Master Thesis
CUSTUM MADE AND TAILORED TO
YOUR NEEDS!
How social advertising hides its persuasive intent and the role of
privacy concerns and intensity of Facebook use in the effectiveness of
tailored advertisements
Author: Vincent Kersten
Student number: 5603285
Supervisor: Dr. Daan Muntinga
January 2015
Graduate School of Communication, University of Amsterdam Faculty of Social and Behavioural Science
2
Word of thanks
First of all I would like to thank my supervisor Dr. Daan Muntinga for his guidance during my research and the creation of my Master thesis. You have been an excellent supervisor with your sharp, though constructive feedback during our one-on-one sessions. You gave me useful insights and helped me tackle my different problems along the way. Besides that, you showed me how to approach my research from different angles and how to be eager into investigating even more. I’ve had a lot of fun during our meetings, the discussions we’ve had
and the things we learned from one another. I wish you all the best and of course a lot of happiness with your daughter.
Second of all I want to thank my wife Raymonda, for her support, patience and unconditional love. Especially while I was busy writing and doing my research and therefore neglecting all my shores. You have been very kind to me and made my life these last few months a lot easier!
Finally a word of thanks to my parents, James and Loes. Thanks for your support and encouragement. Dad, thanks for being my sounding board to bounce my many ideas off and mom for doing the worrying for me.
Kind regards,
3
SUMMARY
Facebook gives marketers access to a large amount of information about their 1.2 billion users. What they like, what they do, who their friends are and much more. Marketers can use this information to create tailored advertisements based on their online behavior within the Facebook platform. Facebook executives state that their users want and request these tailored advertisement (Financial Times, 2014), but this assumption has never fully been researched. The current study tries to give more insight in what the effects are of tailored advertisements on Facebook users’ brand attitude, brand recall and brand recognition. This study also investigates if tailoring affects persuasion knowledge and if in turn persuasion knowledge mediates the effect of tailoring on brand attitudes, brand recall and brand recognition. Besides these effects, the moderating roles of privacy concerns and intensity of Facebook use are also examined.
The results of this research show that tailoring has a positive effect on brand recognition and that it triggers less persuasion knowledge than non-tailored advertisements do. The reason why tailoring induces fewer persuasion knowledge is because tailored advertisements are effective in hiding their persuasive intent. No effects of tailoring for brand attitude and brand recall were found and also no mediating effects of persuasion knowledge, nor moderating effects of privacy concerns and intensity of Facebook use were discovered. An explanation for the lack of significant results for these variables might lie in the absence of self-relevant stimulus material. Conclusions and limitations are further discussed at the end of this research paper.
4
Table of Contents
Theoretical framework ... 8
Social Advertising: advertisements tailored for social media ... 8
How tailored advertisements may affect brand attitudes and brand awareness ... 9
The effects of tailored advertisements on persuasion knowledge ... 11
The mediating effect of persuasion knowledge ... 13
The moderating influence of Privacy concerns ... 14
The moderating effect of intensity of Facebook use ... 17
Method ... 19
Study Design ... 19
Procedure and participants ... 20
Development of stimulus materials ... 20
Measures ... 22
Results ... 24
Controlling for alternative explanations ... 24
Main effect of tailoring ... 25
Effect of tailoring on Persuasion Knowledge ... 25
The Mediating effect of Persuasion Knowledge ... 26
The Moderating Effect of Privacy Concerns ... 27
The moderating effect of Intensity of Facebook use on persuasion knowledge ... 30
Conclusion ... 32
General discussion ... 33
Effects of tailoring ... 34
Effects of tailoring on persuasion knowledge ... 36
Effects of Privacy Concerns ... 37
Effects of intensity of Facebook use ... 39
Limitations and future research ... 40
Managerial implications ... 41
5
Appendices ... 49
Appendix A: Tailored Facebook profile (male) ... 49
Appendix B: Non-tailored Facebook profile (male)... 50
Appendix C: Tailored Facebook profile (female) ... 51
Appendix D: Non-tailored Facebook profile (female) ... 52
Appendix E: Questionnaire ... 53
Appendix G: Background story male ... 59
6
Introduction
During the past 50 years marketers have developed different strategies to advertise their products and services to consumers. One of their newest strategies is the possibility to engage in what is now popularly known as “social advertising”, that is effectively a form tailored advertising. With social advertising, an advertiser uses information from consumers’ profiles on social networking sites (SNSs). With information from this profile, like their interests, their SNS activities, the brand and brand pages they like and more, marketers can now tailor ads for highly specific consumers (Bakshy, Eckles, Yan, & Rosenn, 2012). For example, they can tailor an advertisement based on their interests, incorporate the brands they like and use headlines which reminds them what their friends also like. With approximately 1.2 billion users, the SNS Facebook has a vast amount of personal data available for such types of social advertising. Facebook sells this information to marketers based on the premise that users prefer seeing advertisements in their timeline which are tailored to their needs, likes, shares and personal interests, over ads that are not (Financial Times, 2014).
Despite the premise Facebook makes, this does not mean that those tailored ads do not evoke some kind of reaction from Facebook users which could affect the advertised brand. Indeed, previous research has examined different effects of advertising on SNSs, such as the effects of non-tailored social advertising (Hadjia, Barnes, & Hair, 2012), the effects of likes, shares and number of followers on brand attitude and word-of-mouth (Phua & Ahn, 2014) and what influences consumers’ perceptions and attitudes towards advertising on Facebook
(Yaakop, Anuar, & Omar, 2013). But none of these studies has researched the direct effects tailored advertisements have on brand attitudes and brand awareness. Therefore in this thesis I will focus on the effects of tailored advertising on the social networking site Facebook and examine how this affects brand attitudes, brand recall and brand recognition.
7 I will also investigate if consumers who have relativity more persuasion knowledge have different brand attitudes and brand awareness then those who not, and if persuasion knowledge mediates the effect of tailoring regarding those brand attitudes and brand awareness. Persuasion knowledge is developed within consumers through experience with previous persuasion attempts, such as advertising and shapes consumers’ reaction to new persuasion attempts (Friestad & Wright, 1994). Till date, no research has focused on what the effects are of tailoring on SNSs for persuasion knowledge. While SNSs are becoming more and more integrated into our daily lives (Lovejoy, Horn & Hughes, 2009), the persuasion knowledge users develop about tailored advertisements tactics used on SNSs should increase. Therefore it is important to examine if users’ current persuasion knowledge about social advertising is sufficient to mediate the outcome of those tailored advertisements.
Besides the effects and mediating role of persuasion knowledge, I will also look at existing privacy concerns and if these privacy concerns moderates existing persuasion
knowledge for tailored advertisements. Previous research has focused on privacy concerns of tailored advertisements and the effects on attitudes towards the brand (Debatin, Lovejoy, Horn, & Hughes, 2009; Turow, King, Hoofnatle, Bleakley, & Hennessy, 2009) but never with respect to persuasion knowledge, nor brand recall and brand recognition. It is possible that privacy concerns affects peoples existing persuasion knowledge when they are exposed to tailored advertisement, which then influences their brand attitudes, recall and recognition.
