• No results found

Time is ticking, but is anyone watching? : the influence of advertisement length and thematic (in-)congruity in online video environmentsDrost

N/A
N/A
Protected

Academic year: 2021

Share "Time is ticking, but is anyone watching? : the influence of advertisement length and thematic (in-)congruity in online video environmentsDrost"

Copied!
81
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

Time is ticking, but is anyone watching?

The influence of advertisement length and

thematic (in-)congruity in online video

environments

By

Raphaëla Drost

Student ID. 6277292 /10000120 Student card no. 0020058564 University. University of Amsterdam

Master’s thesis. Master’s Programme Communication Science

Track. Graduate School of Communication Science, Persuasive communication Supervisor. mw. dr. H.A.M. Voorveld

(2)

Abstract

This study investigated the effects of ad length (3-seconds versus 5-seconds versus 15-seconds) and thematic congruity (congruent versus moderately incongruent) between the ad and the ensuing video, on brand awareness and brand attitude in an online video

environment. Findings from the self-administered online experiment showed that

advertisement length did have an effect on brand recognition. The 15-second advertisement led to a higher recognition of Coca-Cola than the 3-second advertisement. However, no differences were found between the other conditions on brand recognition. Likewise, the three advertisement lengths did not generate a different response in brand recall.

The results further demonstrated that ad length did not positively or negatively affect viewers’ brand attitude. Additionally, the level of congruity between the advertisement and ensuing video did not moderate the effect of advertisement length on brand awareness and/or attitude. As ad length and level of congruity did not influence viewers’ brand recall or brand attitude, it could be argued that it would be more profitable for marketers to merely produce short video advertisements of 3- or 5-seconds instead of 15-second or longer advertisements. Longer ads which are often more expensive, are not necessarily better than shorter ads when the advertisement objectives are to increase brand recall or brand attitude.

Keywords: Advertisement length, Thematic congruity, Online video advertisements, Brand awareness, Brand attitude.

(3)

Time is ticking, but is anyone watching? The influence of advertisement length and thematic (in-)congruity in online video environments

Online video-platform YouTube offers its users a wide spectrum of different types of videos, ranging from DIY amateur videos to professional music videos and advertisements, uploaded by both businesses and individuals. It might not come as a surprise that the platform is immensely popular among internet-users. The video community has over one billion users worldwide, which makes it the most popular online video platform globally ("Statistics," n.d.; Pashkevich, Dorai-Raj, Kellar, & Zigmond, 2012). For example, in the Netherlands with a population of almost 17 million inhabitants, 6,8 million people use the online video website, of which 1,2 million are active on a daily basis (Jassies, 2015).

YouTube and its capacity to reach millions of potential customers, offers an

interesting online platform for marketers. YouTube’s business model provides marketers with the opportunity to advertise via in-stream ads, which are advertisement clips shown prior to the user’s selected video. However, the question arises what type of advertising activities can satisfy both marketers and YouTube users, as research by Pikas and Sorrentino (2014)

indicated that 81% of YouTube users were actually annoyed by online video advertisements. With the rise of online advertisements and thereby annoyance among viewers,

commercial efforts could have negative outcomes for both advertiser and user. For example, viewers’ irritation towards the advertisements can result in negative attitudes towards both YouTube and the advertised brand and reduce purchase intentions due to feelings of

intrusiveness by the advertisement (Goodrich, Schiller, & Galletta, 2015). As a consequence, consumers respond adversely to the persuasive messages by installing ad blockers or

switching tabs during the commercial, which makes it more difficult for advertisers and YouTube to successfully reach their target audience. Therefore the challenge is to find a socially accepted level of advertising for both YouTube users and advertisers.

