When are Instagram ads effective?
The influence of personalized ads and the use of influencers on
brand attitudes and brand engagement.
Master’s Thesis Marketing Management
Author: José Maria Calheiros Ponces Magalhães
Adress: Praça Ilha do Faial Nº1 4ESQ, 1000-168 Lisboa, Portugal E-mail: j.m.calheiros.ponces.magalhaes@student.rug.nl
Phone number: +351914759146 Student number: S4124677
University of Groningen
Department: Faculty of Economics and Business Master: Marketing Management
2
Acknowledgement
About a year ago I was setting up for my arrival in Groningen, not knowing the craziness that 2020 would bring to the world, and how it would come to impact life in general. It was an unexpected shift, that forced us to adapt to a whole new situation, from online lectures and exams and working mostly from home. Nevertheless, it is a year I will take with me and for sure won’t forget. Today more than ever I am absolutely certain that the option of doing my Master’s abroad and most importantly in the University of Groningen was the perfect choice. I am grateful for the opportunity to have taken the Marketing Management programme, and I walk towards the end of this journey highly excited of what’s to come next. I would also like to express my gratitude and appreciation for Dr. Judith de Groot for her support and guidance throughout this process, her feedback was essential in the path to the conclusion of my thesis. I would like to thank my entire family for their love and support. To my parents for their unwavering love, for always having my back, independently of success or failure they were always there for me supporting me in whatever challenge I decided to take on next, and for the values and principles they passed onto me, making me who I am today. To my brother, for always being that little voice that always challenges me to get more of myself, and for a few years ago putting that little bug in mind that made me want to experience what studying abroad was like. I must say I was fortunate enough to have had the ability to have had two of these experiences mostly thanks to him that made me realize how much of a unique
opportunity Erasmus or a Master’s abroad can be. To my grandfather Manel, wherever he may be, may he now scratch on his little notebook another accomplishment of one of his
grandsons.
To my friends which I have to thank in three separate groups, to those that have walked with since the early years of my life, you are always the ones I can rely on wherever the hour or day and whatever the issue, thanks for always being there for me. To the little family I got the opportunity to create in Graz, although now separated throughout several countries, you guys made me challenge myself in trying to keep up with your successes and your support was also key in this weird year. Finally, to the family I got the opportunity in this wonderful city. Specially Bahare, Laura, Pablo, Pari and Zarifa you guys made this experience so much more unique, even with a pandemic we all decided we would try and make the most of it. I will cherish the moments we’ve had throughout the year, and cannot wait for those that come in the future.
Finally, thank you to all those that took the time to complete my survey and share it, you were all essential for my success.
Thank you,
3
Abstract
Over the recent years, brands have started to shift their advertising strategies into the social media context (i.e., Instagram), aiming to increase levels of brand attitudes and brand engagement. Several strategies into implementing advertisements successfully in the social media context (i.e., Instagram) have arisen such as the use of personalization and influencers. This study investigated the extent to which the use of personalized advertisements can impact attitude towards the brand and brand engagement. Further, how the interaction between personalization and influencer can also affect brand attitudes. Lastly, this study aimed at assessing the relationship between brand attitudes and brand engagement, and if brand attitudes mediate the relation between personalization and brand engagement. A 2x (personalization vs non-personalization) x 2 (influencer vs no influencer) between-subject experimental design (n=151) showed that stronger brand attitudes lead to higher levels of brand engagement (Hypothesis 1). Furthermore, this study showed that the use of personalization in Instagram advertisements leads to stronger brand attitudes (Hypothesis 2), and also the brand attitudes partially mediate the relation between personalization and brand engagement (Hypothesis 3). Finally, this study showed that the interaction between the use of personalization and influencer was not significant (Hypothesis 4), although both factors showed to be impactful in achieving stronger brand attitudes.
Keywords: Brand Engagement, Brand Attitudes, OBA/Personalization, Influencer, Social Media,
4
Index
1. Introduction ... 6
2. Literature Review ... 9
2.1 Brand Engagement as specific type of consumer engagement ... 9
2.2 Brand Attitude as a step towards brand engagement ... 10
H1: Positive brand attitude will lead to higher brand engagement. ... 11
2.3 OBA in a social media context ... 11
2.3.2 OBA through Instagram ... 12
2.3.3 Personalized advertisement, taking Instagram as the chosen platform ... 12
2.4 The role of Instagram influencers on the relationship between personalization and brand attitude ... 14
2.5 Conceptual Framework ... 17
3. Research Method ... 18
3.1 Participants and sampling strategy ... 18
3.1.1 Excluding criteria ... 18 3.1.2 Sample size ... 18 3.2 Research Design ... 19 3.3 Materials ... 20 3.3.1 Instagram Timeline ... 20 3.3.2 Manipulations ... 20 3.4 Procedure ... 22 3.5 Measures ... 22 3.5.1 Brand Engagement ... 22 3.5.2 Brand Attitude ... 22 3.5.3 Manipulation Checks ... 23 3.6 Plan of Analysis ... 23 3.6.1 Manipulations ... 23 3.6.2 Hypotheses ... 24 4. Results ... 26 4.1 Manipulation Checks ... 26 4.1.1 Personalization ... 26 4.1.2 Influencer ... 26
4.2 Reliability and Validity ... 27
5
4.2.2 Brand Attitudes ... 28
4.3 Analyses of the relationship between the use of personalization and brand attitudes and brand engagement in a social media context. ... 28
4.3.1 Positive brand attitude will lead to higher brand engagement. ... 28
4.3.2 Exposing customers to a personalized advertisement on Instagram will lead to a more positive brand attitude than exposing customers to a non-personalized Instagram advertisement. ... 30
4.3.3 The relationship between personalization and brand engagement will be mediated by brand attitudes. ... 31
4.3.4 When ad is perceived as personal, using influencers will result in more positive brand attitudes than not using influencers. When an ad is perceived as not personal, using influencers will not be more effective to change brand attitudes than not using influencers. ... 32 5. Discussion ... 35 5.1 Theoretical Implications ... 37 5.2 Managerial Implications ... 38 5.3 Limitations ... 39 5.4 Future Research ... 39 Reference List ... 41 Appendix A – Conditions ... 47 Appendix B – Questionnaire ... 48
Appendix C – Scales of Variables ... 58
Appendix D – Assumptions ... 59
6
1. Introduction
Over the years, advertisers regard the use of Online Behavioral Advertising (OBA), through the presentation of personalized ads as one of the most important ways to reach a target audience (Tucker, 2012). OBA can be defined as “a type of digital advertising targeting method that tracks and compiles individual Internet users’ online behavioral data, such as what websites they visit, how long they stay there, and what they do (e.g., shopping; searching; surfing)” (Ham, 2017). In the online industry in particular, personalization of ads has become more common because it increases an ad’s effectiveness (Li, 2016), while less money is wasted on advertising (Nyheim et al., 2015). The development of social media platforms has further accelerated the use of personalized information in targeting efforts (Keyzer, Dens & Pelsmacker, 2015). Moreover, because social media users often provide (in)direct information about demographic, geographic, psychological and sociographic attributes, and all the users’ online actions are tracked, it provides vast opportunities for targeting efforts.