Finally I will also try to find out if users who use Facebook more intensive, have different brand attitudes and brand awareness then those who use Facebook less intensive. Till date only one study focused on intensity of Facebook use and the effects it has on brand attitude without researching the possible moderating effects on persuasion knowledge (Phua & Ahn, 2014). It could be that for users who spend a lot of time on Facebook, persuasion knowledge is less of an influence on how they appreciate the use of their personal information
8 for tailored advertisements.
Besides the contribute to the scientific literature, this study is also important from a socialtal perspective. Marketers are paying Facebook a lot of money for using their data for tailored advertising. For Facebook it is therefore important to know if their tailored
advertising model is working and for marketers it is important to know if their advertising budget is not spend on the wrong media channel or harming their brands in any way.
On the following pages I will dive deeper into the theoretical background of the
different theories and variables. Based on the theory I will formulate hypothesises. After that I will explain the conducted experiment in detail in the method section. Finally I will present the results of the experiment, discuss these findings and draw conclusions regarding the hypothesises.
Theoretical framework Social Advertising: advertisements tailored for social media
Tailoring is an advertisement strategy where messages are created based on unique information about the recipient. The information used for these messages is gathered via surveys, transaction data, online behaviour, and other sources (Kreuter, Lukwago, Bucholtz, Clark, & Sanders-Thompson, 2003). When tailoring is implemented on social media it is called social advertising (Bakshy, et al., 2012). With social advertising, tailoring is based on user information from the SNS about themselves, their peers and peer affiliations (with brands, organizations or products).
The SNS used in the current study is Facebook. Facebook fits Boyd an Ellison’s
(2008) definition of an SNS, namely, a web-based services which allow users’ to (1) construct a public profile within a closed system, (2) the possibility of creating a list of users with whom they share a connection and (3) view the connections of those users and the
9 connections they made with others within the system. Because Facebook fits this definition, this study views tailoring as a form of social advertising (Bakshy et al., 2012).
How tailored advertisements may affect brand attitudes and brand awareness
The goal of social advertisements is to convince the user into taking action, such as clicking on the displayed advertisement. The advertised messages therefore has to be appealing or convincing enough to evoke the willingness to take action within the user. The way an ad affects the recipient can be explained by the Elaboration Likelihood Model of persuasion (ELM). The ELM examines how a persuasive message influences attitudinal changes and explains why some aspects of the persuasive message are influencing attitudinal change and behaviour in various ways (Petty & Cacioppo, 1986). The model has been used in a variety of research studies (Bragge, Kallio, & Sunikka, 2008), including the field of message
personalization (Tam & Ho, 2005) and can help explain the effects of social advertising. According to the ELM, consumers process information either via the central or the peripheral route. Which route will be taken depends on the motivation and the ability the recipient has to elaborate on the message. When the recipient has the motivation as well as the ability to elaborate on the message (the elaboration likelihood is high), message processing occurs via the central route. Via this route the information processing within the recipient is based upon critical thinking and the message content is given due consideration. This route of processing requires more cognitive effort than the peripheral route. When the recipient does not have the motivation or the ability to process the message content in detail, the peripheral route is taken (the elaboration likelihood is low) which requires less thoughtful processing and cognitive effort. Because of the lack of motivation or ability, the recipient focuses on simple cues of the message instead of the message content itself and the decision making is based on rules or heuristics (Bragge et al., 2008; Petty & Cacioppo, 1986; Tam & Ho, 2005).
10 When marketers can heighten user elaboration, they can influence the choices users make regarding tailored advertisements. As mentioned above, when users have to motivation to elaborate on the message content, processing occurs via the central route of the ELM. Therefore it is key that marketers are able to match users’ preferences and personality traits within their message content, to increase the motivation to elaborate on that message content itself, and not to focus on simple cues. By tailoring products or services in social
advertisements that match the preferences of consumers, consumers are more motivated to elaborate on the message content and decision making occurs through the central route of persuasion in the ELM (Tam & Ho, 2005). A way to increase motivation and elaboration is by incorporating a piece of personalization, such as the recipient’s name, which makes the
message more relevant. This in turn leads to more attention and motivation to process the message content because of its relevancy to the recipient. Therefore in this study, social advertising is expected to affect elaboration via the central route. According to the ELM, when the motivation and ability to process the message are high, the cognitive activity is high. This will lead to more central processing of the message itself and evokes longer lasting effects (Maslowska et al., 2011; Tam & Ho, 2005) of recall and recognition of the brands used in the tailored advertisements. Based on this information the following hypothesises are formulated:
H1a: Tailored advertisements leads to more brand recall then non-tailored advertisements. H1b: Tailored advertisements lead to more brand recognition then non-tailored
advertisements.
Previous studies have shown that tailored messages gain more attention and evoke more processing within the respondent (Briñol & Petty, 2006; Kreuter, Bull, Clark, & Oswald, 1999). Advertisements which are not tailored to recipients needs are likely to be avoided and are evaluated more negatively then tailored advertisements (Kelly, Kerr, &
11 Drennan, 2010; Hassan et al, 2013). Tailoring of advertisements increases evaluation of those advertisements. This evaluation creates a cognitive response within the recipient. A cognitive response is a thought generated in response to persuasive communication (Petty, Ostrom, & Brock, 1981). In this case the persuasive communication is a tailored advertisement.
According to the cognitive response theory, the thoughts an individual has about a message affects their attitudes. Positive thoughts lead to more persuasion and negative thoughts lead to less persuasion (Petty, Briñol, & Priester, 2009). So when tailoring leads to more elaboration, this increased elaboration leads to more cognitive thoughts. Depending on the evaluation of the tailored advertisement, these thoughts are either positive or negative (Petty et al., 1981). Based on the positivity bias however, which is the tendency that people have to be more biased towards positive thoughts than to more negative ones, more positively biased thoughts are produced by the recipient when they encounter personal relevant information (Burnkrant & Unnava, 1995). Thus tailoring leads to more elaboration, resulting in more cognitive thoughts which are positive due to the positivity bias, which in turn positively affects someone’s attitudes.
Because tailored advertisements heighten elaboration due to the personal relevance of tailoring and therefore could evoke more positive biased thoughts, the following hypothesis is formulated:
H1c: Tailored advertisements leads to more positive brand attitudes then non-tailored
advertisements.
The effects of tailored advertisements on persuasion knowledge
Over time consumers develop knowledge about the tactics used in persuasion attempts by marketers. They use this knowledge to identify how, when and why marketers are trying to influence them, which helps them to cope with those persuasion attempts. This knowledge about persuasion attempts is called persuasion knowledge (Friestad & Wright, 1994). People
12 gain persuasion knowledge in different ways, from first-hand experience in social interactions with friends or family, from observing marketers, from information gained from the media about advertising and marketing tactics or in school. As a consequence of this gained
knowledge, people react differently when confronted with a persuasive attempt by salespeople or advertisers, which will influence their behaviour and attitudes toward the marketer or the brand (Friestad & Wright, 1994). Some people have more persuasion knowledge then others and because of this knowledge they react different when confronted with a persuasion attempt. Those with more persuasion knowledge are more aware of the different tactics used and are more experienced in coping with them.