(4)

To lessen irritation and avoidance behaviour among consumers on forced lengthy ads, YouTube introduced 5-second skippable advertisements. Additionally, with the introduction of TrueView, advertisers only pay for the ad when a YouTube user decides to watch the entire advertisement or at least for 30-seconds of it ("TrueView," 2011). The new shorter 5-second format has lower abandonment rates than longer advertisements and gives consumers a greater feeling of empowerment by being able to decide whether or not to watch the

commercial. However, research has shown that 88% of YouTube users always use the skip function and decide not to watch the advertisement, indicating there is still need for an effective online video ad strategy ("Non-skippable in-stream," n.d.; Pikas & Sorrentino, 2014).

Therefore, this study examines whether a short new 3-second non-skippable

advertisement snippet can provide an equal or improved awareness level as a longer 5- or 15-second non-skippable advertisement clip. Additionally it is examined whether shorter

advertisements can create a more positive brand attitude than longer advertisements.

Furthermore, previous research has focused on the effects of advertisement length on either brand awareness and/or brand attitude, but did not specifically distinguish whether the advertisement was congruent with the video it preceded (e.g. Goodrich et al., 2015).

The effect of short versus long advertisements (ad length) on brand awareness and attitude might differ depending on the level of congruity between the advertisement and the ensuing video. For example, Russell (2002) researched the effect of product placement congruity and the plot connection of a television show, and showed that congruent products elicit more positive brand attitudes than incongruent products. Other research by Dahlén, Rosengren, Törn and Öhman (2008) found that thematic incongruence led to more positive advertisement awareness. Therefore, (in-)congruity could influence brand awareness and brand attitude differently depending on media context and advertisement characteristics such

(5)

as advertisement length.

To find the right balance between effective advertising and maximum user experience, this thesis researches the effects of video advertisement length on brand awareness and brand attitude. Furthermore it is examined whether the ads’ (in-)congruity with the ensuing video has a moderating effect on YouTube viewers’ brand awareness and attitude. Therefore the following research question is proposed:

RQ. What is the effect of forced exposure to short 3-second advertisements in comparison to longer 5-second and 15-second advertisements on brand awareness and brand attitude in an online video environment? And what is the role of congruity between the advertisement and ensuing video?

Theoretical framework: Theoretical and empirical evidence Video advertisement formats on YouTube

YouTube currently offers several video advertising formats, including skippable and non-skippable in-stream advertisements ("YouTube advertising formats," n.d.).

Non-skippable ads are advertisement clips shown prior, during or after the user’s selected video. The user must watch the ad before the video continues and the ad can be up to 30-seconds long. Skippable advertisements are based on the TrueView principle, where viewers have the choice to watch the entire advertisement or to skip the ad after 5-seconds. In this case the advertiser only pays when a consumer decides to watch the entire commercial or a minimum of 30-seconds ("TrueView," 2011).

Although non-skippable in-stream advertisements have greater abandonment rates than skippable in-stream ads, they do generate greater revenue ("Non-skippable in-stream," n.d.). Skippable advertisements have a lower abandonment rate, but various studies have

(6)

shown that the majority of viewers choose to skip the advertisement nonetheless (Dehghani, Niaki, Ramezani, & Sali, 2016; Saputra & Fachira, 2014). Dehghani, Niaki, Ramezani and Sali (2016) concluded that more than 73% often skip the in-stream video advertisements on YouTube, while Pikas and Sorrentino (2014) found that 88% of the respondents always skip commercials on YouTube.

These numbers indicate that most consumers are willing to expose themselves to an advertisement only for a maximum of 5-seconds. Research by Saputra and Fachira (2014) confirms this line of thought, as their study showed that the top reason for skipping advertisements was the length of an ad (46%). As a consequence of viewers’ skipping behaviour, few consumers are actually exposed to the complete advertisement. Thus the first five seconds that cannot be skipped provides a timeframe opportunity for advertisers to get their advertisement message across (Saputra & Fachira, 2014).

