Jason S. O’Reilly
June 17, 2019
Blending In Or Uninvited Guest -
Master Thesis (MSc) - Marketing Intelligence
Blending In Or Uninvited Guest -
Intrusiveness and Effectiveness of Native Video Advertising on Facebook
Faculty of Economics & Business
Rijksuniversiteit Groningen
Student: Jason S. O’Reilly
Student Number: s3712133 Phone Number: +49 (0) 163 2580 543 Address: Ludwig-Gausebeck-Straße 7 48351 Everswinkel Germany Email: j.s.oreilly@student.rug.nl
Completion Date: June 17, 2019
Supervisor 1: prof. dr. T.H.A. Bijmolt
Supervisor 2: dr. A. Bhattacharya
Supervisor 3: dr. M. Vroegrijk
1 Management Summary:
For the over two billion monthly users on Facebook, the social media platform has changed daily lives - how we communicate with each other, how we work and how we entertain ourselves. Naturally, marketers have entered the platform as a way to engage with (potential) customers. One way in which marketers engage with customers is through native advertising. Simply put, native advertising is a technique used to make advertisements blend in with other content online.
The purpose of this graduation thesis is to investigate the relationship between video length used in native video advertising campaigns on Facebook and brand awareness and whether this relationship is, in fact, mediated by perceived intrusiveness and number of seconds watched. The thesis discusses previous findings regarding other advertising channels and perceived intrusiveness. These earlier findings have not been applied to the context of native video advertising extensively, thus, resulting in a clearly defined gap in literature. In order to reach this goal, a unique data set was employed that captures the responses of 5,157 participants across the United Kingdom, the Netherlands, and Germany, which was used to answer the following research question: “To what extent do perceived intrusiveness and number of seconds watched mediate the
relationship between length and brand awareness in native video advertising on Facebook?”. Each
respondent was shown two Facebook mock-up news feed with six posts, including a post done by a hypothetical friend from within the hypothetical respondent’s network and other liked pages. Additionally, a video was inserted in each feed to simulate native video advertising on Facebook. By using survey data, different insights compared to controlled experimental settings can be obtained by mirroring real-life situations to a larger extent.
2 The results suggest that video length on its own has a negative influence on brand awareness and increasing perceived intrusiveness while also having a positive influence on the number of seconds watched. Perceived intrusiveness itself is negatively affecting brand awareness while the number of seconds watched has a positive effect on brand awareness. Additionally, there are country differences and heterogeneity in socio-demographic groups and their level of brand awareness.
On the basis of these findings, it can be implied that the relationship between video length and brand awareness in the context of native video advertising on Facebook is mediated by perceived intrusiveness and number of seconds watched. Put differently, marketing managers need to be aware of both perceived intrusiveness and number of seconds watched when designing their native video advertising campaigns. While length in itself has a negative influence on brand awareness, an increased length is negatively influencing brand awareness through perceived intrusiveness while positively influencing brand awareness through number of seconds watched. Thus, marketing managers need to evaluate how they can design their length in order to engage users to spend a higher amount of time watching while being non-intrusiveness at the same time when optimizing brand awareness is to be achieved.
3 Preface:
Before you lies the thesis Blending In Or Uninvited Guest - Intrusiveness and Effectiveness of
Native Video Advertising on Facebook, which is the final product of my master programme. The
basis of this research is a survey conducted by DVJ-Insights across the United Kingdom, Germany, and the Netherlands.
Six months ago, I started this project as a marketing effectiveness study, but it turned out to be a journey into social media advertising, consumer research and native advertising. This thesis is the result of this long process. It cannot express the hours spent batting ideas around with my fellow scientists and friends, the hope for good results, the joy for data crunching, and the sadness with each failed attempt.
As the world moves further and at an ever-increasing pace into the digital age, new ways how marketers and consumers interact arise. I am drawn to new processes and ways to optimize existing processes. This thesis has helped me to dive deeper into native video advertising on Facebook and identify the main drivers of advertising performance.
4 Table of Contents: 1 Introduction 5 2 Theoretical Framework 8 3 Methodology 18 3.1 Data Source 19 3.2 Dependent Variable 21 3.3 Independent Variable 22 3.3.1 Mediators 22 3.3.2 Advertisement Characteristics 23 3.3.3 Control Variables 24 4 Results 25 4.1 Descriptive Results 25 4.2 Mediation Analysis 26
4.2.1 United Kingdom Results 28
4.2.2 Germany Results 28
4.2.3 Netherlands Results 29
4.2.4 Hypotheses Tested 29
5 Conclusion & Discussion 33
5 1 Introduction:
With the ongoing rise of social media and its mobile usage, spending on native advertising is forecasted to reach approximately $41 billion in the US by 2019. While there is no clear consensus regarding the definition of native advertising, Campbell & Marks (2015) have defined it as “a term
used to describe a spectrum of new online advertising forms that share a focus on minimizing disruption to a consumer’s online experience by appearing in-stream”. Put differently, the
advertisement is designed in such a way that it appears naturally on the platform. In the example of Facebook, such a sponsored post would be placed in a user’s news feed in the same way as posts by their network or any liked page. In more critical terms, this could also mean that users are deceived by disguising marketing messages to look like unpaid content. Globally, native advertising spending consequently amounted to 60% of total display spending in 2018 while traditional advertising is losing ground at an increasing pace (Benes, 2018).
Especially Facebook is seen as a valuable player that can compete with traditional advertising channels (Nelson-Field et al., 2013) by allowing advertisers to potentially reach one in seven people across the globe (Lin & Kim, 2016), with 86.3% of US marketers using Facebook for marketing purposes (Benes, 2018). While no exact statistics are known for European marketers, there seems to be reason to expect more and more firms to incorporate social media into their marketing communications. Further, the advent of social media has heavily influenced the way marketers and consumers can engage online, giving access to a wide range of new tools. To that end, social networking sites (SNS) such as Facebook have the potential to be(come) a valuable asset for marketers. Clearly, the trend is hinting at an increasing importance of native advertising and ways to measure their impact on a brand level and ultimately on driving sales.
However small the disruption for user experience may ultimately be, intrusive advertisements can lead to avoidance by users. Arguably, consumers express a higher degree of goal-orientation while browsing, thus, native advertisements may appear more intrusive than on other media platforms (Li et al., 2002).
6 For the scope of this research, native video advertisements, embedded in a mobile Facebook news feed, will be used. Based on their close link to traditional video advertisements, certain variables have been found to influence the relationship between advertisement characteristics, intrusiveness, and advertisement performance significantly (Goodrich et al., 2015; Campbell et al., 2017). With length being put forward as the most well-supported advertising characteristic influencing performance throughout.