The advantages of OBA have been acknowledged by online retailers and consumers, especially in relation to developing more positive brand attitudes (Keyzer, Dens & Pelsmacker, 2015) and a stronger brand engagement (Walrave et al., 2016). However, OBA has also been associated with disadvantages. For example, personalized ads can be perceived as intrusive (Tucker, 2012), including feelings of privacy violations (Moore et al., 2015), and, vulnerability (Chen et al., 2018). These consequences of OBA decrease the effectiveness of such strategies (Chen et al., 2018; Bol et al., 2018), which consequently results in weaker attitude towards the brand and less strong brand engagement. This problem is summarized in the “personalization paradox,” which refers to the two-sided effect of personalization (Lee & Rha, 2016). On the one hand, “greater personalization typically increases service relevance and customer adoption,” however, on the other hand “…it also may increase customers’ sense of vulnerability and lower adoption rates” (Aguirre et al., 2015a, p. 34).
7 personalization outweigh the costs of it, and vice versa. However, the present study argues that the effectiveness of OBA would be more fully explained by accepting that the consumer is not a rational decision maker (Feigenbaum, Caliendo & Gahramanov, 2011), and, that the consumer is affected by bounded rationality (Simon, 1982). More specifically, consumers are social beings and influenced by the people around them (Bonabeau, 2004) ever more so on, an online context, particularly a social media one. An important cue, to avoid the possible negative outcomes that arise from the personalization paradox (Lee & Rha, 2016), is the use of influencers. Influencer marketing has been described as a type of native advertising, branded entertainment, or highly credible electronic word of mouth, because the commercial posts usually are woven seamlessly into the daily narratives that social-media influencers share with their followers (Breves, Liebers, Abt, & Kunze, 2019). Influencers can therefore have a role in how effective an OBA strategy can be.
A lot of studies on the effectiveness of OBA have been in general online contexts, such as mobile advertising (Chen & Hsieh, 2012), direct e-mailing (White et al. 2008), and advertising personalization on websites (Kox, Straathof & Zwart, 2017). However, less is known about how OBA works in a social media context (Tran et al. 2020). This is surprising since the personalization paradox might be more relevant in a social media context because the environment in which ads are shown includes the personal space of the receivers (Kelly, Kerr & Drennan, 2010). This space is designed to appear to be fully controlled by the receiver and receivers might be more strongly confronted with a sense of limited control of their own space when receiving personalized ads resulting in stronger feelings of reactance (Chen et al., 2018). Hence, the extent to which consumers may develop brand attitudes and engagement in a social media context might even be more strongly relying on their bounded rationality.
8 allows for a better connection with consumers, on such targeted advertising, that can ultimately lead to higher brand attitude and engagement. Because of such, this study will focus solely on Instagram as it is the most suited platform for companies to engage with consumers, and that ultimately will lead to the higher brand attitude and engagement.
Instagram is the platform that is most used by opinion leaders (influencers), due to the sense of immediacy that is generated and because of its creation of communities (Casaló, Flavián, & Ibáñez-Sánchez, 2020). Influencers are opinion leaders because of the communication they have with a sizable group of people who follow them (De Veirman, Cauberghe, & Hudders, 2017) Studies have shown that digital influencers can affect attitudes (Torres, Augusto, & Matos, 2019). When comparing brand posts and influencer posts, studies have shown that influencer posts lead to stronger/more positive brand attitudes (Jin & Muqaddam, 2019). They can therefore play a major role in helping brands attain brand engagement and improve attitudes towards the brand through social media platforms.
Therefore, not only personalized advertising plays a major role on marketing strategies of today, but there is also a role influencers can play in these strategies. With this study we aim to, not only gather from the current knowledge existent from the topics of social media advertising, online personalized advertising and influencers. But to also study the relation between the use of personalized ads on social media and influencers, and how this interaction can impact consumers brand attitudes and engagement.
9
2. Literature Review
2.1 Brand Engagement as specific type of consumer engagement
In recent years, the interest in OBA has grown considerably (Boerman et al. 2017). Personal data, such as visited websites, read articles, watched videos and made purchases, are used to create a personalized ad (Kox, Straathof, & Zwart, 2016). In the online industry, OBA has become more common, amongst others, because it may increase ad effectiveness (Li, 2016). The effectiveness of personalization has been measured in different ways, including click-through rates, attitudes towards the ad or brand, purchase intention, and actual purchase behavior (Boerman et al. 2017). The present study focuses on brand attitude and brand engagement as a key outcome measure for ad effectiveness, but firstly we will focus on brand engagement.
From the last few years, the concepts related to engagement have gained more relevance. Several types of engagement have been studied recently, yet the focus of this study is on brand engagement, as it is a very important tool for businesses. More specifically because it acts as a first step towards other brand goals such as brand loyalty (Khan, et al. 2019), purchase intentions (Barger et al.,2016), and a lot of other outcomes presented in the study of Gómez and Molina (2019), such as brand attachment, involvement and satisfaction. Although these terms are not the aim of this study, it shows the importance of brand engagement for brands to achieve competitive advantage.
Kumar and Pansari (2017, p.295) defined customer engagement as “the mechanics of a customer’s value addition to the firm, either through direct or/and indirect contribution”. They further develop on this definition by stating that the direct contributions are related to purchases while the indirect relate to referrals from customers. While for the specific context of brand engagement we can identify it as the interactive experiences between consumers and the brand, and/or other members of the community (Brodie, Ilic, Juric & Hollebeek, 2013). Brand engagement therefore, is a deepening of the term of consumer engagement with regard to the connections created between brands and consumers, in the several levels such as emotional and behavioral.
10 particular brand interactions”. However, in a revision made by Obilo, Chefor and Saleh (2019), to the conceptualization of Hollebeek (2011) showed that brand engagement is solely based on the behavioral aspect and does not include the emotional and cognitive areas. Other scholars have also conceptualized brand engagement on completely different scopes, due to this lack of consensus between scholars; this study will focus on an assessment of brand engagement on a general scope only.
In recent years, the growth of social media has been widely studied, especially regarding the engagement of consumers towards these platforms, as study from 2018 showed that people spent an average of 144 minutes daily on social media (Statista, 2018). Such numbers make these platforms essential for businesses to promote their brands. This makes it a vital place for brands to reach higher brand engagement, and gain competitive advantage towards other brands. As brand engagement is deeply connected to a gain of competitive advantage. Brand engagement is an essential tool that allows for brands to create connections with its consumers, even more so in a more and more digital world. Instagram is a platform that incentivizes engagement between brands and consumers, because it is the platform where users are more heavily engaged (Phua, Jin & Kim, 2017). It is therefore, the ideal platform to achieve higher levels of brand engagement, or as explained by its definitions, the creation of connection on behavioral, emotional and cognitive levels.