Advertisement length

As YouTube and other media offer various advertisement length options to

advertisers, the relationships between the ad length and brand awareness and/or brand attitude have been widely researched in various media such as in television, on mobile phones and on the Internet (e.g. Newell & Henderson, 1998; Coursaris, Sung, & Swierenga, 2010;

Goldstein, McAfee, & Suri, 2015). However, most research either merely focussed on the effects of ad length on brand awareness or on brand attitude (e.g. Wang, Shih, & Peracchio, 2013; Goldstein et al., 2015). Yet, the effects on both consumer responses of brand awareness and brand attitude to an ad can differ per advertisement length (Li & Lo, 2015). For example, Goodrich, Schiller and Galletta (2015) did compare the impact of ad length on both brand awareness and attitude of online video advertisements and showed that longer ads generate a higher brand recall and a more positive brand attitude than shorter ads.

(7)

Effect of ad length on brand awareness. Brand awareness is a consumer’s familiarity with a brand and a consumer’s ability to identify or memorize a brand under specific conditions. In this research, brand awareness consists of both brand recall and brand recognition (Barreda, Bilgihan, Nusair, & Okumus, 2015). With brand recall a consumer recalls a brand from memory after given a cue of, for example, the product category. This can be aided by, for example, having a consumer choose from multiple-choice options or unaided by asking an open question. When a consumer recognizes a brand due to previous exposure to it, this is referred to as brand recognition. The greater a consumer’s brand awareness, the more salient the brand becomes for the customer during purchasing choices (Percy & Rossiter, 1992).

Previous research shows that advertisement length is positively related to brand awareness. Goodrich et al. (2015) found that longer ads in online video environments increased brand recall and Li and Lo’s (2015) laboratory study showed that long online YouTube ads enhanced brand recognition. Research in other media such as in television also confirms this line of thought (e.g. Jeong, 2015; Singh & Cole, 1993; Newell & Henderson, 1998). These findings can be explained via the Elaboration Likelihood Model (ELM) on people’s processing and learning abilities.

The Elaboration Likelihood Model by Petty and Cacioppo (1984) explains two routes of persuasion to change attitudes: the central and peripheral route. When a consumer puts a high cognitive effort into processing a message, the message is processed deeply and the central route is taken. On the other hand, when the consumer puts little cognitive effort into the processing of a message, the peripheral route is taken. In this case, peripheral cues such as colour or sound play a role in people’s judgments (Kitchen, Kerr, Schultz, McColl, & Pals, 2014). Several factors such as motivation, but also the ability to process the information, influences which persuasion route is taken. Although the ELM is normally used to describe

(8)

changing attitudes, it can also be used to explain a change in a consumer’s brand awareness (Newell & Henderson, 1998).

Advertisement length influences how much processing time consumers have to elaborate on the advertisement message. Research has shown that if consumers do not have enough time to process information, the ability to cognitively elaborate on the message decreases and thus the peripheral route is taken. An opposite effect has been found for longer processing time (Lee & Schumann, 2004). Thus, the longer the advertisement, the more time viewers are able to process the given information, which in turn leads to a greater brand awareness (Newell & Henderson, 1998).

Additionally, advertisers are able to repeat information more often in a longer commercial. The repetition of the key message also enables the consumer to process the advertisement better. As a consequence, consumers become more familiar with the brand and the given information and so improves brand awareness (Newell & Henderson, 1998).

Research by Newell and Henderson (1998) confirmed this by showing that 30-second advertisements were more effective in generating brand awareness than 15-second

commercials. Theory and empirical findings therefore suggest that advertisement length and brand awareness are positively related.