Therefore, it is important to investigate the influence of native advertising characteristics on advertising performance measures, such as brand awareness (Keller, 1993), and whether this relationship is in fact mediated by intrusiveness. In practice, advertisers decide deliberately on the length of video advertisements. In reality, however, users can then skip the video whenever they please by scrolling further. Therefore, a distinction has to be made based on the actual amount watched by users. By doing so, the relationship between video length as designed by advertisers and consumed length as decided by users can be studied more extensively. Arguably, longer videos might be more successful in raising brand performance when consumed fully, however, longer videos could also be met with avoidance behaviour, rendering them less potent than short videos. As opposed to experimental settings, respondents were allowed to freely browse Facebook news feeds at their own pace, thus, not being forced to watch the total length. Ultimately, a second mediator capturing the exact time of the video that has been watched was used additionally.
Examining this relationship is not just a matter of academic interest, but also key to understand the mechanisms that can effectively improve user experience. Additionally, it is yet to be seen whether brands are trading on their consumers’ trust to gain reach when using native advertising techniques to disguise their messages, which could negatively affect reception in the long run.
Overall, it will be crucial for companies to assess the impact of native advertising campaigns on customer-based brand equity. Thus far, extensive experimental research has been carried out to examine the effect of advertisement characteristics on completion rates. However, the effect of intrusiveness on performance has not been the predominant centre of attention.
7 awareness in more depth. This very data set allows for a mostly unexplored method of dual mediation analysis by investigating the link between native advertising video length, perceived intrusiveness, watched amount of the video, and ultimately, brand awareness. Additionally, light will be shed on the cross-country effectiveness, which could prove useful to gain insights into the trade-off between standardization and differentiation in native advertising video campaigns. Therefore, answering the following research question:
To what extent do perceived intrusiveness and number of seconds watched mediate the relationship between length and brand awareness in native video advertising on Facebook?
This research has been carried out in cooperation with DVJ-Insights, a marketing research agency based in Utrecht, the Netherlands. One of their main lines of work is related to
advertising pre-testing. DVJ-Insights has kindly allowed the use of data that was gathered as part of their benchmarking procedure in the field of advertising pre-testing on social media for this thesis.
8 2 Theoretical Framework:
Extending prior research on intrusiveness and native advertising (e.g. Pikas & Sorrentino, 2014; Goodrich et al., 2015; Lee et al., 2016) by combining earlier findings and introducing two mediators, this thesis sets out to study the hypothesized effects of video length on brand
awareness, mediated by perceived intrusiveness and amount watched. A graphical representation of the model can be seen in figure 1 below.
Figure 1 - Conceptual Model
9 Social Networking Sites:
Social networking sites (SNS) can be defined as “web-based services that allow individuals to (1)
construct a public or semi-public profile within a bounded system, (2) articulate a list of other users with whom they share a connection, and (3) view and traverse their list of connections and those made by others within the system” (Boyd & Ellison, 2007, p. 211). Given how Facebook,
one of the largest social media platforms, works, native advertisements on this platform give advertisers the benefit to having users engage with the content directly via “liking” or “sharing” said content or brand pages itself. Additionally, the user’s network is able to see these actions in their own news feed, which gives them the opportunity to react, as well. Thus, Facebook may act as a credibility check of brands for its users due to the fact that every user can see reactions made by others (Dehghani & Tumer, 2015). Therefore, Facebook can be considered as an important medium for both customers and advertisers alike.
Further, Barreto (2013) identifies three main functions of SNS such as Facebook: Firstly, a personal function to allow users for self-expression. Secondly, a social function related to the sharing of content within one’s social network to enlarge their reach or strengthen existing ties. And lastly, SNS can be considered an “infomercial” medium, which allows brands to share content with users that have allowed them in their personal network, e.g. by liking their page. Marketing on SNS can give marketers the opportunity to engage in novel techniques, such as viral marketing, crowdsourcing, and using user-generated content (Cvijikj & Michahelles, 2013). However, SNS were not solely created for marketing purposes.
When looking at native advertisements, the “infomercial” function immediately lends itself as the most important one. However, it may interfere with the other functions when users are not actively looking for information during their visit by seeing native advertisements of brands not invited in their personal network. Therefore, it is important to critically assess the performance of preceding advertisement forms on SNS to understand whether native advertising can elevate user experience to the next level by minimizing disruption caused to user experience.
Banner Ads
10 insights generated in the field of banner advertising might therefore also be applicable for native advertising in general.
While banner displays have gained momentum, their impact on consumer engagement remains vague. Naturally, advertisements need to be noticed to impact users. Scholars have found contradicting results that point to banner blindness (Benway & Lane, 1998, Dréze & Hussherr, 2003, Burke et al., 2005) or towards high reception (Bayles, 2000, Hervert et al., 2011). However, most of these studies do lack comparability due to different methodologies employed. For example, including the goal-orientation of users influences their receptiveness towards advertisements positively (Barreto, 2013).
Moreover, click-through rates (CTR) of banner advertisements, in general, remain considerably low with less than 1/1000 due to avoidance behaviour displayed by users (Lee & Ahn, 2012). The authors argue that animated banners are less likely to engage users as they are actively bypassing attention-grabbing stimuli in favour of their initial goals while browsing. However, the authors have not conducted that experiment using SNS. Therefore, users may behave differently while browsing SNS compared to news pages.
Based on the experimental study by Barreto (2013), native advertising done by any brand on SNS platforms has a weak impact on the number of displays seen. Nonetheless, the study may suffer from utilizing a small sample size (n=20) which hinders generalizability dramatically. Ultimately, it is yet to be seen whether native video advertising can engage users in an improved way. Especially since native advertisements appear in-stream without users actively visiting a brand page on Facebook as part of the aforementioned experimental study. Thereby, native advertising can potentially combat avoidance behaviour found regarding banner advertising by not requiring additional search effort by users. However, it remains unclear whether these findings will actually hold true for native advertising on Facebook.
Given this undesirable outcome regarding banner effectiveness, more and more attention has been given to other types of advertising such as video advertisements as a form of engaging users, partly due to increased traffic related to video content provided on the Internet. Some of the shortcomings of these two types have been reduced or even overcome by taking it to the next level when going native.
11 With Internet video streaming and downloads amounting to 82% of all consumer traffic by 2022 (Cisco, 2019), videos and their monetization will be critical. Therefore, video providers have focused attention on pre-, mid-, and post-roll video advertisements with significantly different completion rates in favour of mid-rolls (Krishnan & Sitaraman, 2013). Apparently, video advertisements have diverging success rates when placed in different contexts. Overall, one could argue that video advertisements try to monetize on increased traffic due to abandoning banner displays at the expense of disrupting user experience with a forced advertisement roll. However, completion rates are also linked to characteristics of the initial video such as length, among other factors, in which the advertising roll is placed (Krishnan & Sitaraman, 2013). The authors have come to this insight by analysing a vast dataset consisting of 65 million unique viewers across 33 different video providers. By doing so, the authors clearly link advertisement and video characteristics to a performance measure in terms of completion rates. Their approach may benefit from examining the impact of advertisement characteristics on a wider range of performance measures when videos are used for going native on SNS.