The way tailoring affects persuasion knowledge for brand attitude, brand recall and brand recognition can be explained by the Persuasion Knowledge Model (PKM). The model describes how the knowledge a consumer has about the goals of a marketer influence their response to the persuasion attempt (Friestad & Wright, 1994). Based on that knowledge, the consumer is able to recognize, analyse, interpret, evaluate and remember previous persuasions attempts and select a coping tactic which is believed to be effective and appropriate for the given persuasion attempt. Friestad and Wright state that there is a difference between the acceptance of information in a non-persuasive situation and when the consumer realises the persuasion attempt made by the marketer (Friestad & Wright, 1994). When the consumer is aware of the persuasive attempt, they will evaluate the given information more critical and judge for themselves if they will accept or refute the persuasive attempt.
It is important to understand the effects of persuasion knowledge for social
advertising. Marketers try to bypass the persuasive intent of their advertisements by tailoring it to the user and thereby pretending that the commercial message in the advertisement is a natural part of the offered content (De Pelsmacker & Neijens, 2012). Therefore it is possible that users are not aware of the persuasion tactics applied by the marketer and as a result their
13 persuasion knowledge is not activated by the tailored advertisement. With social
advertisements the content of those advertisements is tailored based on personal information of the SNS user. Because personal information is used in tailored advertisements which also looks like they are part of the natural context, users themselves can be unaware of the
persuasive attempts of those advertisements. When they are unware of the persuasive attempt, their existing persuasion knowledge will not be activated. Thus tailoring leads to less
activation of persuasion because it ‘hides’ it persuasive intent to the user. Based on this
assumption, the following hypotheses is formulated:
H2: Tailored advertisements activate less persuasion knowledge than non-tailored
advertisements.
The mediating effect of persuasion knowledge
Different studies have focused on the effects of persuasion knowledge on attitudes towards the advertisement in online tailored environments (Bright & Daugherty, 2012), in different advertising formats (Van Reijmersdal, Neijens, & Smit, 2009), advergames (Van Reijmersdal, Rozendaal, & Buijzen, 2012) and sponsorship (Boerman, Van Reijmersdal, & Neijens, 2012). Several findings from those studies indicate that persuasion knowledge negatively affects attitudes and positively affect cognitive responses such as brand recognition. When people have more persuasion knowledge, they are more aware about the tactics used by marketers’
and react more reserved to the persuasive attempt. When they become aware of those tactics and feel they are misled, this awareness affects their attitudes towards the brand and heightens their attention towards the advertisement. This heightened attention results in better brand recall and brand recognition (Bright & Daugherty, 2012).
Based on theory and empirical results of the effects persuasion knowledge has on brand attitudes, brand recall and brand recognition, it is expected that persuasion knowledge acts as a mediator between tailored social advertising and brand attitude, brand recall and
14 brand recognition. Tailoring affects persuasion knowledge, which in turn mediates the effect tailoring has on brand attitude, brand recall and brand recognition. The more persuasion knowledge a consumer has, the more aware he or she becomes of the tactics used by the marketer in social advertising, resulting in more negative effects on brand attitudes. Therefore the following hypothesis is formulated:
H3a: Persuasion Knowledge mediates the impact of tailored advertising. When tailoring
induces high persuasion knowledge, it in turn has a negative effect on brand attitudes.
In line with previous empirical results it is also expected that the brands used in social advertisements are remembered better when persuasion knowledge is high, because of the heightened attention these tailored social advertisements receive and the repeating of the brand name in the advertisement (Bright & Daugherty, 2012; Boerman et al., 2012). Social advertisements are tailored based on the activities of the Facebook user. So when a user has mentioned or interacted with a brand and after that interaction is being exposed to a tailored advertisement from that same brand, this multiple exposure enhances brand recall and brand recognition. Based on learning theories and associative memory theory (Anderson, 1995; Robinson, 1995), the repeated use of the brand therefore functions as an additional prime which enhances recall of the brand (Boerman et al., 2012). Based on this theory and the previous empirical findings the following hypothesis is formulated:
H3b: Persuasion Knowledge mediates the impact of tailored advertising. When tailoring
induces high persuasion knowledge, it in turn has a positive effect on brand recall and brand
recognition.
The moderating influence of Privacy concerns
SNSs, and Facebook in particular, are becoming more and more integrated in users’ daily
lives through specific rituals and routines (Lovejoy, Horn & Hughes, 2009). SNSs facilitate users’ with features such as sharing of information, online interaction and the development of
15 new relationships. These features such as sharing of information and interacting with others have raised new privacy concerns among users’ (Dhami, Agarwal, Singh, & Minj, 2012). Two reasons why these privacy concerns have arisen are (1) because SNSs represent a social sphere, which stores large amounts of personal information (Govani & Pashley, 2005) and (2), a lot of the information provided on SNSs can easily be used by third parties or it can be copied, replicated, forwarded, or even taken out of context by others (Boyd, 2006). This side of using SNSs can affect the way users’ interpreted tailored social advertisements, which use
personal information based on their activities within Facebook. It is possible that those who have high concerns of privacy are more affected by the use of personal information for tailored advertisements then those who are not. In this paper we define privacy concerns based on the work of Xu, Dinev, Smith and Hart (2008) as “concerns about possible loss of privacy as a result of information disclosure”.
A lot of research has focused on the privacy concerns of users on SNSs and Facebook in particular (Debatin et al., 2009; Fogel & Nehmad, 2010; Hoy & Milne, 2010; McDonald & Cranor, 2010; O’Brien & Torres, 2012). Predominantly the body of work shows a more negative attitude towards privacy concerns for using SNSs (Fogel & Nehmad, 2008; Hoy & Milne, 2010) because people see tailored advertising based on their online behaviour on SNSs as an invasion of their privacy (McDonald & Cranor, 2010; Smit, Van Noort, & Voorveld, 2014).
Despite the negative attitude towards privacy concerns however, research shows that there is a privacy trade-off between users’ and SNSs. Users are willingly giving up their privacy and let go of their privacy concerns for experiencing the benefits of using SNSs for free. Some of their activities are driven by the desire of social acceptance and not by privacy concerns (Lovejoy et al., 2009; O’Brien & Torres, 2012).
16 for free, their attitude towards the use of personal information for tailored advertising is predominantly negative. This negative attitude can affect the evaluation of tailored social advertising. Even though the users’ accept the trade-off, this does not mean they have a
positive attitude towards tailored advertisements.