Effect of ad length on brand attitude. However, the effect of advertisement length might be different for brand attitude. Brand attitude refers to what, and the strength of what, people think (cognitive) and feel (affective) about a brand. This can range from positive to negative and be based on both cognitive and/or affective motivations and resonance. Brand attitude can guide a consumer’s behaviour towards advertisements (Percy & Rossiter, 1992). Previous research on the effects of banner advertisements, showed that longer ad lengths had a negative influence and decreased respondents’ brand attitudes as opposed to the influence of longer ads on brand awareness (Wang et al., 2013). Edwards, Li and Lee (2002)

(9)

studied consumers’ perceived intrusiveness of pop-up advertisements and its consequences on brand attitude. The results showed ads that were perceived as intrusive led to negative consumer responses such as negative attitudes (irritation) and avoidance. Perceived

intrusiveness and irritation of an advertisement can result in advertisement avoidance and/or negative attitude formation. As stated earlier, advertisements longer than 5-seconds have an abandonment rate of about 73% to 88% and indicated that advertisement length was the main reason for skipping ads (Dehghani et al., 2016; Pikas & Sorrentino, 2014; Saputra & Fachira, 2014). Thus, longer advertisements are likelier to be avoided than shorter advertisements and be perceived as intrusive.

Generally advertisements interfere with consumers’ goals to view a video online (Edwards, Li, & Lee, 2002). An individual takes action to go online and to view a video (goal), however an advertisement interrupts the consumer in achieving this goal.

Consequently, if an advertisement is too long, consumers can feel irritated (negative attitude) and avoid the ad. However, Edwards et al. (2002) did not find evidence to substantiate that longer ads are perceived as more intrusive than shorter ads. Goodrich et al. (2015) even found a reverse effect of ad length on brand attitude, and found that shorter ads were perceived as more intrusive.

Nonetheless Goodrich et al. (2015) propose to further research short ad formats to limit consumer annoyance with advertisements. Most YouTube videos are short and under 10 minutes in length (Davidson et al., 2010). Therefore, shorter advertisements might stimulate a higher consumer tolerance of advertisements. As skipping and avoidance behaviour could indicate negative attitudes, it is therefore expected that shorter advertisements lead to a more positive brand attitude than long ads. The following hypotheses are therefore proposed:

(10)

H1. The length of an advertisement positively influences brand awareness but negatively influences brand attitude:

H1a. A 30-second advertisement leads to a greater brand awareness but to a lower brand attitude than a 5-second or 3-second advertisement.

H1b. A 5-second advertisement leads to a greater brand awareness but to a lower brand attitude than a 3-second advertisement.

Thematic congruity

Although Goodrich et al. (2015) did study the effect of ad length on both brand awareness and attitudes of online video advertisements, the authors omitted to determine whether the advertisement was congruent with the video that followed. Apart from the characteristics of an advertisement such as length, contextual factors such as thematic congruity can also affect viewers’ brand awareness and/or brand attitudes in a similar or different manner. Thematic congruity or incongruity is also referred to as the

consistency/inconsistency between an advertisement and its context. A congruent in-stream video advertisement thematically matches the video it precedes, while an incongruent in-stream video ad thematically mismatches with the succeeding video on, for example, topic content (Li & Lo, 2015).

Advertisement congruity has been researched a great deal among marketing and advertisement studies. However, the findings on the effectiveness of (in-)congruence are inconclusive and show different results on awareness and attitudes. The contrasting results could be explained by the different measurement methods of dependent variables, but also by the different research methods and media platforms (Moorman, Neijens, & Smit, 2002).

For example, studies have shown and proposed that incongruity increases consumers’ brand awareness in online video environments and for sponsorship deals (Li & Lo, 2015;

(11)

Jagre, Watson, & Watson, 2001). However, other authors have claimed that not incongruity, but congruity increases brand awareness in, for example, magazines (Moorman et al., 2002). Similarly, for brand attitude a study by Edwards et al. (2002) showed that pop-up ads were perceived as less intrusive when they were congruent with the editorial content of a website. On the other hand, in a different field of study on sponsorship marketing, Jagre, Watson and Watson (2001) suggested that events, which were incongruent with the company, were evaluated more favourably by consumers than congruent sponsorships.