Therefore, the environment in which the advertisement is placed seems to be important, hence, hinting at the potential for native advertising in general when utilizing techniques to blend in. Ultimately, insights generated from previous advertising channels may also be applied to native advertising.
Intrusiveness
12 not receptive to advertising on Facebook. Therefore, their findings regarding the perceived intrusiveness of an advertisement could be related to its performance.
Still, the authors were unable to quantify feelings and perceptions indicated in the survey, hence, leaving room for specifying perceived intrusiveness as a mediating factor in the investigation of advertisement performance based on advertisement characteristics. Moreover, the authors did not specifically distinguish between native advertising and other forms of online advertising. The general notion with respect to intrusiveness is supported by Teixeira et al. (2012), arguing that especially video advertising encounters strong opposition and avoidance. In this context, advertising avoidance can be defined as “all actions by media users that differentially reduce their exposure to ad content” (Speck and Elliott 1997, p. 61). Based on their findings, emotional appeals (joy/surprise) in video advertising can combat this behaviour expressed by users. This could implicate that native advertising’s ability to evoke certain emotions may influence user avoidance similarly.
Nonetheless, this notion has to be put into perspective when accounting for findings based on Hyman et al. (2017), indicating that only a minority (37%) of the respondents (n=896) correctly identified native advertisements as paid content, compared to 81% for regular advertisements. Based on their model, age and self-reported ability to distinguish paid/unpaid content are positively influencing the identification of paid content. Therefore, the authors argue that native advertising is more effective than other advertising channels in avoiding initial scepticism put towards regular adverts but comes at the price of potential perceived deception that may result in a backlash.
13 When applying this to the SNS environment with respect to brand awareness, similar results are expected.
Native Advertising
Consequently, Kim (2015) investigated the relationship between advertisement similarity in the context of native video advertising and consumer attitudes on video sharing platforms. Consumers express more positive feelings towards native advertisements due to their less interruptive nature. Furthermore, the similarity between the advertisement and the video where it is placed increase perceived relevance and generate more positive attitudes. Therefore, the similarity in the context of native video advertising proves to be a potent technique to engage users. Arguably, the strength of embedding native advertisements in-stream is therefore linked to perceived similarity. This is in line with Hegner et al. (2016) who have found marginally significant results in favour of contextual overlap between the video advertisements and where it is placed, positively influencing attitudes. With this in mind, one may argue that going native might be a similar form of contextual overlap that could benefit advertising performance measures. In conclusion, it is yet to been seen how native advertisements fare on this aspect.
14 between countries in more depth. Further, they did not test for mediation effects of intrusiveness or amount watched between advertisement characteristics and performance, hence, it is yet to be seen whether the results change once the model is re-specified in accordance with Goodrich & Schiller (2011) on online video advertising and intrusiveness as a mediating factor. Further, Goodrich et al. (2015) support the notion that increased perceived intrusiveness results in higher abandon rates. When combining this with aforementioned findings regarding the relationship between length and completion rates (Krishnan & Sitaraman, 2013), a second mediator (sequential mediation) can be added to their model, and hence, accounting for the potentially mediating influence of how much of the video has been consumed.
Country Effects:
Additionally, Möller & Eisend (2010) have investigated country effects based on culture when determining banner advertising campaigns. Although banner advertisements are inherently different from native advertising, it could still be taken as an indication that there might be antecedents of advertisement performance, irrespective of individual or advertisement characteristics. From a theoretical perspective, Hofstede’s multi-dimensional view of culture has been used to quantify cross-country differences in advertising effectiveness. However, the study suffers from generalizability flaws due to the survey only being administered in English in all 26 countries. Therefore, non-English speaking respondents have been neglected systematically.
Applying these insights to native advertising could mean that it may be crucial to investigate the perceived intrusiveness of native advertising, which is in turn related to their impact on obscuring other content. Arguably, the unique design that makes them appear natural in the SNS environment, is resulting in a minimized impact on other content. Therefore, it is still unclear whether insights generated on intrusiveness as observed in online video campaigns translates to native video campaigns on SNS. Interestingly, intrusive adverts have had lower recall rates than non-intrusive ads for online videos (Goodrich et al., 2015).
Performance Measures:
15 forward as insightful dimensions to assess the effectiveness of marketing instruments, rather than merely looking at the direct impact of advertising spending on sales. Therefore, by building on existing knowledge outlined by Pikas & Sorrentino (2014) in terms of quantifying user responses towards advertising on Facebook, these feelings can now be assessed more thoroughly. Additionally, using a broader definition of performance measures allows for deeper insights into the workings of native advertising campaigns that could potentially shed light on which aspect of brand awareness can be influenced the most via native video advertising on Facebook if any. Based on Keller (1993) and data availability, the performance of native video advertising will be assessed by examining aided recall and recognition (brand awareness).
Hypotheses:
When putting all of the aforementioned insights into perspective, it can be argued that the length of video advertising is clearly linked to performance measures in terms of brand awareness (Goodrich et al., 2015). Luarn et al. (2015) who argue that vividness of brand posts, appearing in a user’s feed after actively liking that brand page, helps engagement in terms of likes, comments, and shares further support this notion. Moreover, Newstead & Romaniuk (2010) found that longer television advertisements score higher likeability and recall rates although the television landscape might not be fully transferable to the SNS environment support this line of reasoning.
When addressing the lack of attention in light of native advertising, the relationship between advertisement length and brand awareness, mediated by intrusiveness and number seconds watched, can be assessed thoroughly. Consequently, the research question will be answered by engaging in the following hypotheses:
● H1a: The length of native video advertisements positively influences aided recall ● H1b: The length of native video advertisements positively influences brand recognition
16 whereas participants of this survey were allowed to scroll through the news feed at their own pace, thus, also potentially skipping advertisement immediately. Moreover, since respondents are easily aware of a video’s length, longer videos might be considered more intrusive since they are interfering with other content that could be consumed instead.
Similarly, longer videos might also be less interesting for users scrolling through their Facebook news feed when they were not actively looking for such information. Therefore, longer videos could result in a reduced amount of seconds watched when the video is skipped shortly after being encountered. The user behaviour could work similarly to avoidance behaviour displayed with television commercials where consumers are known to skip them whenever possible (Krugman, 1983). Arguably, longer videos will, therefore, make users more inclined to skip the video before it is finished, rather than completing it.
● H2: The length of native video advertisements positively influences perceived intrusiveness ● H3: The length of native video advertisements negatively influences the number of seconds
watched of native video advertising
Moreover, Goodrich et al. (2015) argue that intrusive advertisements have higher abandon rates. Therefore, users seem to avoid video advertising when it is perceived as intrusiveness.
