Credibility Towards the Medium (b = -.22, SE = .09, p =.014, 95%BCBCI [-.39, -.04]). Thus, intrusiveness was negatively related with both Attitude and Credibility towards the medium, accepting the third hypothesis.
The fourth hypothesis (H4) was tested to see if Intrusiveness mediated the effect between Advertising Type and Website Evaluation. Again, this analysis was done for both Attitude and Credibility Towards the Medium.
For Attitude, the mediation model was found significant and showed a significant indirect effect of Advertising Type on Credibility Towards the Medium through Intrusiveness (b = -.14, SE = .08, 95%BCBCI [-.33, -.01]). In other words, Advertising Type (retargeted ads) was positively related with intrusiveness (b = -.21, SE = .08, p =.012, 95%BCBCI [-.38, -.05]), which in turn led to a lower Attitude Towards the Medium (b = -.18, SE = .08, p =.027, 95%BCBCI [-.35, -.02]).
For Credibility, the mediation model was found significant and showed a significant indirect effect of Advertising Type on Credibility Towards the Medium through Intrusiveness (b = -.17, SE = .08, 95%BCBCI [-.35, -.04]). In other words, Advertising Type (retargeted
20
ads) was positively related with intrusiveness (b = 3.25, SE = .97, p =.001, 95%BCBCI [1.32, 5.17]), which in turn led to a lower Credibility Towards the Medium (b = -.18, SE = .08, p =.027, 95%BCBCI [-.35, -.04]). As the mediation was found significant for both Attitude as dependent variable and Credibility as dependent variable, the fourth hypothesis was supported by the data.
In addition, no significant direct effects were found between Advertising Type and Attitude Towards the Website (b = -.01, SE = .13, p =.987, 95%BCBCI [-.27, .26]) and between Advertising Type and Credibility Towards the Website (b = -.02, SE = .14, p =.868, 95%BCBCI [-.31, .26]). This means the mediation was partial.
Additional analyses
Additional analysis in PROCESS (Hayes, 2013; Model 7) showed that the moderated mediation model, including Advertising Type as independent variable, Advertising Knowledge as moderator, and Intrusiveness as mediator, was significant when tested with Attitude Towards the Medium as dependent variable (SE = .03, 95%BCBCI [.01, .11]), as well as when tested with Credibility Towards the Medium as dependent variable (SE = .08, p =.027, 95%BCBCI [.01, .12]).
Lastly, the analysis did not show a significant effect of Advertising Knowledge on Intrusiveness (b = .99, SE = .07, p =.159, 95%BCBCI [-.04, .24]).
Conclusion and Discussion
The purpose of this study was to examine the effect of retargeted ads on website evaluation. The first point of interest was to see how retargeted ads affect perceived intrusiveness of Internet users. The second point of interest was to see if higher levels of intrusiveness lead to a more negative evaluation of the medium the ads were presented on.
21
The third point of interest was to see whether intrusiveness mediates the effect of retargeted ads on website evaluation. Lastly, the effect of advertising knowledge was analyzed to see if it has a moderating effect of retargeting on intrusiveness.
First of all, the findings of the current study show that there is no direct effect of Advertising Type on Website Evaluation. Retargeted ads did not affect attitude and credibility towards the NRC website. Second, retargeted ads were shown to lead to higher levels of Intrusiveness, confirming the first hypothesis. Third, Advertising Knowledge was found to have a moderating effect of Advertising Type on Intrusiveness. However, the significant interaction effect was contradictory to expectations: Lower levels of Advertising Knowledge were associated with higher levels of Intrusiveness compared to higher levels of Advertising Knowledge. Thus, the second hypothesis was rejected. Next, the analyses showed that Intrusiveness had a significant effect on Website Evaluation. Both attitude and credibility towards the website were negatively affected by Intrusiveness, confirming the third hypothesis. Next, a partial mediation effect was found for the model using Attitude Towards the Website as dependent variable, as well as for the model using Credibility Towards the Website as dependent variable, confirming the fourth hypothesis. Lastly, the moderated mediation model as a whole did not appear to be significant.
Comparison with the literature
While no research to date has investigated the effect of retargeting on website evaluation, it was hypothesized that retargeted ads would lead to a less well evaluated website because of higher levels of intrusiveness. The effect of retargeting on intrusiveness was found significant in the sense that retargeted ads lead to higher levels of intrusiveness. This is in line with previous studies that found that personalized or retargeted ads are privacy invasive and can therefore induced a sense of intrusiveness (van Doorn & Hoekstra, 2013;
22
White, Zahay, Thorbjørnsen & Shavitt, 2008; Goldfarb, 2014). In addition, several studies found intrusiveness to be negatively related to purchase intention (Bleier & Eissenbeis, 2015; van Doorn & Hoekstra, 2013; Lambrecht & Tucker). Additionally, Ha (1996) found that ad clutter led to a less positive attitude towards the medium.