It is likely that those users’ who have high privacy concerns, see tailored advertising
as an invasion of privacy. Their privacy concerns therefore moderate their evaluation of the tailored advertisement and the advertised brand. Also, because of their privacy concerns, when they see the tailored advertisement based on their personal information, they pay more attention to the tailored advertisement due to the use of personal information. When they pay more attention towards the tailored advertisement because of their privacy concerns, they will have a more positive brand recall and recognize the brand better. Therefore the following hypothesises are formulated:
H4a: The effect of tailored advertising on brand attitude is moderated by privacy concerns.
When privacy concerns are high, tailoring has a more negative impact on brand attitude then
non-tailored advertisements.
H4b: The effect of tailored advertising on brand recall and brand recognition is moderated
by privacy concerns. When privacy concerns are high, tailoring results in more positive brand
recall and brand recognition then non-tailored advertisements.
Besides the moderated role privacy concerns has for tailored advertisements on brand attitude, brand recall and brand recognition, it is also expected that privacy concerns
moderates the relationship between tailoring and persuasion knowledge. When marketers make use of the personal information found on SNSs to tailor their advertisements to fit the user, the usage of that information exceeds the intentions the user originally had when joining a SNS. Namely, the development and supporting of social connections and not the use of their personal information for tailored advertisements (Hoy & Milne, 2010). Users who are
17 concerned about their privacy could realize that marketers are using their personal
information for tailoring advertisements to their needs. This realization can trigger their existing persuasion knowledge. As mentioned before, there is a difference between the acceptance of non-persuasive information and when the user realizes the persuasive intent from the marketer (De Pelsmacker & Neijens, 2012). When consumers become aware of the persuasive intent they will cope with this persuasion attempt (Friestad & Wright, 1994). Privacy concerns thus work as a trigger for their existing persuasion knowledge and by that it moderates the effect of persuasion knowledge. Based on this assumption, the following hypothesis is formulated:
H4c: The effect of tailored advertising on persuasion is moderated by privacy concerns. When
privacy concerns are high, tailoring induces more persuasion knowledge then non-tailored
advertisements.
The moderating effect of intensity of Facebook use
Despite the vast amount of studies in the field of SNSs and social advertising, almost no studies have researched the effects of intensive Facebook use for brand attitude, brand recall and brand recognition regarding social advertising (Ellison, Steinfield, & Lampe, 2007; Ross et al., 2009; Valenzuela, Park, & Kee, 2009). Only one study conducted by Phua and Ahn (2014) has focused on the intensity of Facebook use and the effects it has on brand attitude. Their results showed that the more a person uses Facebook, has it incorporated in their daily routine and feels out of touch when he or she has not logged in for a while, the more likely they have a positive attitude towards the brands advertised on Facebook brand pages. Users who use Facebook more intensively are in general more experienced users. Their experience makes that they notice, interpreted and evaluate content different than less experienced users (Phua & Ahn, 2014).
18 advertisements. Research has shown that more than half of the users’ of SNSs almost never notice the social advertisements because they focus on other content (Hadjia et al., 2012). When intensive Facebook users do not notice tailored social advertisements, their existing persuasion knowledge cannot be triggered by the persuasive intent of those advertisement. Thus, the intensive use of Facebook moderates the relationship between persuasion
knowledge and brand attitude, brand recall and brand recognition. More persuasion
knowledge should lead to more negative brand attitudes and more positive brand recall and brand recognition, except for intensive Facebook users. Their intensive use of Facebook triggers less persuasion knowledge, because they do not notice the tailored advertisements due to their intensive use of Facebook. Therefore their persuasion knowledge will not affect their attitudes towards the advertised brand. Also, because those intensive users do not notice the social advertisements, they will not be able to recall or recognize the brand. Based on this assumption, the final hypothesises are formulated:
H5a: The effect of persuasion knowledge on brand attitude is moderated by the intensity of
Facebook use. When the intensity of Facebook use is high, the effect of persuasion knowledge
on brand attitude becomes less strong, resulting in more positive brand attitudes.
H5b: The effect of persuasion knowledge on brand recall and brand recognition is moderated
by the intensity of Facebook use. When the intensity of Facebook use is high, the effect of
persuasion knowledge on brand recall and brand recognition become less strong, resulting in
19
Figure 1: Conceptual model of this study
Method
Study Design
A 2 (tailoring: tailored vs. non-tailored) between-subjects experimental design was chosen to test the hypotheses because this allowed me to manipulate the independent variable in two conditions: (1) a condition of a Facebook profile which has a tailored social advertisement displayed and (2) a condition of a Facebook profile which had a non-tailored social advertisement displayed. The experimental element was incorporated in a survey and distributed among participants.
The survey consisted of three parts. In the first part of the survey participants were asked about certain demographics (e.g. age, education) and their intensity of Facebook usage. After answering these questions participants were directed towards the second part of the survey, which contained the manipulation itself. They received instructions about what they were expected to do. First they would read a background story about a Facebook user and afterwards they would be directed to the Facebook profile of that user. Participants were asked to try to put themselves in the position of the Facebook user and look at and think of the Facebook profile as if they were that Facebook user themselves. After reading the background
20 story, participants were randomly assigned to one of the two conditions, a Facebook profile with the tailored social advertisement or a Facebook profile with a non-tailored social advertisement (see Appendices A-D). After studying the Facebook user profile, participants were directed to the final part of the research. This part started with four filler questions and after those they had to answer the questions about privacy concerns, brand attitude, persuasion knowledge, brand recall and brand recognition (see Appendix E).
Procedure and participants
Respondents were recruited via a social media (Facebook, LinkedIn and Twitter) and e-mail through a non-probability sampling technique with the request to participate in this study. This created a diverse group of respondents. The message contained a link that directed respondents to a webpage where they could start the experiment. To hide the reason behind the experiment, a cover-up story was used. Respondents were told that they would participate in a research that focused on marketing and privacy on SNSs and that their input was
necessary for future research and recommendations for the University of Amsterdam. A total of 172 respondents started the survey. From the initial 172 respondents, 56 (32.56%) failed to complete the questionnaire. After deleting the cases with missing data, 116 participants (66.4%) remained that completed the questionnaire. These participants ranged in age between 16 and 69 years old (M = 29.38, SD = 1.09). From those participants 61 where female (52.6%) and more than half of the participant finished a form of higher education (78.5%) and held a fulltime job (57.8%).
Development of stimulus materials
Two types of Facebook user profiles were created. One condition which incorporated the tailored advertisement and one condition with a generic, non-tailored advertisement. The layout of the user profiles were created to form exact replicas of those on Facebook. This was done in order to create a setting which was as close as possible to real life. The Facebook
21 profiles were adjusted by using a fictional user name, fictional friends and messages. Because participants were asked to place themselves in the Facebook user whose profile they were about to see, a version for males and a version for females was created. By creating a profile for both genders, it was expected that this would help the process of transitioning themselves in the position of the Facebook user. The advertisements used for males or females, as for the tailored and the control condition, were exactly the same. The only differences were that in the female condition the Facebook user was a woman named Sandra Bakker and in the male condition the Facebook user was a man named Sander Bakker. Also the content on the Facebook profile was the same for males and females, only some words were adjusted which would appeal more to men or woman For instance in the male condition the word ‘sneakers’ and in the female condition the word ‘pumps’ was used to indicate shoes. Besides the
adjustments on the Facebook profile, the background stories were also adjusted for men and woman. The stories themselves were the same, only some words were interchanged so they would fit a female or male user more. For example, in the background story about the male the word ‘sneakers’ and for the female the word ‘pumps’ is used to indicate his or her interest
in online shoe shopping.