● H4: Perceived Intrusiveness negatively influences the number of seconds watched of native
video advertising
Conversely, once a video has been consumed to a large extent, the user is unlikely to perceive it as being intrusive. The time watched can, therefore, be taken as an indication of whether respondents are interested or at least unbothered by the specific native video advertising.
● H5: The number of seconds watched negatively influences perceived intrusiveness of native
video advertising
17 have found that intrusive adverts that interrupt user experience are evaluated less favourable. Additionally, Truong & Simmons (2010) argue that this applies even stronger in the mobile environment since screen sizes are smaller, giving a greater perceived focus on intrusive advertising, and typical usage patterns with high degrees of goal-orientation. Moreover, McCoy et al. (2008) found that users respondent negatively to advertisements that have been forced upon them. In combination with the aforementioned literature, a clear case for the negative influence of perceived intrusiveness on advertising performance can be made (McCoy et al., 2008; Pikas & Sorrentino, 2014). With respect to the number of seconds watched, Newell & Henderson (1998) have found that aided recall and brand recognition have been higher when respondents in their experimental study viewed longer advertisements. While respondents were forced to watch the entire video, respondents of this survey were not. Thus, it can be argued that the seconds watched aids the recollection of information contained in the adverts.
18 Methodology:
Plan of Analysis:
For carrying out this research, several linear and nonlinear regressions were used to test for the hypothesized mediating effect of perceived intrusiveness and the number of seconds watched between video length and brand awareness, following Baron & Kenny (1986). As extensively outlined by the authors, four steps are needed to establish mediation and will be covered in the following chapter together with the results. Firstly, one needs to establish a relationship between the independent and dependent variable that can effectively be mediated (path c). Secondly, variation levels in the independent variable significantly account for variations in the mediator (path a and d). This step can be tested by treating the mediator as a regular outcome variable in an OLS regression model. Thirdly, the mediator should affect the dependent variable, thus, variation in the mediator(s) should account for variation in the dependent variable (path b and e). Fourthly, when controlling for path a, b, d, and e, the effect of the independent variable on the dependent variable, when controlling for both mediators, should be zero (path c’). The effects for step three and four can be established in the same model that uses the dependent variable as its outcome variable and the independent variable, as well as both mediators as predictors. If all four steps are met, the mediators completely mediate the relationship between the independent and dependent variable. However, when only the first three steps are met, partially mediation is indicated instead. Following Shrout & Bolger (2002), bootstrapping with a sample size of 5,000 will be used to acquire confidence intervals and significance levels while avoiding normality issues that have been found. Unfortunately, a decomposition of total effects into direct and indirect effects is foregone due to computational issues when using logit models as the final outcome models. Due to the fact that one cannot identify parameter estimates and error variance separately but only their ratio, decomposition is not advisable (Breen et al., 2013).
19 Moreover, consumer perception is at the core of this research, thus, making survey data the most suitable. Due to the cross-national setting, the survey has been translated into three languages, namely English, German, and Dutch. The survey was administered solely by DVJ-Insights.
All things considered, previously mentioned findings regarding country-effects can be taken as an indication that there might be potential differences between countries, hence, justifying the use of country models by utilizing country-specific advertisements and surveys (Möller & Eisend, 2010). This allows for further insights into the differences between a small set of countries that are arguably rather similar based on cultural dimensions (Hofstede, 1980). Therefore, each country was tested individually instead of testing them on an aggregated level.
As outlined in more detail in the following paragraphs, advertisement performance is split-up into brand awareness, consisting out of aided recall and recognition, while perceived intrusiveness is made up out of several items.
3.1 Data Source Survey:
In order to test this thesis’ hypotheses, a mobile-only survey was administered showing a total of two Facebook news feeds and three news website outlets to approximately a total of 5,224 respondents across three countries (Germany, the United Kingdom, and the Netherlands). The survey consisted of questions aimed at gaining insights into the respondent’s background in terms of demographics and in-context exposure to the digital advert, followed by a distraction task and ultimately questions related to their evaluation of the advertisement and its perceived intrusiveness. Initially, a short video clip was shown that guarantees that only respondents that can view and listen to the video are included in the final sample size by asking about content and music used in the video. DVJ Insights has kindly allowed the use of the data that was gathered for other purposes than this research.
20 By showing the advertisement embedded in a real mobile Facebook news feed, the respondent was able to scroll through it on his/her own pace, thus, reflecting the reality in which respondents are exposed to native advertising when using Facebook on their own devices. This is a key feature in determining the impact of the advertisement to grab the user’s attention and communicate the brand and message. Thereby, ecological validity is expected to be high by mirroring real-life situations as much as possible.
Mock Feed:
As shown in the appendix (figure 1-6), a total of six mock-up Facebook news feeds were used for this study. Each respondent saw two feeds and two news websites, corresponding to their country of origin. For the scope of this thesis, only data related to the Facebook news feeds were used, hence, ignoring data gathered regarding the news websites also shown to respondents. The feeds were put together by using six posts, excluding the native advertising, out of a pool of nine posts. The number of posts shown to each respondent was kept low in order to decrease fatigue despite potentially harming the realism of the feed by being limited in size. The posts that make up the feed are real posts done by pop cultural pages coming from comedy, music, entertainment, and lifestyle but also news/sports pages, and a post done by a hypothetical friend in the respondent’s network. The friend’s post was emulated by using a picture of a beach/animal (one feed per country and respondent is using the holiday post while the other is using the zoo post) for the post with a short caption. Additionally, the post done by a news site was changed to include a comment from a friend in the user’s network, thus, making the feed seem more realistic. Each of the posts from within the hypothetical network has been modified in such a way that they use a realistic name and profile picture to be as authentic as possible. The names have been slightly adapted to fit each country. For each news feed, the user interface language was adjusted to match each country.
21 “own” profile picture has been changed to the default picture, which is gender-neutral, thus, making the feeds less realistic but allowing to use the same feeds for all respondents.
The native advert is inserted as the third post in the feed in order to capture user engagement in terms of the amount of time the video has been “in-focus” (i.e. visible to the respondent). As soon as the video’s main part becomes visible, the video will play, thereby, a counter can be used to measure the time passed until the video is scrolled by. Moreover, the other six posts are inserted in a random order between respondents to ensure no order effects are interfering. For each post, the video used to simulate native video advertising on Facebook is chosen from a pool of over 30 videos per country randomly.
3.2 Dependent Variables
The dependent variables for the scope of this research are related to customer-based brand equity that can be defined as “the differential effect of brand knowledge on consumer response to the marketing of the brand” (Keller, 1993). The key focus here is on the awareness of the brand (recall and recognition) to assess the native advertising’s performance. This is supported by Huang & Sarigöllü (2014) who use also use customer mindset metrics to assess consumer brand attitudes via brand awareness and brand recognition.