2.2 Brand Attitude as a step towards brand engagement
11 specifically, brand attitudes are the creation of favorable or unfavorable responses towards brands that will impact how people act towards such.
Brand attitudes are important to achieve brand engagement, as brand attitudes will ultimately be the guiding principle of the creation of a positive or negative response of consumers towards a brand based on the type of advertisement they receive (Loureiro, Gorgus & Kaufmann, 2017). But also due to the existent connection between the levels of brand engagement and brand attitudes. Based on the constructs of brand engagement (Hollebeek, 2011) and brand attitudes (McLeod, 2018), we can draw the similarity between the components of brand attitudes and the emotional and cognitive levels of brand engagement. Brand attitudes create the basis of a customer’s brand engagement since it will create the emotional and cognitive tools for how customers will behave towards the brand which is the ultimate level of brand engagement. Ultimately, this will impact how consumers will act towards brands. That is, the more positive the attitude towards the brand, the stronger the brand engagement.
Brand attitudes play an important role in determining how consumers perceive a brand, and this will impact not only what they think about it but also how they act towards it due to the relation between brand attitudes and brand engagement. In a society where social media has been growing constantly and brands are able to reach consumers much easier. It further shows the major importance social media will have for brands to project positive brand attitudes onto consumers. It is therefore important that brands make use of social media platforms properly so they can reach more positive brand attitudes and ultimately higher brand engagement. Therefore, we hypothesize:
H1: Positive brand attitude will lead to higher brand engagement.
2.3 OBA in a social media context
12 2009). Therefore, social media platforms have a role to play in the successful implementation of OBA in online contexts.
A lot of studies on the effectiveness of OBA have been in general online contexts, such as mobile advertising (Chen & Hsieh, 2012), direct e-mailing (White et al. 2008), and advertising personalization on websites (Kox, Straathof & Zwart, 2017). However, the influence of OBA in a social media context (Tran et al. 2020), has not been the focus of many studies. The limited research in social media contexts is surprising because personalization in such environments will be more accurate. Personalization in a social media context will lead to a lesser negative reaction to said personalized content (Kelly, Kerr & Drennan, 2010). OBA in the form of personalized advertisements is therefore expected to work better in a social media context than in other online contexts.
2.3.2 OBA through Instagram
Studies of OBA in social media context are scarce, and have mostly focused on Facebook (Tran et al., 2020). However, Instagram is one of the main social media platforms that exist today, with a reported user base in 2019 of 788.4million in 2019 which acquaints to 20.6% of internet users as reported by eMarketer (2019). Phua, Jin & Kim, (2017) showed that from the existing social media platforms, as the likes of Facebook, Twitter, Instagram, etc. Instagram showed to be the platform where users showed highest brand engagement, with these being the ones more likely to participate in brand activities, follow community rules and being the ones that retain loyalty toward the brand over a longer period of time. Finally, several reports show how businesses are facing the social media platform in their strategically planning for their businesses, with several reports that can be found on eMarketer showing that companies planned to continue to increase the amount of money dedicated to social media marketing, not only focusing on social advertising but also in the talent and tools necessary to make such processes function properly (eMarketer, 2009b). Since we wish to evaluate brand attitude and engagement of social media advertising it makes only sense to focus our study in the platform where its users have been described as the more engaging user base with regard to this specific social media platform. Even more so, when as shown before that engagement is based on the interactive experiences that happen between consumers and brands.
2.3.3 Personalized advertisement, taking Instagram as the chosen platform
13 to the right consumer at the point in time that maximizes immediate and future business opportunities. (Aguirre, Mahr, Grewal, et al., 2015). Although these strategies have been known to happen in online and offline platforms, focus on such topic has seen interest spark with the growth of the digital world, and our dependence on the online tools that have been made available to us in the past few years.
When it comes to personalization of advertising in the online environment, advertisers have taken the opportunity to utilize data that is made available by users, even sometimes data that users are not aware they are giving away, to create personalized offers and opportunities for their consumers. Such data can include websites visited, articles read, and videos watched, as well as everything searched for with a search engine (Ham, 2017).
With regard to personalization in the social media context, some studies have been made regarding the impact that personalized ads can have on social media platforms (Taylor, Lewin & Strutton, 2011; Tran, 2017). The study of Li (2016), showed that the majority of studies related to personalized ads lead to a positive impact of the use of personalized ads on consumers likening and also providing more favorable effects. Furthermore, the study of Walrave et al. (2016) and Keyzer, Dens and Pelsmacker (2015), show the positive impact that the use of personalized advertisement on social networking sites has on brand engagement and also on attitudes towards the brand. The use of personalized ads in social media will be impactful towards the creation of brand attitudes and brand engagement. It is therefore expected that the use of personalized advertisement on Instagram will allow us to achieve more positive brand attitudes than non-personalized advertisement. This in turn will lead to a stronger brand engagement.
Therefore, we hypothesize:
H2: Exposing customers to a personalized advertisement on Instagram will lead to a more positive brand attitude than exposing customers to a non-personalized Instagram advertisement.
14 concept. By showing them a product, which is in line with their current wants, we are impacting the individual’s self-concept. Such will lead to higher levels of congruence which result in better attitudes towards the brand. Ultimately these higher levels of congruence that lead to better attitudes will end in better brand engagement, as theorized by Loureiro, Gorgus and Kaufmann (2017, p.990) “The match between consumers’ self-image and the brand image is regarded a motivational factor for thinking more proactively about the brand, feeling good about it and using it, hence, being more engaged with the brand.” Therefore, high levels of congruence are important in achieving strong brand attitudes, which ultimately impact how customers engage with the brand.
Therefore, we hypothesize:
H3: The relationship between personalization and brand engagement will be mediated by brand
attitudes.
2.4 The role of Instagram influencers on the relationship between
personalization and brand attitude
The mediating role of attitude towards the brand between personalization and brand engagement explains the positive impact of personalized information/OBA in social media. However, it fails to explain why personalization is sometimes extremely effective while at other times it is not effective at all, or sometimes even counter-effective, referring to the personalization paradox (Lee & Rha, 2016). The presence of social influencers on social media platforms, such as Instagram, might provide an explanation for the (in)effectiveness of personalization.
15 hence, people are expected to be less resistant to the messages that originate from such said influencers, when in comparison with highly known celebrities.