These inconclusive results indicate a knowledge gap in congruity research. One can deliberate under which circumstances thematic (in-)congruence effectively influences brand awareness and brand attitude. And more specifically, it could be argued that a combination of thematic (in-)congruence and a certain advertisement length could lead to positive brand awareness and brand attitude in an online video environment. Current research on the combined effect of congruity and length is scarce and does not provide an all-encompassing reflection of the effects.

The effects of ad length and the effect of ad-context congruence and advertisement position have been previously researched in online video environments (Li & Lo, 2015). The findings showed that ad length has a positive effect on brand name recognition and that the effect of thematic (in-)congruity on brand name recognition is dependent on the position of the ad (pre-roll, mid-roll or post-roll). Thus, Li and Lo (2015) did not research whether there was an interaction between ad length and congruence and as a result the relation between ad length and thematic congruency remains unclear. The proposed interaction effect between ad length and thematic congruity can be explained via Mandler’s Schema Incongruity Theory on people’s information processing which is (in-)congruent to his/her expectations.

Mandler’s Schema Incongruity Theory. As described above, it is expected that congruent ads influence the effect of ad length on brand awareness and brand attitude in a

(12)

different manner than incongruent advertisements. It is proposed that the level of congruity affects specific schemas which an individual retrieves when being exposed to an ad. Schemas are cognitive structures of stored memory, which comprises of prior knowledge of,

experiences with, and attitudes about, for example, a product category or a brand. These schemas guide people’s perceptions and thoughts when they encounter new knowledge about this category or specific brand. Ergo, schemas enable processing and when new information becomes available, existing schemas allow the processing to be retrieved easier and faster due to the existent memory links (Hsieh, Lo, & Chiu, 2016; Jagre et al., 2001).

Therefore, when an individual is exposed to an advertisement, existing schemas are stimulated. These advertisements can either conform to the knowledge schemas (congruent) or differ from the expectations based on the schemas (incongruent). When new information is congruent with a consumer’s expectations, his/her previous perceptions are reinforced. As people generally appreciate things that correspond to their expectations, congruent

advertisements therefore generate positive attitudes. (Hsieh et al., 2016; Noseworthy, Di Muro, & Murray, 2014). However, due to congruent ads’ conformity to expectations, they do not require a lot of engagement and thus limit awareness (Noseworthy et al., 2014).

Incongruent advertisements do not meet consumer’s expectations and, as a consequence, they enhance cognitive processing. The incongruent advertisement leads to responses such as arousal due to the surprise of incongruity. This enhances the memorability and persuasion of a message as consumers actively elaborate on trying to add the new information to the existing schemas (Dahlén, Lange, Sjödin, & Törn, 2005). Thus, recall and recognition (awareness) can be favourably affected by incongruent messages.

The choice of advertisements with thematic congruence or incongruence with the ensuing video is dependent on the advertising objectives (Kononova & Yuan, 2015). In this research it is examined whether the weaknesses of short ads (low brand awareness) and long

(13)

ads (low brand attitude) can be compensated for by combining it with different levels of ad thematic (in-)congruity. It is expected that the shortness of the ad leads to a positive brand attitude. However, the brand awareness level is expected to be weaker for shorter ads due to the short processing time to elaborate on the ad message. By linking ad length and thematic congruity, it is expected that the thematic incongruence between the ad and ensuing video enhances brand recall and recognition of short advertisements. Likewise it is expected that for longer advertisements, by incorporating a congruent ad with the ensuing video, the attitude level can be increased while the longer length of the ad increases brand awareness. Therefore the following hypotheses are proposed:

H2. The level of congruity moderates the effect of advertisement length on brand awareness and brand attitude:

H2a. 15-second advertisements congruent with the ensuing video generate more positive brand attitudes than 15-second advertisements incongruent with the ensuing video. H2b. 3- and 5-second advertisements incongruent with the ensuing video generate greater brand awareness than 3- and 5-second advertisements congruent with the ensuing video.