Brand awareness:
Based on Keller (1993), brand awareness can be split into two concepts, namely brand recall and brand recognition. This is in line with Rossiter & Percy (1987) who argue that brand awareness is closely linked to the consumer’s ability to identify a brand. While brand recall is thus related to the likelihood of a brand being retrieved after exposure, brand recognition is about confirming exposure to an advertisement that has been consumed previously. When combining this with insights based on Petty & Cacioppo (1986) on consumer decision-making, brand awareness may be crucial in low involvement contexts. Arguably, browsing Facebook could be considered a low involvement task. According to Huang & Sarigöllü (2014), brand awareness is likely to influence market performance.
22 when given a list of multiple brands to choose from regarding their appearance (“Could you
indicate whether the brand appeared on the websites you browsed through?”). This is in line with
Unnava & Burnkrant (1991) and Goodrich et al. (2015). Moreover, brand recognition is equal to one when the respondent selects “yes, definitely” or “maybe” (3-point scale) to the question concerning the ability to indicate whether a shown advert has appeared while browsing (“Can you
indicate for this advert whether you saw it on one of the websites?”). The respondent was shown
multiple video advertisements and asked to indicate whether each video was part of the previously shown Facebook news feed. For both closed questions, all brands were displayed randomly to rule out order effects.
3.3 Independent Variables
In order to capture the indirect effect of native advertising video length on brand awareness through perceived intrusiveness and the number of seconds watched, both concepts need to be included in a dual mediation analysis.
3.3.1 Mediators
Perceived Intrusiveness:
As outlined in earlier stages, Goodrich & Schiller (2011), Pikas & Sorrentino (2014), and Lee et al. (2015) have made a clear case for the inclusion of perceived intrusiveness when assessing the performance measures of native advertising. In order to operationalize perceived intrusiveness, the existing survey has been modified to account for an additional seven-level Likert scale with the following five items: distracting, disturbing, forced, interfering, and intrusive. This scale has been put forward by previous research carried out by Li et al. (2002) and Goodrich & Schiller (2011). Previous research used two additional items, however, given the length of the survey and the possibility of fatigue, a slightly reduced scale has been employed. A factor analysis has been carried out to aggregate the responses on each scale to unify respondents’ perception into a one-factor variable that can be used as the mediating one-factor to test the aforementioned hypotheses.
23 of sphericity p < 0.00). Lastly, perceived intrusiveness in the Dutch sample can be described with an alpha of 0.84 (Kaiser-Meyer-Olkin factor adequacy = 0.84, Bartlett’s test of sphericity p < 0.00). Based on Harkness & Schoua-Glusberg (1998), a back-translation was used to translate the English items into German and Dutch. Two fluent colleagues have helped and translated the German/Dutch version back into English with no problems encountered. The resulting translation used was, therefore: ablenkend, störend, gezwungen, eingreifend, aufdringlich (German), and
afleidend, verstorend, gedwongen, ingrijpend, opdringerig (Dutch). Ideally, the translation would
be tested with bilingual respondents to assess the quality to check for equal responses despite the survey being administered in different languages. Unfortunately, this was not possible due to sampling issues and time constraints.
Number of Seconds Watched:
Due to the nature of the survey as administered by DVJ-Insights, the respondents are able to scroll through the Facebook news feed according to their own pace. Therefore, no respondent was forced to watch any of the native advertising videos. With this in mind, it would be difficult to examine the effect of length on aided recall and brand recognition without investigating how much of the video has actually been consumed.
Based on data availability, it is only possible to measure the seconds watched for the scope of this study. Thus, the potential relationship between the length of the video, perceived intrusiveness, seconds watched, and brand awareness can be assessed by adding a second mediator. Therefore, using the number of seconds watched also gives insights into how much respondents are willing to watch when knowing themselves how much would be left to watch. This model specification is mostly unexplored with respect to native video advertising on SNS.
3.3.2 Advertisement Characteristics Length:
24 3.3.3 Controls
25 Results:
4.1 Descriptive Results:
Table 1 presents the mean values of all variables used in the analyses going forward, as presented in the aforementioned section. A more detailed representation can be found in the appendix.
The analyses carried out on country-level were based on slightly different sample sizes due to listwise deletion based on the aforementioned hardware checks. For the UK sample, it can be seen that there were 36 observations with missing values for the number of seconds watched (Watched Length). This may arise when respondents could not answer the initial survey questions regarding sound and video quality, thus, they had to be excluded. Since each respondent is present twice, once for each news feed seen, there are 18 respondents in total that could not adhere to the aforementioned hardware requirements of sound and video quality. In total, 1750 respondents (n=3500/2) were used. Similarly, 32 observations (16 respondents) in the German and 56 observations in the Dutch sample had to be excluded due to similar issues, leaving 1735 respondents in the German sample and 1672 for the Dutch sample.
26 respectively. Further, only a minority of the total sample identifies as Other in terms of their gender. All samples are slightly skewed towards female respondents.
When looking at the video advertisements used for native video campaigns, it can be seen that the category split between FMCG (25%), Durables (25%), and Non-Durables (50%) is almost identical across the three samples. However, mean video length has been shorter in the UK sample (M = 19.38), compared to German and Dutch levels. Moreover, the average watched length is also highest in the UK sample (M = 3.92). Additionally, perceived intrusiveness levels were relatively consistent across all three countries, ranging from 11.84 to 12.09. These findings may be based on using videos that were published by the same or at least comparable brand.
Overall, aided recall and brand recognition are highest in the UK sample, with 44% (SD = 0.5) of respondents identifying a seen brand from a list of options while 61% (SD = 0.49) can confirm prior exposure. Aided recall and brand recognition are more similar between the German and Dutch sample than compared to the UK sample, thus, hinting at country differences in terms of brand awareness at first glance. However, since different video advertisements were being used for each country, differences in descriptive statistics could also potentially arise due to differences in the videos itself.
In terms of correlations (see appendix, table 2, 4, 6), there were no major violations with respect to multicollinearity identified despite significant correlations. All correlations are below 0.25 and especially for the logistic regressions models less of a concern (Tabachnick & Fidell, 2012). The mediators are correlated with control variables but this is still considerably low. Additionally, there needs to be a correlation between video length and the mediators in order to establish a causal mediation.
4.2 Mediation Analysis:
28 4.2.1 United Kingdom Results
When assessing mediation analysis, it is important to first establish a relationship between the predictor and outcome variable (Baron & Kenny, 1986). For each of the two outcome variables, namely aided recall and recognition, a logistic regression model was used, respectively. It was found that video length is negatively influencing aided recall (b = -0.0086, SE = 0.002, p < 0.01, OR = .9914) and brand recognition (b = -0.0039, SE = 0.002, p < 0.05, OR = 0.9961).