Maio and Olson (2000), introduced a distinction between two groups of people, high self-monitors and low self-self-monitors. High self-self-monitors are characterized by being concerned with the image they project, and adapt themselves to follow the attitudes that reference groups are projecting. Low self-monitors, do not shape their behavior to fit specific reference groups, they focus on themselves. Taking in consideration the study of Phua, Jin & Kim (2017) we are able to see that the users of Instagram are those that use the platform, to keep up with fashion, showing affection and demonstrating sociability. It is therefore likely that in general Instagram users score relatively high on self-monitoring. Furthermore, one of the suggested ways to affect the development of positive attitudes of high self-monitors is by using spokespersons (Maio & Olson, 2000). Spokespersons, especially more attractive ones, have been shown to promote positive attitude change towards the promoted product for high monitors compared to low-monitors (DeBono, & Harnish, 1988). As influencers can provide their opinions and reviews of brands and promote them to their followers (Sokolova & Kefi, 2020), they act as spokespersons for brands. Being that people that follow them on Instagram do so because they are attracted to the influencer, whether because of looks or the content they share. Furthermore, the study of Torres, Augusto & Matos (2019), showed that influencers play a major role in attitude formation. Other studies have also shown how the use of influencers in Instagram posts lead to better brand attitudes, than just being the brand posting such advertisements (Jin & Muqaddam, 2019). We can therefore draw a connection with our study, and predict that the use of influencers will positively impact the attitudes towards the brand since high self-monitors are more influenced by spokespersons.
16 individual. Further, the perceptions about what other people think and do, as presented by influencers, might also provide such cue. High monitors might be even more sensitive for such “social” peripheral cue as they value external opinions even more (Maio & Olson, 2000). Therefore, for high-monitors, such as Instagram users, providing personalized information will be taken even more seriously when presented by a relevant social influencer than presented on its own. Consequently, both the use of personalization and influencer are important cues in attaining stronger brand attitudes. Furthermore, the use of personalization will most likely result in a more positive attitude towards the brand when combined with the presence of an influencer.
Influencers make salient relevant social norms in relation to the brand, as they have the ability to set trends on what is interesting or not. Influencers can therefore act as an overall rule of what to do in social terms, what is trending, what is cool. As we theorized before, users of Instagram can be linked with high self-monitors (Maio & Olson, 2000), because they are highly engaged in adapting to what the trend is of the moment. Influencers that were able to build following bases have therefore the potential to influence those that follow them through emotional bonds created between influencer and followers (Ki, Cuevas, Chong & Lim, 2020). It is therefore clear, the role that influencers play in this study, as together with personalization they can be a very important tool, in facilitating the connection between brands and consumers. Acting as peripheral cues that facilitate the creation of attitudes towards the brand. It is for such reasons that influencers play a major role in the impact that personalized advertisements on Instagram have on brand attitudes and ultimately on brand engagement.
Therefore, we hypothesize:
H4: When ad is perceived as personal, using influencers will result in more positive brand
17
2.5 Conceptual Framework
Based on the previous literature review presented the conceptual framework was originated. Such framework can be observed below in Figure 1, and is the basis of this research.
18
3. Research Method
This study examined if personalized advertisements on Instagram have a more positive relationship on brand engagement, by having a stronger impact in attaining positive brand attitudes, and, whether the use of influencers can positively impact this relationship.
3.1 Participants and sampling strategy
The target group of this research was Instagram users, therefore those not familiar with the platform were excluded from participating. The sampling technique that was used was convenience sampling. Since it allows us to have a large enough varied sample, and also allows us to fully randomize participants within each condition, since the main aim of this study was to assess the impact of how the different relationships between the independent variables impact our dependent variables. The data needed to conduct the research was gathered via an online experiment, and participation was entirely voluntarily. The participants were provided with an online link, which directly sent them to the online experiment that was created via the software Qualtrics.
3.1.1 Excluding criteria
To conduct the experiment properly, two filter questions were asked at the start of the experiment. The first filter question asked whether or not the respondent is a user of the social media platform Instagram. If the respondent answered the question with ‘no,’ the respondent was directly directed to the end of the questionnaire. Because, responses from someone not familiar with Instagram would be not representative and any valid results would not be possible to be drawn. The second filter question pertained if the participant knew any influencers. If the response given was “no”, the respondent was directly directed to the end of the questionnaire. This decision was made due to the fact that for two of the conditions presented to participants, an influencer name provided by the participants was required. If we were to allow participants that did not know any influencers to proceed, we would be introducing bias into the study. Since such participants would only be able to take part in the condition pertaining to no influencer. Therefore, to not impact the validity of the study, the decision was to also exclude these participants. After the respondent answered that they are familiar with the social media platform, they proceeded to the main experiment.
3.1.2 Sample size
19 larger than 30 participants is large enough to measure group differences, therefore we needed at least 30 participants per condition making a total of 120 participants.
In total 294 participants completed the survey. 129 participants were excluded, either because they were not familiar with the social media platform Instagram(n=29), or because they did not know any influencers(n=102). Furthermore, 5 more participants were excluded due to the fact that when asked to name one influencer, they replied “no”, and 2 for being under 18 years old. One participant was excluded because this person did not answer any of the remaining questions after providing answers on the excluding criteria. In addition, while assessing for the assumptions of the chosen tests to perform, six more outliers were found and removed from the dataset, the criteria for these outliers is presented in the Plan of Analysis. Such was done in order to prevent any impact from these outliers in the outcomes of our dataset.
The final analyses included 151 participants. The sample consisted of 53.00% Male, 45.70% Female, one participant identified as Other .70% and another participant preferred not to state gender .70%. The participant’s ages ranged from 18 years old to 64 years old (Mage=30.65,
SD=11.12). The majority of the participants was of Portuguese nationality (74.20%), followed by
UK (4.60%), Germany (3.30%), Netherlands (3.30%), Austria (2.60%), Italy (2.00%), the remaining nationalities were all below the 2.0% margin, equivalent to 2 or less participants.
3.2 Research Design
This study used a 2 (personalized vs. non-personalized) x 2 (Use of influencer: yes vs. no) between factorial design, which is displayed in Table 1. The effects of the manipulation variables (personalization and influencer) were hypothesized to influence brand attitude, which was the dependent variable. Brand attitude was included as the mediator variable between personalization and brand engagement, which was the dependent variable of the overall model. Table 1
Condition Type of Ad Use of Influencer
1 Personalized Yes
2 Personalized No
3 Non-Personalized Yes
20
3.3 Materials
3.3.1 Instagram Timeline
The Instagram posts used in the experiment were either featuring the brand promoting the product the respondents are deemed to want, or a general online retailer. The product chosen to be presented as a need for the respondents will be a pair of Bluetooth headphones. This product was chosen because it can be deemed as a more common product for the majority of the population but also because its price can generally be placed within a very wide range, which gives room for participants to assess the value of the product by themselves, and therefore the study does not impact their judgment. Furthermore, it is a need that could be generalized for both the male and female audience, since it is a gender-neutral product.