For the present study, it was therefore expected that the negative effects of retargeting could be projected onto the medium. Accordingly, Intrusiveness was found to negatively affect Website Evaluation. Despite the fact that the hypotheses for the present study were based on results mainly associated with the relationship between retargeting, intrusiveness, and directly related aspects such as purchase intention, the leap of projecting these results onto possible effects towards the medium was justified by the results. In addition, a mediating effect was found of retargeting on website evaluation through intrusiveness.
The fact that retargeted ads indirectly affect website evaluation has important implications for websites that allow for programmatic display advertising. As no studies to date have investigated this effect, company owners might not be aware of the potential negative effects of banners and other forms of programmatic display advertising. For news websites in particular, one could argue that maintaining a good image is important. The results of the present study can therefore assist website owners in making an informed decision on whether to implement programmatic display advertising on their website.
Regarding the moderation, the interaction effect contradicted the hypothesis, as lower levels of advertising knowledge led to higher levels of intrusiveness. This contradicts expectations and is not in line with previous findings which suggested that a higher understanding of the persuasive intent of prominent ads is positively related to the level of irritation consumers experience (Tutaj & van Reijmersdal, 2012), which could lead to higher perceived intrusiveness (Li et al., 2002), particularly from a privacy invading perspective (Baek & Marimoto, 2012).
23 Limitations
The present study aimed to explore the effects of retargeted advertising on the advertising medium. To date, no studies have focused on the relationship between online behavioral advertising and the possible effect this modern form of advertising can have on perceived attitude and credibility towards the medium. The hypotheses used for this study were mainly based on findings regarding personalization, intrusiveness, purchase intention, and conversion rates. Projecting these findings onto certain expectations towards the advertising medium can, despite significant findings, be considered a big leap. Literature assessing the spillover effect of for example attitude towards advertisements onto attitude towards the advertising medium is scarce.
Another limitation for the present study might be the uniformity of the sample. Participants were collected solely through social media networks and mainly consisted of students or recent graduates, meaning that education level and age did not vary much. Taking these facts into consideration, the external validity of the present study could be compromised. Furthermore, while the field of education was not part of the survey, it is expected that a large part of the sample were Communication Science students, which could first of all explain the high mean score for advertising knowledge. Second, a large part of the sample being Communication Science students or graduates would also mean participants had extensive knowledge on marketing techniques such as retargeting. Thus, the sampling might have compromised both internal and external validity. Furthermore, the survey questions did not include a question as to which participants owned a Ziggo membership or owned a Huawei phone. It could be argued that ownership of one or both of these products might have influenced participant’s attitude towards one of the examined variables. Lastly, this study was limited to creating a retargeted experimental condition through a scenario only.
24 Recommendations
Future studies could focus on the different aspects that underlie the possible effects of retargeting on the advertising medium. In other words, the steps towards investigating the potential effects of online behavioral advertising on advertising media should initially focus on how attitude towards online behavioral advertising can be projected onto advertising media. Further avenues of research could compare different forms of online advertising and how they affect both attitude and credibility towards advertising and advertising media. Regarding advertising knowledge, future research could investigate what the effects are when the sample is more representative of society as a whole, meaning age, income, and education are more varied.
This study aimed to contribute to the literature about online behavioral advertising by investigating the effect of retargeting on medium evaluation, through intrusiveness. As many results turned to be insignificant, this might imply that assumptions and hypotheses have to be formulated more gradually in this relatively new advertising landscape.
25 References
Allen, M. (2013). What was Web 2.0? Versions as the dominant mode of internet history. New Media & Society, 15(2), 260-275.
Athanasiadis, E., & Mitropoulos, S. (2010). A distributed platform for personalized advertising in digital interactive TV environments. Journal of Systems and Software, 83(8), 1453-1469. Barnard, L. (2014). The cost of creepiness: How online behavioral advertising affects consumer
purchase intention (Doctoral dissertation, The University of North Carolina at Chapel Hill). Baek, T. H., & Morimoto, M. (2012). Stay away from me. Journal of advertising, 41(1), 59-76. Bijmolt, T. H., Claassen, W., & Brus, B. (1998). Children's understanding of TV advertising: Effects
of age, gender, and parental influence. Journal of Consumer Policy, 21(2), 171-194.
Bleier, A., & Eisenbeiss, M. (2015). The importance of trust for personalized online advertising. Journal of Retailing, 91(3), 390-409.
Bock, de, K., & Poel, van, den, D. (2010). Predicting website audience demographics forweb advertising targeting using multi-website clickstream data. Fundamenta Informaticae,
98(1),49-70.