In the first ordinary least squares (OLS) regression model, video length was significantly related to higher perceived intrusiveness scores, b = 0.02, SE = 0.004, p < 0.01. Every additional second of video length increases perceived intrusiveness therefore by 0.02. The second OLS regression model tested the relationship between video length and watched length. It was found that video length positively influences time watched, b = 0.03, SE = 0.004, p < 0.01. Similarly, each additional second of video length increases time watched by 0.03 seconds. Therefore, the second requirement for mediation following Baron & Kenny (1986) is met by establishing a relationship between the predictor variable (video length) and each mediator.
Lastly, the two logistic regression models from the first step were repeated but modified accordingly by incorporating both mediators simultaneously. With these additional predictors, the impact of video length on aided recall (b = -0.01, SE = 0.002, p < 0.01, OR = 0.9898) and brand recognition (b = -0.01, SE = 0.002, p < 0.01, OR = 0.9955) has increased. The odds ratios therefore imply that it becomes less likely that respondents have positive outcomes in terms of brand awareness should video length increase. For perceived intrusiveness, its impact on aided recall and brand recognition was found to be negative (b = -0.03, SE = 0.006, p < 0.01, OR = 0.97). For watched length, its impact on aided recall was positive (b = 0.09, SE = 0.008, p < 0.01, OR = 1.09) while being even stronger on brand recognition (b = 0.17, SE = 0.013, p < 0.01, OR = 1.19). Thus, an additional second of watched length results in an increased likelihood of 1.19 times of observing a favourable brand recognition. Moreover, only age has been found to be a significant control. Age is negatively influencing aided recall (b = -0.01, SE = 0.003, p < 0.01, OR = 0.99) and brand recognition (b = -0.014, SE = 0.003, p < 0.01, OR = 0.99).
4.2.2 Germany Results
29 video length and aided recall that was statistically different from zero (b = -0.0017, SE = 0.001, p = 0.14, OR = 0.996). Therefore, only the model outcomes for brand recognition will be considered going forward.
Video length is affecting both mediators equally (b = 0.01, SE = 0.003, p < 0.01). With longer videos, respondents consider the video more intrusive while also watching a longer amount. For the final logistic regression model, video length is negatively influencing brand recognition (b = -0.003, SE = 0.001, p < 0.01, OR = 0.997) when both mediators are present in the model. Perceived intrusiveness was found to negatively influence brand recognition (b = -0.04, SE = 0.008, p < 0.01, OR = 0.96) while watched length positively influences brand recognition (b = 0.28, SE = 0.016, p < 0.01, OR = 1.32). For control variables, age was found to negatively influence brand recognition (b = -0.01, SE = 0.003, p < 0.01, OR = 0.99). Moreover, male respondents have significantly lower brand recognition scores than females (b = -0.15, SE = 0.075, p < 0.05, OR = 0.86).
4.2.3 Netherlands Results
Similarly, the relationships between video length and aided recall/brand recognition were tested, respectively. It was found that video length is only significantly influencing aided recall (b = -0.03, SE = 0.001, p < 0.01, OR = 0.97). Hence, brand recognition will be ignored due to parameter estimates not statistically different from zero.
As in the other two countries, video length is positively influencing perceived intrusiveness (b = 0.01, SE = 0.02, p < 0.01) and watched length (b = 0.03, SE = 0.04, p < 0.01).
In terms of the final logistic regression model, video length is found to negatively influence aided recall (b = -0.05, SE = 0.001, p < 0.01, OR = 0.95). Considering the mediators, perceived intrusiveness is found to negatively influence aided recall (b = -0.04, SE = 0.008, p < 0.01, OR = 0.96) while watched length is positively influencing aided recall (b = 0.11, SE = 0.01, p < 0.01, OR = 1.12). Only age has been identified as a significant control variable, which negatively influences aided recall (b = -0.02, SE = 0.003, p < 0.01, OR = 0.98).
4.2.4 Hypotheses Tested
30 each of the other two samples, the relationship between mediators and brand awareness is consistent across all samples. Moreover, the significant findings between length and brand awareness are running in the reverse direction. Therefore, hypotheses 1a and 1b cannot be confirmed. Overall, the results can nevertheless be taken as an indication that the length of native video advertisements is affecting brand awareness. Put differently, each additional second in length reduces aides recall and brand recognition. The findings contradict prior research carried out in experimental settings with regard to the effect of video length on brand awareness (e.g. Goodrich et al., 2015). Considering the country differences, one could argue that the highest number of seconds watched in the UK sample could also explain parts of the phenomenon. However, the number of seconds watched is relatively close in the German and Dutch sample despite opposite significance findings with respect to the relationship between length and aided recall or brand recognition. Moreover, a brand’s country of origin might be influencing the receptiveness for attention-grabbing stimuli that are exerted on users online. Most brands used for the scope of this study are from the USA or UK. Therefore, respondents from outside the UK might be drawn less to each video, rendering the relationship between length and brand awareness insignificant in those samples. Nonetheless, the use of an additional dummy variable capturing whether brands are local does not consistently capture significant effects that are different from zero, nor changes earlier findings.
Moreover, hypothesis 2 is confirmed across all countries, thus, hinting at a different relationship between video length and perceived intrusiveness than found by Goodrich et al. (2015). Apparently, respondents perceive longer native video advertising as more intrusive when being able to skip it in more realistic environments, compared to control experimental settings. However, video length has positively influenced watched length consistently across all three countries, hence, running in the inverse direction (hypothesis 3). Apparently, longer videos are left in-scope longer, compared to shorter videos.
31 at watching only content that is interesting and therefore enjoyable which reduces perceived intrusiveness.
Lastly, perceived intrusiveness and watched length were both influencing aided recall and brand recognition as hypothesized in all three samples. Overall, the aforementioned findings regarding the relationship between perceived intrusiveness and brand awareness can be confirmed. In terms of watched length, a priori expectations have been met while there are no comparable studies available as of yet.
Overall, four out of six relationships between length and brand awareness have been mediated. In all cases but brand recognition in the German sample, the direct effect is larger than the total effect. These findings point towards inconsistent mediation due to the fact that the sum of the multiplied paths (ab, de, afe, dgb, see figure 1) are opposite in sign to c’ (MacKinnon et al., 2007). Contrary to Baron & Kenny (1986), Zhao et al. (2010) argue that a significant direct effect should not be taken as an ultimate requirement. Instead, finding only indirect mediation as in the German and Dutch sample can still provide some insights despite length not significantly influencing aided recall or brand recognition with respect to the influence of both mediators on brand awareness.
32 needed to make users stick in order to benefit brand awareness. By finding the right technique to engage users, the other path is benefitting brand awareness simultaneously. Ultimately, the findings suggest that native advertising video campaigns perform worse in terms of brand awareness with increasing length, despite positively influencing watched length. Consequently, shorter videos may be a more risk-averse technique to acquire positive effects on brand awareness through lower perceived intrusiveness that could then, in turn, result in an increased watched length and increased brand awareness overall.