Furthermore, we used a fictional brand named “Musik” for the headphones rather than an existing brand to avoid confounding effects of pre-built attitudes towards existing brands. For the same reasons as for the product, we also used a fictional online retailer named “ShopOn” in the experimental condition including the general online retailer.
Finally, all four experimental conditions (i.e., Instagram posts) looked similar (e.g., same lay-out of the ad, same price per product, etc.) to ensure treatment equivalence. Appendix A represents the final four experimental conditions, which were created using the website zeoob.com/generate-Instagram-post.
3.3.2 Manipulations
A scenario-based experiment was used in which participants were asked to take the perspective of one specific person. The scenario portrayed someone who was currently searching for a pair of new Bluetooth headphones, in the following way:
“Imagine yourself as being someone who is currently looking for a new pair of headphones, especially a Bluetooth pair of headphones. You have already been searching on the internet for a pair, but haven’t yet been able to find what you want. Recently you logged on to Instagram and came across the following ad post on your timeline…”
Two different ads were integrated into the specific scenario to manipulate personalization of an ad: a post from a made-up brand connected with products related to sound reproduction (personalization) or a post from a made-up online retailer shop (no personalization).
21 for this online. The online retailer ad was shown to the non-personalized group, since an ad to an online retailer brand was not immediately fulfilling the current want from the person. The second manipulation variable, influencer (present or not present), was manipulated in the scenario by manipulating the information present on the post. The post containing influencers were identified with the name of the influencer as mentioned in the pre-questions by the participant including a comment of this influencer inviting those who follow him/her/them to try this new brand. We used an influencer mentioned by the participant rather than choosing one beforehand ourselves to avoid (1) unfamiliarity of participants, and, (2) any pre-built negative attitudes, towards a pre-specified influencer. For the “no influencer” experimental condition, a normal person was used to identify the post and a simple comment describing that this was the brand said person now used. Furthermore, the difference in the likes present on the post was also made significant to differentiate the influencer (i.e., significantly more likes) from the regular person. Figure 1 below gives an overview of the four conditions introduced.
22
3.4 Procedure
The survey started by showing an introduction text to participants explaining the goal of this study. Following that participants were asked socio demographic questions, regarding age, gender and nationality. Next, the survey started with asking whether participants had an Instagram account. When participants indicated that they did not own an Instagram account or they did not know any influencers, they were thanked and excluded from further participation (n=129).
Then participants were randomly assigned to one of the four experimental conditions: they either received the personalized ad with influencer (n=42), or, they received the low personalized ad with no influencer (n=38). For the non-personalized condition, they either received a non-personalized ad with influencer (n=36), or, a non-personalized ad with no influencer (n=35). Next, they received questions regarding personalization, and influencers. The order in which the participants received the questions about personalization and influencers were randomized. Following this, attitude towards the brand was measured. The items regarding personalization, and influencer were measured prior the attitude items, because this gave participants the opportunity to form an attitude towards a brand they were not familiar with. Finally, they were presented with the questions regarding brand engagement. The study concluded by thanking participants and providing form of contact if any information is required by participants. The questionnaire is shown in Appendix B.
3.5 Measures
3.5.1 Brand Engagement
Ten items will be used to measure brand engagement. These 10 items are taken from a validated scale by Hollebeek, Glynn & Brodie (2014). All items will be measured using a seven-point Likert scale (See Appendix C).
3.5.2 Brand Attitude
23
3.5.3 Manipulation Checks
To assess the manipulation check personalization, the scale validated by Li (2016) was used. This scale consisted of two items, “The advertisement seems to be designed specifically for me” and “The advertisement targets me as a unique individual”. Such was measured on a seven-point Likert Scale. Due to the lack of existent validated scales to measure our manipulation regarding influencer, the measurements to check this manipulation were created by us. To assess the manipulation for the use of influencer or not two items were introduced, namely “I am familiar with the person that posted the Instagram post” and “The person in the post can be deemed as an influencer”. This measure was also measured on a seven-point Likert scale.
3.6 Plan of Analysis
The software that has been used to analyze the data from the questionnaire is SPSS, version 26.
3.6.1 Manipulations
To verify that the manipulations worked as we wanted an independent sample t-test was used. Personalization and use or not of an influencer were introduced as independent variables and the manipulation check questions were used as dependent variables. Before we were able to perform the independent samples t-test, assumptions were checked.
Assumptions t-test
24 other. To test such a Levene’s test for equal variances was executed. Personalization (1=personalized, 2=non-personalized) was introduced as the independent variable, and the manipulation check as the dependent variable. Levene’s test for equality of variances revealed a p>.05 namely p=.12, which means equal variances were assumed. For the Influencer independent variable, a Levene’s test for equal variances was executed. Influencer (1=yes, 2=no) was introduced as the independent variable, and the manipulation checks as dependent variables. For the first manipulation check regarding influencer, Levene’s test for equality of variances revealed a p>.05 namely p=.22. For the second manipulation check regarding influencer p=.18. Such means that both manipulation checks were assumed as having equal variances.
3.6.2 Hypotheses
In order to test our hypotheses different methods were used, a Linear Regression Analysis and a two-way ANOVA. The Linear Regression analysis allowed us to test the relationship between our independent and dependent variables, brand attitude and brand engagement (H1). The impact of the use of personalized or non-personalized ad towards brand attitudes (H2) was measured through a two-way ANOVA. The mediating effect of brand attitude on the relationship between personalization and brand engagement (H3), was tested through Hayes Macro, Process model 4. Further, to assess the outcomes we used Baron and Kenny (1986) method. The impact that the use of influencers has towards brand attitudes in personalized or non-personalized advertisements (H4), was measured through the same two-way ANOVA introduced for the testing of hypothesis 2, this allows us to compare the differences between the four different conditions we have presented in this study. Certain assumptions had to be met to proceed with this analysis.
Assumptions Linear Regression Analysis
25 dataset. The fourth assumption states that there should exist independence of observations, such was tested through a Durbin-Watson statistic. The test confirmed that we have independence of observations since the value we should obtain should be around 2, and the Durbin-Watson statistic returned as 2.09. So, this assumption was met. The fifth assumption states that there should be homoscedasticity of data. This was confirmed through visual inspection of the plot of standardized residuals versus standardized predicted values (see Appendix D). This assumption was also met. The sixth assumption states that residual errors of the regression line should be approximately normally distributed. Based on the previous information introduced regarding the Central Limit Theorem (Pallant, 2005), since we have n>30 for each condition, normality of distribution can be assumed.
Assumptions Two-Way ANOVA
26
4. Results
4.1 Manipulation Checks
In order to check if the two manipulations, personalization (yes vs. no), and use of influencer (yes vs. no) had the desired outcome, two manipulation checks were conducted. These manipulation checks were done to evaluate if our manipulations were successful.