In the background story participants read about the current social life of the Facebook user and their online behaviour. After reading the background story participants were
randomly assigned to either the experimental or the control condition.
For the tailored advertisement the brand Zalando was used, which is an online department store for clothing. Zalando is a well-known brand that regularly advertises on Facebook and other media platforms. This brand was chosen in order to create a setting that is as close to real life as possible. The advertisement was tailored to the activity of the user on Facebook. The brand in the advertisement (Zalando) was mentioned in one of the posts on the Facebook profile. Also the first name of Facebook user (either Sander or Sandra) was used in
22 the advertisement message which was placed above the advertisement. Depending on the male or female profile, the sentence of that message was: ‘Pssst… Sander [Sandra] we have new sneakers [pumps] available for you!’. In the control condition the advertisement was not
based on information of the Facebook profile and also the name of the user was not
incorporated in the advertisement. Appendices [F] and [G] display the background story used in this study.
Measures
Dependent variables. In this study we define brand attitude as ”the overall internal
evaluation of an object such as a branded product” (Mitchel & Olson, 1981). Brand
recognition and brand recall are two measures of brand awareness. Brand recognition is the consumers’ ability to confirm prior exposure to the brand, whereas brand recall is the unaided
retrieval of the brand from memory by the consumer (Hoeffler & Keller, 2002).
Brand attitude was measured trough six items used by Phua and Ahn (2014). The six items asked respondents to indicate on a seven-point semantic differential (ranging from 1 to 7) how they felt about the brands in the advertisements on the Facebook profile. The
questions were among others: ‘unappealing/appealing’ and ‘unpleasant/pleasant’. Factor
analysis (Principal Component Analysis) showed that the six items loaded on one factor and together formed a sufficiently reliable scale for brand attitude (EV = 3.34; R2 = .56;
Cronbach’s α = 0.84; M = 3.95; SD = 0.92).
Brand recall was measured by asking participants which brand they remembered from the Facebook profile. When they did not remember any brands, did not remember them correct, or when they remembered any brands but not the brand from the tailored
advertisement is was coded as 0. When they remembered the tailored brand or a combination of the tailored brand with any other brand it was coded as 1.
23 which they had to indicate if they had seen the brand on the Facebook profile. Five brand logos were present on the Facebook profile (Zalando, HEMA, Albert Heijn and Adidas) and four were not (Apple, Samsung, Essent and ANWB). The five present brand logos were chosen to further enhance a real life setting. These brands are all well-known Dutch brands with a lot of followers on Facebook. The recognition measure was corrected for false
responses by recoding the scores. When they did not remember any brand (correct) or if they remembered a brand correct, but not the brand in the tailored advertisement is was coded as 0 and when they remembered the tailored brand or a combination of the tailored brand with any other brand is was coded as 1.
Persuasion knowledge.The proposed mediator persuasion knowledge was measured with six items taken from Tutaj and Van Reijmersdal (2012), on a scale ranging from 1 (strongly agree) to 7 (strongly disagree). Examples of these items were: ‘The aim of this advertisement is to sell products or services’ for the selling intent and ‘The aim of this
advertisement is to influence your opinion’ for the persuasive intent. Finally two items were used as filler questions referring to the informational intent of the advertising format. Factor analysis (Principal Components Analysis) showed that the items loaded on one factor and together formed a reliable scale for persuasion knowledge (EV = 2.98; R2= .50; Cronbach’s α = .80; M = 4.89; SD = 1.16). Persuasion knowledge was then transformed into a dichotomous variable trough a median split, consisting of groups with low persuasion knowledge (0) and high persuasion knowledge (1).
Privacy concerns and intensity of Facebook use. The proposed moderator privacy concerns was measured with seven items based on the work of Krasnova, Kolesnikova and Günther (2010). Respondents were asked to indicate on a seven-point scale how concerned they were (1 not concerned at all, 7 very much concerned) with the sharing of personal information on Facebook. The questions were asked based on the statement: ‘How concerned
24 are you that the information you leave on Facebook…’ and participants had to answer
questions such as: ‘… can be used in a way I did not foresee’ or ‘… can be misinterpreted’. A factor analysis (Principal Components Analysis) showed that the seven items loaded on one factor and together formed a reliable scale for privacy concerns (EV = 4.79; R2 = .68;
Cronbach’s α = .92; M = 4.11; SD = 1.41). Privacy concerns was then transformed trough an
median split into a dichotomous variable consisting of groups with low privacy concerns (0) and high privacy concerns (1).
Intensity of Facebook use was measured by using the Intensity of Facebook Scale (Ellison et al., 2008). Five items were used to assess how respondents were emotionally connected to Facebook and how the integrate Facebook into their daily activities. Participants were asked to indicate on a seven-point scale ranging from 1 (strongly disagree) to 7 (strongly agree) what their opinion was for questions such as: ‘I am proud to tell others that I use Facebook’ and ‘Facebook is part of my daily activity’. Factor analysis (Principal Components
Analysis) showed that the five items loaded on one factor and together formed a reliable scale for intensity of Facebook use (EV = 2.52; R2 = .50; Cronbach’s α = .74; M = 3.89; SD = 1.23). Intensity of Facebook use was then transformed into a dichotomous variable consisting of groups with low intensity of Facebook use (0) and high intensity of Facebook use (1).
Results Controlling for alternative explanations
To check for any alternative explanations before conducting the necessary analysis’s a
bivariate correlation analysis was conducted. The results showed I had to control for age when conducting analyses for persuasion knowledge, because that variable significant correlated with the variable persuasion knowledge (r = .21; p = .023). I also had to control for gender when conducting analyses for intensity of Facebook use, because that variable significant correlated with the variable intensity of Facebook use (r = .23; p = .015).
25
Main effect of tailoring
The first hypothesis stated that tailored advertising would lead to higher brand recall (H1a), higher brand recognition (H1b) and more positive brand attitudes (H1c), than non-tailored advertising. To test the main effect of tailoring on brand recall and brand recognition a logistic regression analysis was conducted. This analysis was chosen because the dependent variables brand recognition and brand recall were dichotomous. The results showed there was no main effect of tailoring on brand recall, Wald (χ2 (1) = .12, p = .725. Participants in the tailored condition did recall the advertisements slightly better (M = .58; SD = .06) than the non-tailored condition (M = .54; SD = .07), but the difference was not significant. Therefore hypothesis one cannot be supported for brand recall.