H2c. 3-second advertisements incongruent with the ensuing video generate greater brand awareness than 5-second advertisements incongruent with the ensuing video.

Based on these theories and empirical findings the conceptual model shown in Figure 1 is studied.

(14)

Figure 1. Conceptual model.

Method section Design

Using a 3-by-2 factorial between-subjects design with factor length of advertisement and factor level of congruity was conducted. Factor length had three levels namely: 3-second advertisement, 5-second advertisement and 15-second advertisement. Factor level of

congruity had two levels, namely: congruity and moderate incongruity. Each participant was assigned to one experimental condition and the groups were compared to one another. The aim was to achieve moderate incongruence and not extreme incongruity, as previous research has shown that consumers have adverse attitudes towards extreme incongruence (e.g. Jagre et al., 2001; Noseworthy et al., 2014). The processing of extreme incongruent messages requires greater cognitive effort than congruent or moderate

incongruent messages. When an individual is incapable of solving an extreme incongruent message, it results in more negative responses towards the message (Dahlén et al., 2005; Hsieh et al., 2016). Therefore, moderate incongruence between advertisement and the ensuing video was established in this research.

(15)

Sample

Due to limited resources such as time, money and accessibility, the participants were gathered via convenience sampling by way of social media. A total of 276 respondents participated in this study, which is above the required minimum of 180 participants (at least 30 participants per experimental group) necessary for statistical power.

The sample was controlled for missing data and outliers on the dependent variables. A participant who did not answer the items on brand awareness and attitude, and/or with two or more outliers (value ≥ 2.20) was considered unusable. As a result, 64 participants were removed from the sample. The final sample N = 212 was comprised of both females (N = 118) and males (N = 87) between 18-65 years of age (Mo = 18-25 years old). As a result of the convenience sample, there was a mixture of several nationalities, however the majority of the respondents were Dutch (N = 118) and had a bachelor’s degree or an equivalent level of education (N = 99).

Adolescent individuals, below 18 years of age, were excluded from participation in this study. It was considered unethical to show minors commercials and moreover parental consent would be required ("Kinder- en Jeugdreclamecode," n.d.). Age shows a negative correlation with YouTube usage and social media usage in general (Nielsen, 2007; "Frequency at which," n.d.; Smith, 2016). Participants older than 65 years of age are least (active) of all age groups between 18-65+ on YouTube and are less likely to encounter online video advertisements on video platforms such as YouTube (Jpriestley, 2011; Blattberg, 2015; Chaffey, 2016). Participants older than 65 of age were therefore excluded and both males and females aged 18-65 were included in the sample.

Stimuli

Gender-neutral content for both the advertisement as the ensuing videos were chosen, so the content was equally interesting for both genders to view. Therefore the ad and ensuing

(16)

videos fell within the food and drinks categories, as the statistic-planning tool by OpenSlate showed that this type of content is viewed by 43% females and 57% male ("OpenSlate tool online," n.d.).

An advertisement of the international brand Coca-Cola was used as ad stimuli. In this research a familiar existing brand was used for two reasons. First, consumers are often already previously exposed to the brand which is being advertised. It was assumed that although consumers can always encounter new brands, they are more likely to encounter an advertisement by a large corporation, than one for an unfamiliar brand by a start-up company. Second, an international brand was chosen as the sample consisted of a mixture of

nationalities. Coca-Cola is the best-known brand in the world, which made it a suitable familiar existing brand for this sample ("How many countries," n.d.).

Length. The independent variable in this research was advertisement length, which consisted of three levels: a 3-, a 5- and a 15-second advertisement. The existing Coca-Cola advertisement "Anthem" which was made available on the company’s YouTube channel on January 19th, 2016 was used as it included the new tagline "Taste the Feeling" (Coca-Cola, 2016). The same original 60-second advertisement was slightly edited and used for each condition, so the content of the advertisements mostly overlapped. This is important because differences in the outcomes between the conditions should be assigned due to the

manipulation and not due to differences between two different advertisements in content. All three advertisements included the same shots of the brand’s product, name, logo and tagline. The 5- and 15-second ad also included a frame where the product was being consumed and the 15-second ad also included a shot where the Coca-Cola was poured into a glass.