33 5. Conclusion & Discussion:
While marketing and advertising managers are facing major changes when it comes to advertising in the digital age that forces them to re-evaluate existing strategies and techniques, scholars are looking for adapted ways to measure and explain marketing effectiveness in the digital age. The primary research aim was to investigate whether perceived intrusiveness and number of seconds watched mediate the relationship between video length and brand awareness in the context of native video advertising on Facebook. Although earlier studies have incorporated perceived intrusiveness as a mediating factor, no prior research has used survey data in combination with Facebook news feeds allowing for investigating the role of number of seconds watched to generate new insights outside of controlled experimental settings. This study fills the gap by using a more realistic setting that allows for examining the role of the number of seconds watched as a second potential mediator.
5.1 Concluding Findings
Overall, native video advertising on Facebook has considerable potential for marketers to engage with customers. The findings presented in this thesis clearly show that perceived intrusiveness and number of seconds watched are not only affecting aided recall and brand recognition but also mediate the relationship between length and both brand awareness measures. Arguably, these findings can be taken as support for Lee et al. (2016) by showing that perceived intrusiveness is related to advertisement performance in the context of native advertising, thereby, also able to quantify the impact of perceived intrusiveness on brand awareness, as suggested by Pikas & Sorrentino (2014). Number of seconds watched has been shown to be a crucial factor in the assessment of brand awareness and especially in determining the relationship between video length and brand awareness. Despite not being the predominant centre of attention in previous research, its influence on brand awareness can be taken as a first indication of importance when mirroring real-life environments on Facebook in advertising effectiveness studies.
34 those environments. Additionally, the results give support for Krishnan & Sitaraman (2013) in showing that video characteristics in terms of length are important for performance measures, even outside of completion rates but also brand awareness in specific. Moreover, age and gender differences have been found across the three samples. The findings reported in this paper indicate that there are both similarities and differences between the countries. The most pronounced difference between the countries are results related to the total effect model between length and brand awareness with the mediation only being significant in both cases for the UK sample compared to only one in the German (recognition) and Dutch (aided recall) sample.
5.2 Recommendation
35 advertising campaigns in the United Kingdom, Germany, and the Netherlands with regard to video length and its impact on brand awareness to a certain extent. Thus, Facebook users from all three countries respond similarly to video length in terms of aided recall and brand recognition. However, the country differences that have been found may warrant a case-to-case evaluation of video length as designed by marketers.
5.3 Limitations
36 5.4 Further Research
Despite these limitations, this study can still be taken as an insightful study on the effectiveness of native video advertising campaigns. This, in turn, can result in interesting avenues for further research. Arguably, the cross-country setting can be extended by including even more countries to then incorporate culture as an explanatory factor or investigating the relationship between standardization and localization with respect to related costs in more depth. Moreover, there is potential to improve the chosen videos for native advertising campaigns by using an even greater sample size to draw from. Thereby, more brands can be included to examine a potential difference between global and domestic brands more closely. By doing so, scholars and marketing managers may be able to generate more insights into the trade-off between standardization and localization. Since this study was only focusing on Facebook in terms of SNS, different platforms could be incorporated, as well. Especially due to the fact that different platforms generally attract different target groups. Taking Instagram as an example, a study into the effectiveness of native video advertising could be carried out with a usually younger target group on this platform. Linking this back to the negative effect of age, marketing managers could potentially find a more favourable environment on that platform for their native video advertising campaigns. Logically, designing follow-up research could then ensure that questions are tailored to the research at hand, thus, featuring more questions aimed at usage behaviour, editorial content, and motives, as outlined earlier. This opens up ways to investigate further techniques to make the transition for users from “normal” content to native advertising on Facebook as seamless as possible. Overall, more research is needed to investigate which length design optimizes brand awareness via watched length and perceived intrusiveness simultaneously. When reflecting on the ambiguous finding regarding the two paths, future research could examine the optimal length in more detail. By doing so, however, the model could incorporate length as a squared term but the non-linear nature of brand awareness may complicate matters further. Especially since both mediators are affecting each other. Nonetheless, one could specify a model without this relationship as a starting point for further research.
38 Bibliography:
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(6), 1173.
Barreto, A. (2013). Do users look at banner ads on Facebook?. Journal of Research in
Interactive Marketing, 7(2), 119-139.
Baumann, C., Hamin, H., & Chong, A. (2015). The role of brand exposure and experience on brand recall—Product durables vis-à-vis FMCG. Journal of Retailing and Consumer Services, 23, 21-31.
Bayles, M. (2000), “Just how ‘blind’ are we to advertising banners on the web?”, Usability News, Vol. 2 No. 2, pp. 520-541.
Benes, R. (2018, December 12). Are Marketers Leveraging Facebook? - eMarketer Trends, Forecasts & Statistics. Retrieved April 02, 2019, from https://www.emarketer.com/content/how-many-marketers-in-the-us-use-facebook
Benway, J.P. and Lane, D.M. (1998), “Banner blindness: web searchers often miss “obvious” links”, Internetworking: Internet Technical Group Newsletter, Vol. 1 No. 3, available at: www.internettg.org/newsletter/dec98/banner_blindness.html
Boyd, D. M., & Ellison, N. (2007). Social network sites: Definition, history, and scholarship. Journal of Computer-Mediated Communication, 13, 210−230.
Breen, R., Karlson, K. B., & Holm, A. (2013). Total, direct, and indirect effects in logit and probit models. Sociological Methods & Research, 42(2), 164-191.
Burke, M., Hornof, A., Nilsen, E. and Gorman, N. (2005), “High-cost banner blindness: ads increase perceived workload, hinder visual search, and are forgotten”, ACM Transactions on Computer-Human Interaction (TOCHI), Vol. 12 No. 4, pp. 423-445.
Campbell, C., & Marks, L. J. (2015). Good native advertising isn’ta secret. Business Horizons,
58(6), 599-606.
39 Cisco Visual Networking Index: Forecast and Trends, 2017–2022 White Paper. (2019, February 27). Retrieved April 02, 2019, from https://www.cisco.com/c/en/us/solutions/collateral/service-provider/visual-networking-index-vni/white-paper-c11-741490.html#_Toc532256805
Cvijikj, I. P., & Michahelles, F. (2013). Online engagement factors on Facebook brand pages. Social Network Analysis and Mining, 3(4), 843-861.
Dehghani, M., & Tumer, M. (2015). A research on effectiveness of Facebook advertising on enhancing purchase intention of consumers. Computers in Human Behavior, 49, 597-600.
Dréze, X. and Hussherr, F. (2003), “Internet advertising: is anybody watching?”, Journal of Interactive Marketing, Vol. 17 No. 4, pp. 8-23.