4.1.1 Personalization
The personalization manipulation check consisted of two items introduced and validated by Li (2016), “The advertisement seems to be designed specifically for me” and “The advertisement targets me as a unique individual”. The internal consistency for these two items was tested, Cronbach’s alpha returned as .75, which is above the minimum value of .60 that allows us to combine items into one variable. A new variable Personalization was computed based on this by adding the two items and then dividing them by 2 (M=3.36, SD=1.62).
To analyze if our manipulation regarding personalization was successful, between the participants that received personalized ads and those who did not, we performed an independent samples t-test. Personalization was entered as a dependent variable, and as independent variable, the conditions (1=personalized, 2=non-personalized) each participant was assigned to. The participants who received a personalized advertisement (M=3.57, SD=1.68) perceived a slightly higher level of personalization, than the participants that received a non-personalized advertisement (M=3.11, SD=1.52). The independent samples t-test showed a borderline significant outcome, t(153)=1.81, p=.079. Therefore, our manipulation personalization was somewhat successful, although it could have been stronger.
4.1.2 Influencer
27 To analyze our manipulation regarding the use of influencer or not was successful, we performed an independent samples t-test. The variable influencer was introduced as the dependent variable, and as the independent variable, the conditions (1=influencer, 2=no influencer) for which each participant was assigned to. Participants who were placed in the condition of influencer (M=5.10, SD=1.27) perceived the person in the post as a significantly stronger influencer than those who received the no influencer condition (M=3.25, SD=1.19),
t(150)=9.18,p<0.001. Therefore, we can conclude that our manipulation regarding influencer
was successful.
4.2 Reliability and Validity
For the dependent variable brand engagement, the scale developed and validated by Hollebeek, Glynn and Brodie (2014) was used; this scale consisted of 10 different items. For the dependent variable brand attitude, the scale developed and validated by Spears and Singh (2004) was used; this scale consisted of 5 different items. For both dependent variables a Principal Component Analysis (PCA) was conducted, in order to reduce these items into fewer variables. A PCA allows for the reduction of a large dataset by taking into account the variability in the pattern of correlations (Pallant, 2005). Furthermore, for both dependent variables, a Cronbach’s alpha test was conducted to analyze if we were allowed to merge the several items in both cases into one new variable.
4.2.1 Brand Engagement
Two PCA’s were conducted for the 10 items regarding brand engagement, one for each of the results obtained for the response sets of each brand, based on the scale of Hollebeek, Glynn and Brodie (2014). The PCA regarding the set of items for “Musik” revealed a KMO=.89, and Bartlett’s test of sphericity showed a significant score of p<.001. The communalities were all above .40. For “ShopOn”, the PCA showed a KMO=.90 and Bartlett’s test of sphericity showed a significant score of p<.001. The communalities were all above .40. Therefore, for both cases, we proceeded with all 10 items.
28 by 10. “Musik” (M=3.8, SD=1.21), “ShopOn” (M=3.31, SD=1.43). These two were then combined, creating the final dependent variable “EngagementTotal” (M=3.55, SD=1.33).
4.2.2 Brand Attitudes
Two PCA’s were conducted with all 5 items intended to measure brand attitude, taken from the scale of Spears and Singh (2004), one for “Musik” and one for “ShopOn”. The Kaiser-Meyer-Olkin (KMO) revealed sufficient in both cases to proceed, “Musik” KMO=.84 and for “ShopOn” KMO=.86. Furthermore, the Bartlett’s test of sphericity showed significant results in both cases of p<.001. With regard to the communalities for both “Musik” and “ShopOn”, these were all above .40, so we proceed with all 5 items for both cases.
In addition to the PCA analysis, a Cronbach’s alpha test was conducted on the 5 items regarding brand attitude. The Cronbach’s alpha test revealed for the case of “Musik” a value of .94 meaning that the 5 items regarding brand attitude are highly correlated. For the case of “ShopOn” the Cronbach’ alpha test of reliability revealed a value of .92, which also show high correlation between the 5 items. Based on this a new variable was created for each brand by summing the 5 items and dividing by 5 for each case, “AttitudeMusikT “(M=4.33, SD=1.15) and ”AttitudeShopOnT” (M=3.88, SD=1.14). These two where then combine into the final dependent variable “AttitudeTotal” (M=4.13, SD=1.16).
4.3 Analyses of the relationship between the use of personalization and
brand attitudes and brand engagement in a social media context.
The hypotheses were tested using a Linear Regression Analysis and a two-way ANOVA. Hypothesis 1 will discuss if brand engagement can be estimated through the level of brand attitude, specifically in our case we hypothesize that more positive brand attitudes will lead to higher brand engagement, this was tested through a Linear Regression Analysis. For the testing of hypotheses 2 and 4 a two-way ANOVA was used, where we expect for the use of personalization to lead to more positive brand attitudes than non-personalized advertisement, hypothesis 2. Hypothesis 4 expect an interaction between the use of personalization and influencer, leading to more positive brand attitude, when compared with sole use of influencer or personalization. Hypothesis 3 measured if brand attitudes mediates the relationship between personalization and brand engagement and was assessed through Hayes Process model 4.
4.3.1 Positive brand attitude will lead to higher brand engagement.
29 The linear regression while correcting for the remaining two independent variables, personalization and influencer, showed a significant statistical effect (Table 2), R2=.41 F(1, 138) = 32.03, p < .005. Further, the analysis showed that there is a positive relation between brand attitudes and brand engagement B=0.52, t(141)=6.55, p<.001, which shows that there is a positive relation between brand attitudes and brand engagement, which means that higher brand attitudes will lead to stronger brand engagement. Therefore, we can conclude that, not only we have a statistically significant linear regression between brand attitudes and brand engagement, but also that brand attitudes positively impact brand engagement. Therefore, Hypotheses 1 is supported.
Table 2 Regression Analysis Summary
Variable Coefficient Std. Error t-Statistic Sig.
Constant (Brand Engagement) .32 .35 .92 .36
Brand Attitudes .52 .08 6.55 .00
Personalization .18 .06 3.16 .00
Influencer .11 .06 1.87 .06
R .64 R- Squared .41
Adjusted R-squared .39 F-statistic 32.03
S.E. of regression 1.02 Sig. (F- statistic) .00
30
4.3.2 Exposing customers to a personalized advertisement on Instagram will lead to a more
positive brand attitude than exposing customers to a non-personalized Instagram advertisement.
Table 3- Test of Between-Subjects Effects. Interaction effect of Personalization and Influencer on Brand Attitude. R Squared = .068 (Adjusted R Square=.048)
Source df Mean
Square F Sig. Cohen’s d
Personalization 1 7.774 6.071 .015 .42
Influencer 1 3.309 2.584 .110 .27
Personalization*Influencer 1 2.780 2.171 .143 .72
Hypotheses 2 proposed that a personalized advertisement on an Instagram post will lead to more positive attitudes than using a non-personalized advertisement Instagram post. To analyze whether or not brand attitudes were more positively influenced by the use of personalization or not, we performed a two-way ANOVA.