For brand recognition, results showed a significant effect of tailoring, Wald χ2 (1) = 4.67, p = .031. The participants in the tailored condition recognized the advertisement significantly better (M = .34; SD = .06) than the non-tailored condition (M = .16; SD = .05). Therefore hypothesis one is supported for brand recognition.
To test the effect of tailoring for brand attitude, a one-way ANOVA was conducted. The results showed that tailoring a message did not significantly increase brand attitude. Respondents who were exposed to the tailored advertisement had a slightly higher brand attitude (M = 3.99; SD = .90) than those who did not see the tailored advertisement (M = 3.90;
SD = .94), but the difference was not significant, F (1, 114) = .30, p = .896. Therefore
hypothesis one is not supported for brand attitude.
Effect of tailoring on Persuasion Knowledge
The second hypothesis stated that tailored advertising would activate less persuasion knowledge than non-tailored advertising. I controlled for the influence of age. To test the effect I conducted a logistic regression analysis. I chose this analysis because the dependent variable (persuasion knowledge) was binary divided. The results showed that there is a
26 significant effect of tailoring on persuasion knowledge, Wald χ2 (1) = 4.94, p = .026. Tailored advertising activated less persuasion knowledge (M = .48; SD = .07) than non-tailored
advertising (M = .62; SD = .07). Therefore, the second hypothesis is supported. Tailored advertisements are created in such a way that the persuasive intent is not visible to the consumer and that therefore less persuasion knowledge is activated.
The Mediating effect of Persuasion Knowledge
The third hypothesis stated that the mediated effect of persuasion knowledge would have a negative effect on brand attitude (H3a) and a positive effect on brand recall and brand
recognition and (H3b). To test these hypotheses, the four steps presented by Baron and Kenny (1986) for testing mediation were followed. For the first step a regression analysis had to be performed with the independent variable predicting the dependent variable, to see if the independent variable correlates with the dependent variable. The regression analyses showed that tailoring did not significantly influence brand attitude (β = .05; p = .584) and brand recall (β = .03; p = .728) but did significantly, although negatively, affect brand recognition (β = -.21; p = .028). The second step in the Baron and Kenny (1986) method is running a regression analysis with the mediator (persuasion knowledge) as dependent variable and tailoring as the independent variable. The analysis showed that tailoring does not significantly influence persuasion knowledge (β = .16; p = .091). For the third step in the Baron and Kenny method a regression analysis with the independent variable (tailoring) and the mediator (persuasion knowledge) predicting the dependent variables had to be run. The analyses showed that tailoring and persuasion knowledge were no significant predictors of brand attitude (β = .63; p = .798), brand recall (β = .07; p = .764) and brand recognition (β = .21; p = .090). Thus, according to Baron and Kenny (1986), persuasion knowledge cannot mediate the relationship between tailoring and brand attitude, brand recall and brand recognition. A Sobel test was run to confirm these results. The results from the Sobel test confirmed that persuasion knowledge
27 does not significantly mediate the effects of tailoring on brand attitude (z = .06; p = .090), brand recall (z = .06; p = .949) and brand recognition (z = .50; p = .615). Therefore, the third hypothesis is not supported.
The Moderating Effect of Privacy Concerns
The fourth hypothesis stated that privacy concerns would moderate the effect of tailoring, resulting in a more negative brand attitude (H4a) but better brand recall and brand recognition (H4b). To test this an MANOVA was run because there is one binary independent variable and there are several dependent variables (brand attitude, brand recall and brand recognition). The results show no significant moderating effect of privacy concerns on the relationship between tailoring and brand attitude, F (1, 114) = 1.32, p = .209, brand recall F (1, 114) = 1.82, p = .180 and brand recognition F (1, 114) = 1.37, p = .245. Therefore the fourth hypothesis is not supported (see table 1).
Table 1: Moderating effects of privacy concerns
Dependent variable Tailored condition Privacy concerns
Mean Standard deviation F p Brand attitude Yes High 4.13 .16 1.32 .209 Low 3.76 .19 No High 3.94 .16 Low 3.82 .19
Brand recall Yes High .60 .09 1.82 .180
Low .54 .10 No High .46 .09 Low .65 .10 Brand recognition Yes High .66 .07 1.37 .245 Low .67 .09 No High .76 .07 Low .96 .09
Because no significant effect was found for privacy concerns between the tailored and non-tailored advertisement condition, I checked if there were significant effects of privacy concerns within the group for the tailored condition and within the group for the non-tailored condition.
28 To test if these effects within the tailored group were significant, only the cases for the tailored advertisement condition were selected and an MANOVA was run. The results show no significant effects of privacy concerns on tailored advertisements for brand attitude (F (1, 57) = 2.18, p = .146), brand recall (F (1, 57) = .19, p = .663) and brand recognition (F (1, 57) = .01, p = .941).
To test if these effects within the non-tailored group were significant, only the cases for the non-tailored advertisement were selected and an MANOVA was run. For brand recognition, there was a significant effect in the non-tailored group between those with low privacy concerns and high privacy concerns. Participants with low privacy concerns (M = .96;
SD = .08) recognize non-tailored advertisements significantly better than those with high
privacy concerns (M = .76; SD = .06), F (1, 54) = 4.13, p = .047 (see figure 1). For brand attitude (F (1, 55) = 2.86, p = .227) and brand recall (F (1, 55) = 2.54, p = .116) no significant effects were found. It seems that the amount of privacy concerns affects brand recognition for non-tailored advertisements. Low privacy concerns lead to more recognition of brands then high privacy concerns.
Figure 2: Moderation effect of privacy concerns on brand recognition for non-tailored advertisements
The fourth hypothesis (H4c) stated that the effects of tailoring on persuasion knowledge will be moderated by privacy concerns. When the privacy concerns are high
29 within respondents, this would induce more persuasion knowledge. To test this hypothesis an ANOVA was conducted with tailoring and privacy concerns as the independent variables and persuasion knowledge as the dependent variable. The results showed no significant
moderation effects of privacy concerns on persuasion knowledge for tailored advertisements,
F (1, 111) = .50, p = .482.
Because no significant moderating effect was found for privacy concerns on
persuasion knowledge between the tailored and non-tailored condition, I checked if there were significant moderating effects of privacy concerns on persuasion knowledge within the group for the tailored and within the group for non-tailored advertisements.
To test if these effects within the tailored group was significant, only the cases for the tailored advertisement condition were selected and an ANOVA was run. The results show no significant moderating effect of privacy concerns on persuasion knowledge within the tailored condition (F (1, 57) = 3.78, p = .079).
To test if these effects within the non-tailored group was significant, only the cases for the non-tailored advertisement condition was selected and an ANOVA was run. The results show a significant moderation effect for privacy concerns on persuasion knowledge for non-tailored advertisements F (1, 54) = 5.12, p = .006 (see figure 2). Participants with high privacy concerns significantly induced more persuasion knowledge (M = 1.52; SD = .08) then those with low privacy concerns (M = 1.17; SD = .09). It seems that non-tailored
advertisements for people with high privacy concerns induce more persuasion knowledge then for people with low privacy concerns.