Congruence. The level of congruity was a moderator variable in this research and consisted of two levels: congruity and (moderate) incongruity. The congruence manipulated was the ensuing video, not the advertisement. For both conditions, the video included a "How

(17)

to" video to provide participants with a similar type of video and to limit the effects of video content differences. Both videos included an individual’s hands showing the process and written explanations of the steps on screen (white words). There were no verbal explanations as the same instrumental audio was placed underneath the videos. The videos were also edited into a similar video length of approximately 50-seconds.

Both edited congruent and incongruent videos showed an existing "How to video" and were extracted from YouTube. Both videos were chosen for the similar number of views (1.016 and 640 views), making it less likely to have been previously seen by participants and should therefore prevent boredom effects. The congruity video provided instructions on "How to make a Lime Cola". The video was congruent with the Coca-Cola brand as one of the ingredients of the drink was the soda coke and both products were beverages. The moderate incongruity video differed in topic and provided instructions on "How to make Guacamole". The video was incongruent as it did not show a beverage nor did it include Coca-Cola. However, it fell under the food and drinks category, which made the topic moderately incongruent.

Pre-test

The aim to establish congruent and moderate incongruent advertisements and ensuing videos was verified in the pre-test among 10 female participants. The participants rated the similarity between the advertisement and video theme on a 7-point semantic scale by

answering one question with two items as done by Li and Lo (2015): "To what degree do you consider the advertisement prior to the video to be relevant to the video that followed?" (1 = Strongly irrelevant/unrelated, 7 = Strongly relevant/related). A correlation analysis between

the dependent variables, relevance between ad and video and similarity between the ad and ensuing video showed there was a very strong positive association N = 10, r = .99, p < .001. The more relevant one considered the ad and ensuing video, the more one perceived the ad

(18)

and video to be related. Therefore, the two items were computed into the new variable congruity (M = 3.05, SD = 2.59).

The means showed differences between the three levels of congruity (congruent, moderate incongruent, extreme incongruent). Respondents rated the congruent condition (M = 6.67, SD = 0.58) higher than the moderate incongruent condition (M = 1.83, SD = 1.04), and the extreme incongruent condition (M = 1.25, SD = 0.50). The moderate incongruent condition was rated slightly less incongruent than the extreme incongruent condition and was therefore used in the experiment as the moderate incongruent condition. The pre-test thus showed that congruent and moderate incongruent conditions differed from one another and were established successfully and were therefore used for the final experiment.

Procedure

The participants completed a self-administered online experiment and were randomly assigned to one of the six conditions. They were first directed to an introduction text with information about the research and the participant’s rights. If the participant agreed to the terms and conditions, he/she could continue the experiment. Participants were told the experiment was about online videos on the platform YouTube and they would be shown a short YouTube video and would be asked questions about the video. The information did not expose the true intentions of the experiment, so participants’ attention to the advertisements remained unaffected.

Next, the participants were given an introduction text before watching the video by reading a fictional situation for why they were watching the video: e.g. "Imagine you are interested in making a Lime Cola and you go on YouTube to find a video that explains to you how to make one". After watching the video once, they were asked to proceed to fill in a questionnaire. These applied the dependent variables, moderating variables, manipulation checks, control variables and demographic variables. At the end of the study participants

(19)

were thanked and debriefed about the true purpose of the study and were kindly asked to not reveal any information about the study to other participants.

Measurements

Dependent variables.