Table 4 – Marginal Means
The mean scores for the personalized versus non-personalized conditions (Figure 2) showed that participants that were exposed to a personalized advertisement showed more positive brand attitudes (M=4.36, SD=1.15), than participants that were exposed to non-personalized advertisement (M=3.88, SD=1.14). Through the analysis of these values, we can see that the use of personalization has led to more positive brand attitudes than not using personalization. Furthermore, Table 3, shows us that p. value for personalization is significant on brand attitudes, since p<.05, however the using Cohen’s criterion (1988) the expected effect of personalization on brand attitudes is medium, by calculating Cohen’s d we get an effect of .42. The two-way
Personalization Influencer Mean SD
Personalized Influencer 4,37 1,15
NoInfluencer 4,34 1,16
NonPersonalized Influencer 4,18 1,24
31 ANOVA analysis revealed that although the effects are small, personalization has a significant statistical effect. Therefore, we can assume hypotheses 2 as confirmed.
Figure 3 Mean differences between personalized advertisement and non-personalized advertisement on brand attitudes
4.3.3 The relationship between personalization and brand engagement will be mediated by
brand attitudes.
Hypotheses 3 predicted that the relationship between personalization and brand engagement will be mediated by brand attitudes. To analyze such mediation effect, we used Hayes Macro and followed the process introduced by Baron and Kenny (1986).
32 In Step 1 of the mediation model, the regression of personalization on brand engagement, ignoring the mediator, was significant, B = .34, t(139) = 5.47, p = <.001. Step 2 showed that the regression of personalization on the mediator, brand attitudes, was also significant, B = .28,
t(139) = 4.92, p = <.001. Step 3 of the mediation process showed that the mediator (brand
attitudes), controlling for personalization, was significant, B = .64, t(139) = 8.31, p = <.001. Step 4 of the analyses revealed that, controlling for the mediator (brand attitudes), personalization scores was a significant predictor of brand engagement, B = .17, t(138) = 2.96, p = .0036. So, brand engagement depends on the use of personalization, and this relation is partially mediated by brand attitudes. Furthermore, the test showed that zero is not within our interval rejecting the null hypothesis. We can conclude that we have mediation, however since the effect of X on Y (c’) is still significant, we only have partial mediation. Therefore, we can assume hypotheses 3 as confirmed.
4.3.4 When ad is perceived as personal, using influencers will result in more positive brand
attitudes than not using influencers. When an ad is perceived as not personal, using influencers will not be more effective to change brand attitudes than not using influencers.
Hypotheses 4 predicted that using influencers in a personalized Instagram advertisement will lead to more positive brand attitudes than not using influencer. Further, it predicted that for non-personalized Instagram advertisements using influencers or not would not make a difference in changing brand attitudes.
33 To analyze whether there was a significant interaction effect between the use of personalization and the use of influencer, a two-way ANOVA was introduced. The mean scores of the interaction between the use of personalization and influencer versus use personalization and no influencer (Figure 5), showed that participants in the first condition showed a slightly higher level of brand attitudes (M=4.37, SD=1.15), when compared with the second condition (M=4.34, SD=1.16). The small differences in the mean scores do not only go against the expected outcomes, but also show that the interaction effect between the use of personalization and use of influencer did not significantly impact brand attitudes. Therefore, the analysis shows that our interaction effect is not statistically significant. Further, the interaction between non-personalized advertisement and influencer, showed a very different scenario (Figure 5). Participants in the condition of no personalization and use of influencer (M=4.18, SD=1.24), showed higher values towards brand attitudes when compared with those that were in the condition regarding no personalization and no influencer (M=3.59, SD=.97). Further, when assessing for Cohen’s d we can identify a medium effect of .52. Although such was not hypothesized in this study, the data gathered from the analyses shows an interesting trend of results. The results show that brand attitudes can be attained through three different paths, being that only for the case of no presence of personalization and no influencer is highly prejudicial to the brand evaluation of an individual. Which could indicate that personalization is not present, a solution could be the introduction of influencer, and the opposite also applies. Nevertheless, the test of between-subjects effect showed a score of F(1,142)=2.17, p=.143 (Table 3), which rejects our hypotheses 3.
Figure 5. Interaction effect between Personalization and use of Influencer towards Brand Attitude
4,37 4,34 4,18 3,59 0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0
Influencer NoInfluencer Influencer NoInfluencer
34 The table below gives an overview of the results that are found while the hypotheses were tested (Table 5).
Table 5 Hypotheses Overview
Hypotheses Findings
H1: Positive brand attitude will lead to higher brand engagement.
Supported
H2: Exposing customers to a personalized advertisement on Instagram will lead to a more positive brand attitude than exposing customers to a
non-personalized Instagram advertisement. Supported
H3: The relationship between personalization and brand engagement will be
mediated by brand attitudes. Supported
H4: When ad is perceived as personal, using influencers will result in more positive brand attitudes than not using influencers. When an ad is perceived as not personal, using influencers will not be more effective to change brand attitudes than not using influencers.
35
5. Discussion
The present study examined the possibility of positive brand attitudes leading to higher levels of brand engagement, it was also examined in this study the relation between the use of personalization and brand attitudes. Further, this study examined is brand attitudes acted as a mediator of the relation between personalization and brand engagement. Lastly this study focused on analyzing the interaction effect between the use of personalization and influencer towards brand attitudes.
Our results found that a positive brand attitude results in higher levels of brand engagement, which supported Hypothesis 1. These findings are in line with the theorizations put forward by this study, building from the concepts of brand engagement (Hollebeek, 2011) and brand attitudes (McLeod, 2018). Further, this study helps in confirming the definition put forward by Hollebeek (2011), by confirming the three levels of brand engagement and rejecting the revision put forward by Obilo, Chefor and Saleh (2019), that affirmed that brand engagement relied solely on the behavioral level. From these conceptualizations this study was able to draw the connection between attitudes and engagement, where attitudes draw the emotional and cognitive basis for the behavioral aspect of engagement.
With regard to hypothesis 2, this study examined the possibility of a relationship between the use of personalization in advertisements and brand attitudes. Previous research (Keyzer, Dens & Pelsmacker, 2015; Zhu & Chang, 2016; Vernuccio et al., 2012) had already established that the use of personalization in advertisements will lead to positive outcomes, specifically in brand attitudes. Therefore, we expected that the use of personalization in this study would lead to better brand attitudes when compared with non-personalized advertisements.