30
Figure 3: Moderation effect of privacy concerns on persuasion knowledge for non-tailored advertisements
The moderating effect of Intensity of Facebook use on persuasion knowledge
The fifth and final hypothesis stated that intensity of Facebook use moderates the effect of persuasion knowledge on brand attitude, brand recall and brand recognition. Higher intensity of Facebook use will lower the amount of persuasion knowledge induced, which leads to more positive brand attitudes (H5a) and less brand recall and brand recognition (H5b). To test the moderating effect of intensity of Facebook use on persuasion knowledge for brand
attitude, brand recall and brand recognition an MANCOVA was conducted in which I controlled for gender. Persuasion knowledge and intensity of Facebook use were used as independent variables and brand attitude, brand recall and brand recognition as dependent variables. The results show there were no significant moderating effects of intensity of Facebook use on persuasion knowledge for brand attitude (F (1, 110) = .44, p = .251), brand recall (F = (1, 110) = .09, p = .768) and brand recognition (F (1, 110) = .56, p = .456). Therefore the fifth hypothesis not supported (see table 3).
31
Table 3: Moderating effects of Intensity of Facebook use on persuasion knowledge
Dependent variable Persuasion knowledge Intensity of Facebook use Mean Standard deviation F p
Brand attitude High High 3.93 .19 .440 .251
Low 4.02 .17
Low High 3.76 .16
Low 4.08 .18
Brand recall High High .56 .11 .087 .768
Low .62 .09 Low High .47 .09 Low .59 .10 Brand recognition High High .69 .10 .559 .456 Low .75 .08 Low High .69 .08 Low .87 .08
Because no significant moderating effect was found for intensity of Facebook use on persuasion knowledge between tailored and non-tailored advertisements, I checked if there were significant moderating effects of intensity of Facebook use on persuasion knowledge within the group for tailored advertisements and within the group for non-tailored
advertisements.
To test if these effects within the tailored group was significant, only the cases for the tailored advertisement were selected and an MANCOVA was run, in which I controlled for gender. The results show no significant moderating effect of intensity of Facebook use for persuasion knowledge on brand attitude (F (1, 54) = .65, p = .423), brand recall (F (1, 54) = .54, p = .462) and brand recognition (F (1, 54) = .21, p = .642).
To test if these results within the non-tailored group was significant, only the cases for the non-tailored advertisement were selected and a MANCOVA was run, in which I
controlled for gender. The results show there was no significant moderating effect of intensity of Facebook for persuasion knowledge on brand attitude (F (1, 51) = 1.43, p = .237), brand recall (F = (1, 51) = .93, p = .340) and brand recognition (F (1, 51) = .12, p = .731).
32
Table 4: Summary of results for the hypotheses
Hypothesis Independent Dependent Moderator Mediator Supported
H1a Tailoring Brand recall No
H1b Tailoring Brand recognition Yes
H1c Tailoring Brand attitude No
H2 Tailoring Persuasion
knowledge
Yes H3 Tailoring Brand attitude
Brand recall Brand recognition
Persuasion knowledge
No
H4a Tailoring Brand attitude Privacy concerns
No H4b Tailoring Brand recall
Brand recognition Privacy concerns No H4c Tailoring Persuasion knowledge Privacy concerns No H5a Persuasion knowledge
Brand attitude Intensity of Facebook use No H5b Persuasion knowledge Brand recall Brand recognition Intensity of Facebook use No Conclusion
This study investigated the effects of tailored social advertisements on Facebook regarding brand attitudes, brand recall and brand recognition. Two important conclusions can be drawn as result of this study. First, tailored advertisements reduces persuasion knowledge within users of Facebook. Second, tailored advertisements have a positive effect on brand
recognition. These findings have important theoretical and societal implications.
These results confirm a part of the expectations that were drawn up at the beginning of this study. The present study also demonstrates that tailored advertisements on Facebook are effective in their goal of hiding their persuasive intent towards the consumer. The primary goal of those type of ads is to increase purchase behaviour and brand awareness. By hiding their persuasive intent they reduce the possible reserved or negative reactions from the recipient as a result of activating persuasion knowledge. This is especially important for
33 marketers who use Facebook as a medium for promoting their brands and products. By
applying tailoring strategies they can effectively prevent the triggering of persuasion knowledge in their target audience. The findings also showed that tailored advertisement increases brand recognition compared to non-tailored advertisements. This means for
marketers that although there are no direct positive effects for brand attitude and brand recall, tailoring leads to more brand recognition, which can positively affect sales at point of
purchase (Park & Stoel, 2005).
General discussion
The primary aim of this study was to examine how tailored social advertising affects brand attitude, brand recall and brand recognition, and investigate the role of persuasion knowledge, privacy concerns and intensity of Facebook in this relationship. No research so far has
focused on the effect of these different variables in respect to social advertising, therefore this study is important for future researchers and online marketers. Results of this study indicate that tailoring affects brand recognition and persuasion knowledge. However, persuasion knowledge does not mediate the effect of tailoring on brand attitude, brand recall and brand recognition. Finally, privacy concerns and intensive use of Facebook did not moderate the effects of tailoring or persuasion knowledge. In times were privacy has become the vocal point in a lot of media coverage this is also an important fact for marketers. Although privacy concerns and the intrusion of privacy by third parties gets a lot of negative media attention (New York Times, 2013), this will not affect any prior persuasion knowledge within the users of SNSs and thereby reflect on their brand or negatively affect the brand directly.
In this last part of this study the results will be discussed in further detail and conclusions will be drawn. Some of the implications which were encountered will be discussed and suggestions for future research will be proposed
34
Effects of tailoring
Tailored advertisements on Facebook did not affect brand attitude. This finding is contrary to what was predicted and contrary to the findings in previous research (Burnkrant & Unnava, 1995; Petty et al., 2009). Previous studies have shown that tailored messages gain more attention and evoke more processing within the respondent (Briñol & Petty, 2006; Kreuter et al., 1999) and that the tailoring of those advertisements plays an important role in the overall evaluation of those advertisements. When advertisements are not tailored to the recipient’s
needs, they are more likely to be avoided and evaluated more negatively then advertisements which are relevant to the respondent (Kelly et al., 2010; Hassan et al, 2013).
Unlike the current study, these previous studies were focused on tailored advertising outside Social Networking Sites (SNSs). It could be that SNSs themselves play an important role in the processing of tailored advertising. Facebook as a SNS offers users a lot of options while they are using the SNS. They can browse their timeline, chat with friends, place comments, view the activity of others and more. There is a chance that these possibilities distract the users and take away their motivation to process the information from those social advertisements. According to the Elaboration Likelihood Model of persuasion (ELM),
without the motivation to elaborate on the message content, processing of the message occurs via the peripheral route of the ELM. This route requires little cognitive effort and thoughtful processing (Petty & Cacioppo, 1986). Conform the cognitive response theory, the thoughts an individual has about the message affects their attitude (Petty et al., 2009). When message processing takes place via the peripheral route, fewer thoughts are created about the tailored advertisement which affects someone’s attitude towards the brand.