Duffett, R. G. (2015). Facebook advertising’s influence on intention-to-purchase and purchase amongst Millennials. Internet Research, 25(4), 498-526.
Goodrich, K., Schiller, S., & Galletta, D. (2011). Intrusiveness of online video advertising and its effects on marketing outcomes.
Goodrich, K., Schiller, S. Z., & Galletta, D. (2015). Consumer reactions to intrusiveness of online-video advertisements: do length, informativeness, and humor help (or hinder) marketing outcomes?. Journal of Advertising Research, 55(1), 37-50.
Harkness, J., & Schoua-Glusberg, A. (1998). Questionnaires in translation.
Hegner, S. M., Kusse, D. C., & Pruyn, A. T. (2016). Watch it! The influence of forced pre-roll video ads on consumer perceptions. In Advances in Advertising Research (Vol. VI)(pp. 63-73). Springer Gabler, Wiesbaden.
Hervert, G., Guérard, K., Tremblay, S. and Chtourou, M.S. (2011), “Is banner blindness
genuine? Eye tracking internet text advertising”, Applied Cognitive Psychology, Vol. 25 No. 5, pp. 708-716.
Hofstede, G., Culture's Consequences: International Differences in Work-Related Values, Sage Publications, 1980.
40 Hyman, D. A., Franklyn, D., Yee, C., & Rahmati, M. (2017). Going native: can consumers recognize native advertising: Does it matter. Yale JL & Tech., 19, 77.
Keller, K. L. (1993). Conceptualizing, measuring, and managing customer-based brand equity.
Journal of marketing, 57(1), 1-22.
Kim, S. (2015). Effects of ad-video similarity, ad location, and user control option on ad avoidance and advertiser-intended outcomes of online video ads.
Krishnan, S. S., & Sitaraman, R. K. (2013, October). Understanding the effectiveness of video ads: a measurement study. In Proceedings of the 2013 conference on Internet measurement
conference (pp. 149-162). ACM.
Krugman, H. (1983). Television program interest and commercial interruption. Journal of Advertising research, 23 (1), 21-23
Lee, J., & Ahn, J. H. (2012). Attention to banner ads and their effectiveness: An eye-tracking approach. International Journal of Electronic Commerce, 17(1), 119-137.
Lee, J., Kim, S., & Ham, C. D. (2016). A double-edged sword? Predicting consumers’ attitudes toward and sharing intention of native advertising on social media. American Behavioral
Scientist, 60(12), 1425-1441.
Li, Hairong, Edwards, Steven M., and Lee, Joo-Huyn (2002) “Measuring the intrusiveness of advertisements: Scale development and validation,” Journal of Advertising, 31:2, 37-47.
Lin, C. A., & Kim, T. (2016). Predicting user response to sponsored advertising on social media via the technology acceptance model. Computers in Human Behavior, 64, 710-718.
Luarn, P., Lin, Y. F., & Chiu, Y. P. (2015). Influence of Facebook brand-page posts on online engagement. Online Information Review, 39(4), 505-519
MacKinnon, D. P., Fairchild, A. J., & Fritz, M. S. (2007). Mediation analysis. Annual Review of Psychology, 58, 593-614.
McCoy, S., Everard, A., Polak, P., & Galletta, D. F. (2008). An experimental study of antecedents and consequences of online ad intrusiveness. Intl. Journal of Human–Computer
41 Möller, J., & Eisend, M. (2010). A global investigation into the cultural and individual
antecedents of banner advertising effectiveness. Journal of International Marketing, 18(2), 80-98.
Nelson-Field, K., Riebe, E., & Sharp, B. (2013). More mutter about clutter: Extending empirical generalizations to Facebook. Journal of Advertising Research, 53(2), 186-191.
Newell, S. J., & Henderson, K. V. (1998). Super Bowl advertising: Field testing the importance of advertisement frequency, length and placement on recall. Journal of Marketing
Communications, 4(4), 237-248.
Newstead, K., & Romaniuk, J. (2010). Cost per second: The relative effectiveness of 15-and 30-second television advertisements. Journal of Advertising Research, 50(1), 68-76.
Okazaki, S. (2007). Exploring gender effects in a mobile advertising context: on the evaluation of trust, attitudes, and recall. Sex Roles, 57(11-12), 897-908.
Pikas, B., & Sorrentino, G. (2014). The effectiveness of online advertising: Consumer’s perceptions of ads on Facebook, Twitter and YouTube. Journal of Applied Business and
Economics, 16(4), 70-81.
Rossiter, John R. and Larry Percy (1987), Advertising and Promotion Management. New York: McGraw-Hill Book Company.
Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies: New procedures and recommendations. Psychological Methods, 7, 422-445.
Speck, Paul S., and Michael T. Elliott (1997), “Predictors of Advertising Avoidance in Print and Broadcast Media,” Journal of Advertising, 26 (3), 61-76.
Tabachnick, B. G. & Fidell, L. S. (2012). Using multivariate statistics (6th ed.). Boston, MA: Pearson
Teixeira, T., Wedel, M., & Pieters, R. (2012). Emotion-induced engagement in internet video advertisements. Journal of Marketing Research, 49(2), 144-159.
42 Unnava, H. R., And R. E. Burnkrant. "An Imagery-Processing View of the Role of Pictures in Print Advertisements." Journal of Marketing Research 28 (1991): 226-231.
Wang, J., & Calder, B. J. (2006). Media transportation and advertising. Journal of consumer research, 33(2), 151-162.
43 Appendix:
49 Advertisement Brands Table 7:
DE NL UK
1&1 Adidas KFC
Adidas Albert Heijn PlayStation
Audi Audi Spotify
BILD Burger King Monster Energy
Bosch Chanel Microsoft
Burger King Chocomel Red Bull
Calvin Klein Coca Cola Audi
Chanel De Volkskrant McDonalds
Coca Cola DHL Netflix
Deutsche Bahn Durex Kitkat
Deutsche Telekom Fortnite H&M
DHL H&M Adidas
Edeka IKEA Fortnite
Fortnite Jumbo Samsung
H&M Kitkat Mercedes-Benz
IKEA KLM IKEA
KFC KPN Visa
Kinder Kruidvat Slack
Kitkat Mastercard Victoria's Secret
KLM McDonalds KLM
Mercedes-Benz Mercedes-Benz Chanel
Microsoft Microsoft DHL
Morphe Morphe Durex
Netflix Netflix Morphe
PlayStation ns Sainsbury's
Red Bull Philips Tesco
REWE PlayStation BT
Rolex Red Bull Barbour
Rossmann Rolex Virgin Train
Samsung Samsung EE Broadband
Slack Slack TalkTalk
Spotify Spotify The Sun
Unitymedia Telfort Cadbury Dairy Milk Victoria's Secret Victoria's Secret Rolex