Brehm, J. W. (1966). A theory of psychological reactance.
Brehm, S.S., & Brehm, J.W. (1981). Psychological reactance: A theory of freedom and control. New York: Academic Press.
Campbell, M. C., & Kirmani, A. (2000). Consumers' use of persuasion knowledge: The effects of accessibility and cognitive capacity on perceptions of an influence agent. Journal of consumer research, 27(1), 69-83.
Van Doorn, J., & Hoekstra, J. C. (2013). Customization of online advertising: The role of intrusiveness. Marketing Letters, 24(4), 339-351.
Edwards, S. M., Li, H., & Lee, J. H. (2002). Forced exposure and psychological reactance: Antecedents and consequences of the perceived intrusiveness of pop-up ads. Journal of Advertising, 31(3), 83-95.
26 Friestad, M., & Wright, P. (1994). The persuasion knowledge model: How people cope with
persuasion attempts. Journal of consumer research, 21(1), 1-31.
Goldfarb, A., & Tucker, C. (2011). Online display advertising: Targeting and obtrusiveness. Marketing Science, 30(3), 389-404.
Goldfarb, A. (2014). What is different about online advertising?. Review of Industrial Organization, 44(2), 115-129.
Ha, L. (1996). Advertising clutter in consumer magazines: Dimensions and effects. Journal of advertising research, 36(4), 76-85.
Knowles, E. S., & Linn, J. A. (Eds.). (2004). Resistance and persuasion. Psychology Press.
KPMG International (2016). Creepy or Cool? Staying on the right side of the consumer privacy line. Retrieved from: https://assets.kpmg.com/content/dam/kpmg/xx/pdf/2016/11/creepy-or-cool.pdf
Lawlor, M. A., & Prothero, A. (2008). Exploring children's understanding of television advertising– beyond the advertiser's perspective. European Journal of Marketing, 42(11/12), 1203-1223. Li, H., Edwards, S. M., & Lee, J. H. (2002). Measuring the intrusiveness of advertisements: Scale
development and validation. Journal of advertising, 31(2), 37-47.
Liebl, D., Rameseder, S., & Rust, C. (2017). Functional Insights into Google AdWords. Lambrecht, A., & Tucker, C. (2013). When does retargeting work? Information specificity in
online advertising. Journal of Marketing Research, 50(5), 561-576.
Machleit, K. A., & Wilson, R. D. (1988). Emotional feelings and attitude toward the advertisement: The roles of brand familarity and repetition. Journal of Advertising, 17(3), 27-35.
Maximum CPC bid: Definition. Google. (2017) Retrieved from: https://support.google.com/adwords/answer/6326?hl=en
Milne, G. R., & Boza, M. E. (1999). Trust and concern in consumers’ perceptions of marketing information management practices. Journal of interactive Marketing, 13(1), 5-24.
Mooradian, T. A. (1996). Personality and ad-evoked feelings: The case for extraversion and neuroticism. Journal of the Academy of Marketing Science, 24(2), 99-109.
27 Obermiller, C., Spangenberg, E., & MacLachlan, D. L. (2005). Ad skepticism: The consequences of
disbelief. Journal of advertising, 34(3), 7-17.
Rozendaal, E., Buijzen, M., & Valkenburg, P. (2010). Comparing children's and adults' cognitive advertising competences in the Netherlands. Journal of Children and Media, 4(1), 77-89. Samsup, J. (2005). The effect of online media credibility on trust relationships. Journal of Website
Promotion, 1(2), 57-78.
Speck, P. S., & Elliott, M. T. (1997). The antecedents and consequences of perceived advertising clutter. Journal of Current Issues & Research in Advertising, 19(2), 39-54.
Tutaj, K., & van Reijmersdal, E. A. (2012). Effects of online advertising format and persuasion knowledge on audience reactions. Journal of Marketing Communications, 18(1), 5-18. White, T. B., Zahay, D. L., Thorbjørnsen, H., & Shavitt, S. (2008). Getting too personal: Reactance to
highly personalized email solicitations. Marketing Letters, 19(1), 39-50.
Xu, L., Duan, J. A., & Whinston, A. (2014). Path to purchase: A mutually exciting point process model for online advertising and conversion. Management Science, 60(6), 1392-1412.
Ying, L., Korneliussen, T., & Grønhaug, K. (2009). The effect of ad value, ad placement and ad execution on the perceived intrusiveness of web advertisements. International Journal of Advertising, 28(4), 623-638.
Zufryden, Fred S., James H. Pedrick, and Avu Sankaralingam (1993), "Zapping and Its Impact on Brand Purchase Behavior," Journal of Advertising Research, 33 (January/February), 58-66.
28 Appendix I: Scenario used in the experiment
29 Appendix II: Examples of stimuli used in the study