36 With regard to hypothesis 3, this study built on the previously introduced constructs and predicted that brand attitudes would mediate the relationship between the use of personalization and brand engagement. Previous studies showed that when presented with high levels of self-congruence consumers will have better attitudes towards the brand, which will lead to higher brand engagement (Loureiro, Gorgus & Kaufmann, 2017). We therefore, theorized that the use of personalization would increase the levels of self-congruence and prove the existent mediation of brand attitudes in the relationship between personalization and brand engagement. The results yield a significant outcome, which proved that the relationship between personalization and brand engagement is mediated by brand attitudes. However, the results obtained showed that the existent mediation is only a partial mediation. This could be because, already it has been studied that the use of personalization is correlated with higher levels of brand engagement (Walrave et al, 2016). Another explanation for this might be the fact that our manipulation personalization was only borderline significant, meaning that our manipulation might have not been entirely successful in distinguishing between the personalized versus non-personalized experimental conditions.
37 of several studies into the relationship between influencers and brand attitudes (Torres, Augusto & Matos 2019; Jin & Muqaddam, 2019). So, our hypotheses 4 ended by being unsupported, however some interesting conclusions can still be drawn based on the outcomes provided by our analysis. It also confirms that by producing high levels of self-congruency (Sigry, 1982), whether through personalization or influencer it will lead to better brand attitudes (Loureiro, Gorgus & Kaufmann, 2017). Furthermore, it strengthens the idea that both personalization and influencers can act as peripheral cues in the attitude creation process, based on the ELM model by Petty and Cacioppo (1986).
5.1 Theoretical Implications
Our findings build on the existing research, in the concepts of brand attitudes and brand engagement, of OBA/personalization and the use of influencers in the social media context. Further, this study builds on the existing concepts to analyze the possible existent relations between some of these concepts. Brand attitudes leading to brand engagement, personalization leading to stronger attitudes towards the brand, brand attitudes acting as a mediator between personalization and brand engagement, and finally the interaction effect between personalization and the use of influencer towards brand attitudes.
Building on the existing research related with brand attitudes (Allport, 1935; Kotler & Keller, 2006) and brand engagement (Hollebeek, 2011). Our findings contribute by introducing the knowledge of how to reach higher levels of these factors in a social media context. Furthermore, our finding contributes to the knowledge of how brand engagement is highly impacted by brand attitudes, in the direct relationship between these two factors. By further developing both these concepts, our study reached the theoretical implication that brand attitudes are the building construct of brand engagement with regard to its emotional and cognitive levels. Being that brand attitudes create the emotional and cognitive of brand engagement that are completed with the behavioral level that builds on the previous two. But also, how brand attitudes can mediate the relationship between the use of personalization and brand engagement.
38 highly impactful in achieving higher brand attitudes. In addition, also the combination of personalization and influencer together was tested and did not generate enough difference in the outcome of brand attitudes, to prove useful. However, when personalization was not present, influencer was highly significant in improving attitude towards the brand, further supporting the studies of Torres, Augusto and Matos (2019); Jin and Muqaddam (2019).
5.2 Managerial Implications
This study provides useful insights for brands firstly, if a brand wants to increase the levels of engagement towards the brand from its consumers, it should focus on the creation of strategies that allow for the creation of stronger brand attitudes. This study showed that a reliable strategy in achieving stronger brand attitudes is through the use of personalized advertisements on Instagram. Companies should therefore focus on the introduction of personalization in their Instagram advertisements, as such is a strong tool into achieving better brand attitudes. And being that ultimate goal is to achieve better brand engagement, as this study showed brand attitudes are key for that process.
39
5.3 Limitations
This study has some limitations which will be discussed. First, the manipulation check of personalization was borderline significant only, and therefore did not work as desired. It might be the case that the distinction between the conditions personalized and non-personalized was not sufficient in differentiating between both situations for our participants. By introducing an online retailer as the non-personalized target, some participants might have still been led to believe the ad was personalized to them since online retailers sell a large array of products. With some like Amazon also having Bluetooth headphones available for purchase through their website. A stronger differentiation between the personalized and non-personalized condition could have been used to completely avoid the possibility of participants linking the non-personalized condition with personalization.
Second, another limitation of this study was with regard to the sample size; although our sample reached the minimum levels sketched out by our theorization. The sample size proved to be small, specifically due to the excluding criteria included in the study. The inclusion of two excluding criteria although majorly important so that our conditions worked in the expected way, ended being very heavy. As a result of the two excluding criteria 129 participants were excluded from the study, which proved to be almost half of the total of participants that responded to the questionnaire. This could have been possibly avoided by incorporating different strategies in the drawing of the study, firstly instead of asking participants to name an influencer a fictional one could have been created. alternatively, instead of asking participants to name an influencer a list of influencers could have been provided, avoiding a momentary blank of memory when asked to name an influencer.
5.4 Future Research
Future research should aim at providing more insights into the effects of OBA in a social media context because this study showed that in a social media context, specifically Instagram, OBA is highly significant in impacting brand attitudes and engagement. Therefore, the study of how such behaves in different social media platforms would prove interesting to assess if the results are specific only to Instagram, or if they are similar across all platforms. An interesting case would be the YouTube case, based on the fact that so many people now use ad blockers, how would a study assess for what was presented in this study.
41
Reference List
Aguirre, E., Mahr, D., Grewal, D., de Ruyter, K., & Wetzels, M. (2015). Unraveling the personalization paradox: The effect of information collection and trust-building strategies on online advertisement effectiveness. Journal of Retailing, 91(1), 34–49.
Allport, G. W. (1935). Attitudes. In C. Murchison (Ed.), Handbook of social psychology (pp. 798844). Worcester, MA: Clark University Press.
Barger, V. A., Peltier, J. W., & Schultz, D. E. (2016). Social media and consumer engagement: A review and research agenda †. 10(2016), 268–287.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of
Personality and Social Psychology, 51, 1173-1182.
Bleier, A., & Eisenbeiss, M. (2015). The Importance of Trust for Personalized Online
Advertising. Journal of Retailing, 91(3), 390–409. https://doi.org/10.1016/j.jretai.2015.04.001 Boerman, S. C., Kruikemeier, S., & Zuiderveen Borgesius, F. J. (2017). Online Behavioral Advertising: A Literature Review and Research Agenda. Journal of Advertising, 46(3), 363–376. Bonabeau E (2004) “The Perils of the Imitation Age,” Harvard business review, 82(6), pp. 45–54. Bol, N., Dienlin, T., Kruikemeier, S., Sax, M., Boerman, S. C., Strycharz, J., Helberger, N., & de Vreese, C. H. (2018). Understanding the effects of personalization as a privacy calculus: Analyzing self-disclosure across health, news, and commerce contexts. Journal of
Computer-Mediated Communication, 23(6), 370–388. https://doi.org/10.1093/jcmc/zmy020
Breves, P. L., Liebers, N., Abt, M., & Kunze, A. (2019). The perceived fit between instagram influencers and the endorsed brand: How influencer–brand fit affects source credibility and persuasive effectiveness. Journal of Advertising Research, 59(4), 440–454.