Another reason for the lack of motivation to elaborate on the message is because in this study, participants were asked to place themselves in the position of a fictional Facebook user. Because the advertisements were tailored to this fictional Facebook character and not to
35 the respondent, it could be that the displayed advertisements were never really self-relevant to the respondents. When the advertisements are not self-relevant to the respondents, there is a chance they do not have the motivation to elaborate on the message content. As mentioned above, without the motivation to elaborate on the message, fewer thoughts about the advertisement are created which could affect someone’s attitude (Petty et al., 2009).
Tailored advertisements on Facebook did positively increase brand recognition. Respondents who were exposed to a tailored advertisement recognized the brand significantly better than those who were exposed to a non-tailored advertisement. This is in accordance with was predicted and in line with previous studies (Maslowska et al., 2011). Despite this positive effect of tailoring on brand recognition, no effects of tailoring were found for brand recall. This is contrary to what was predicted and contrary to the findings in previous research (Maslowska et al., 2011).
A reason for the positive results for brand recognition, but the absence of positive results for brand recall could be because the respondents did not notice the advertisement consciously or simply because they ignored them. Previous research has shown that one of the goals of SNS users is to avoid or ignore social advertising (Hadjia et al., 2012; McDonald & Cranor, 2010), especially when the advertising is not relevant to the user (Kelly et al., 2010). As explained above, there is a chance that the displayed advertisements in this study were never really self-relevant to the respondents and therefore respondents did not have the motivation to elaborate on the message. Processing of the content thus occurred less consciously via the peripheral route and the respondents were focused on cues, such as the brand logo (Petty & Cacioppo, 1986a). Therefore the respondents noticed the brand and recognized them when they were presented in the questionnaire, but could not recall them from memory when asked.
36
Effects of tailoring on persuasion knowledge
The results of this study show that tailored advertisements induce less persuasion knowledge within the respondents than non-tailored advertisements. This finding is in line with our theoretical assumption. As suggested, tailored advertisements hide the persuasive intent and therefore activate less persuasion knowledge than non-tailored advertisements. This finding is in line with the results of previous studies which focused on tailored advertisements outside SNSs (Van Reijmersdal et al., 2009; Wouters & de Pelsmacker 2011), but these effects are not found for tailored advertisements within SNSs. Tailored advertising and other marketing tactics that try to hide the persuasive intent have been successful in doing this by developing a less intrusive format, which triggers less persuasion knowledge within the recipient
(Cauberghe & De Pelsmacker 2010). Tailored advertisements do not only hide the persuasive intent of the message but, especially in low thinking conditions, are also easier to accept by the recipient because the message ‘just feels right’ due to the tailored aspects of the message (Cesario, Grant & Higgins, 2004). This results in less suspiciousness and thereby less activation of existing persuasive knowledge.
Besides the effect of tailoring on persuasion knowledge itself, this study also examined if there was a mediating effect of persuasion knowledge on brand attitude, brand recall and brand recognition for tailored advertisements. The results indicate that there was no mediating effect of persuasion knowledge. Theory and previous empirical findings suggested that when tailoring induces more persuasion knowledge, it in return has a positive effect on brand recall and brand recognition and a negative effect on brand attitude (Bright &
Daugherty, 2012; Boerman et al., 2012). The more persuasion knowledge someone has, the more they become are aware about the tactics marketers use and respond with an appropriate coping behaviour (Friestad & Wright, 1994). So when somebody is aware of the tactics used in tailored advertisement, it could be that they feel they are misled and tricked by the
37 advertiser. This awareness affects their brand attitudes and also heightens their attention towards the advertisement, which results in a more negative brand attitude and better brand recall and brand recognition (Bright & Daugherty, 2012).
A possible reason for why this mediated effect was not found, is because the tailored advertisements were effective in hiding their persuasive intent. As the results of this study indicated, tailoring induced less persuasion knowledge then non-tailored advertisements. Because tailoring did not trigger persuasion knowledge, persuasion knowledge itself could not act as a mediator and affect the brand attitudes, brand recall or brand recognition of the
respondents. This can also be seen in the results, as the difference between respondents with either high or low persuasion knowledge was minimal for brand attitude. Although some respondents claimed that they have a lot of persuasion knowledge at one’s disposal, this knowledge was not activated through the tailored advertisement and therefore persuasion knowledge could not act as a mediator.
Another reason could also lie in the set-up of this study. As pointed out in the results for tailoring on brand attitude, brand recall and brand recognition, there is a possibility that the advertisements were not self-relevant enough for the respondents in this study. If the fictional Facebook user profile and the tailored advertisements were not relevant enough to the respondents, it could be that effects which are normally evoked by tailored advertisements did not occur. Which in turn means that no existing persuasion knowledge was evoked to act as a mediator. Future research should focus on these non-relevant aspects of this study and try to incorporate more relevant stimulus material for the respondents.
Effects of Privacy Concerns
The results of this study show that there were no moderating effects of privacy concerns for tailored advertisements on brand attitude, brand recall and brand recognition between tailored and non-tailored advertisements. These findings were contrary to my predictions based on
38 previous research. Earlier studies have shown that respondents with high privacy concerns in a tailored condition have a more negative brand attitude then those who have low privacy concerns (Fogel & Nehad, 2010; Hoy & Milne, 2010; McDonald & Cranor, 2010). In this study these effects were not found. This could be because of the existence of a privacy trade-off between users of SNSs and their privacy concerns for using the SNS (Lovejoy et al., 2009; O’Brien & Torres, 2012). Users of Facebook have stated before that they accept tailored
advertisements on Facebook compared to other sources (Ur et al., 2012). It could therefore be, that because Facebook is such an integrated part in many lives, that any existing privacy concerns are unsufficient to overrule the benefits of using Facebook. Facebook has become a habit for many users and maybe because of this, ignoring any privacy concerns they have, has also becomes a part of their daily routine.
Another explanation could again lie in the set-up of this study. It could be that the respondents do have privacy concerns, but those concerns did not apply to this research and the Facebook profile they have seen, simply because it is not their own personal Facebook profile. Any intrusion of privacy in this case does not affect them personally, but a fictional Facebook user with whom they do not have any ties. This could also explain why privacy concerns did not heighten attention towards the advertisement and increase brand recall and brand recognition for those respondents with more privacy concerns.
There were also no moderating effects found of privacy concerns for tailored
advertisements on persuasion knowledge. Contrary to the results, it was expected that when respondents have more privacy concerns, these concerns would induce more persuasion knowledge within the respondent for tailored advertisements. Although having more privacy concerns did lead to more inducing of persuasion knowledge then having less privacy
concerns, these results were not influenced by tailoring. This could mean that tailoring does not affect privacy concerns at all or that, as explained above, the stimulus material in this