Brand awareness. Brand awareness was measured via both recall and recognition. Unaided recall was measured as done by Jeong (2015) and Dahlén et al. (2008) with one open question: "Please name all brands, as specifically as possible, that you remember being advertised before the video".Aided recognition was measured with one multiple question item: "Please identify the brand that was shown in the advertisement before the video". The answer options were: Fanta, Starbucks, Pepsi, Coca-Cola, Don’t know (Li & Lo, 2015; Jeong, 2015). Both variables were recoded as dichotomous with 0 = Incorrect, 1 = Correct (Recall Mo = Correct. 80.7% answered correct, SD = .40, recognition Mo = Correct. 89.6% answered correct, SD = .31).

Brand attitude. Brand attitude was measured via four items on a 7-point semantic scale: "I think the brand Coca-Cola is …"; e.g. items: Bad/Good; Unfavourable/Favourable (1 = Negative evaluation, 7 = Positive evaluation; Kim & Choi, 2012; Dahlén et al., 2005). A principal component analysis (PCA) with Varimax rotation indicated the four items formed a single uni-dimensional scale, with one component with an eigenvalue above 1 (EV = 3.48). All items had a factor loading higher than 0.91 explaining 86.87% of the variance. The reliability of the scale was excellent α = .95 (M = 4.90, SD = 1.51).

Advertisement awareness. Advertisement awareness consisted of advertisement recognition and was measured with one item as done by Voorveld and Valkenburg (2015): "Have you just seen/heard the advertisement shown above?". The answer options were: No

definitely not, Yes maybe, Yes definitely, Don’t know. The variable was recoded as dichotomous with 0 = No, 1 = Yes (Mo = Yes. 79.2% answered yes, SD = .41).

(20)

Advertisement attitude. Advertisement attitude was measured with six items on a 7-point semantic scale: "I thought the advertisement was…"; e.g. items: Bad/Good,

Unpleasant/Pleasant (1 = Negative evaluation, 7 = Positive evaluation; Lee & Thorson, 2009; Rifon, Choi, Trimble, & Li, 2004). All items loaded on one factor that proved to be reliable (EV = 4.48, R2 = .75, α = .93, M = 4.24, SD = 1.39).

Video attitude. Attitude towards the congruent lime YouTube video was measured with three items on a 7-point semantic scale as done by Rifon, Choi, Trimble and Li (2004): "I thought the YouTube video on how to make a Lime Cola was…"; e.g. items: Bad/Good,

Pleasant/Unpleasant (1 = Negative evaluation, 7 = Positive evaluation). All items loaded on one factor that proved to be reliable (EV = 2.56, R2 = .85, α = .91, M = 4.81, SD = 1.44). The

same items were used to measure attitude towards the moderate incongruent guacamole YouTube video. All items also loaded on one factor that proved to be reliable (EV = 2.74, R2

= .91, α = .95, M = 5.30, SD = 1.36).

Perceived intrusiveness. Perceived intrusiveness was measured with four items on a 7-point Likert scale as done by Edwards et al. (2002): "When the advertisement played before the video, I thought it was…"; e.g. items: Distracting, Intrusive (1 = Strongly

disagree, 7 = Strongly agree). All items loaded on one factor that proved to be reasonable reliable (EV = 2.07, R2 = .52, α = .65). However, if one item was removed, the scale

improved significantly to α = .76. The factor loading for this item was also low and therefore this item was removed from the new scale (M = 4.21, SD = 1.35).

Irritation. Irritation towards the advertisement was measured with four items on a 7-point Likert scale as done by Edwards et al. (2002): "When the advertisement played before the video, I thought it was…"; e.g. items: Irritating, Reasonable (1 = Strongly disagree, 7 = Strongly agree). A PCA indicated the scale with four items had two components with eigenvalues above 1 (factor 1 EV = 1.77, R2 = .44, factor 2 EV = 1.09, R2 = .27). As both

(21)

factors only consisted of two items, two correlation analyses between the variables of the two factors were conducted.

The analysis showed that there was a strong positive association between the two items of factor 1 N = 207, r = .56, p < .001. The more one perceived the ad as irritating, the more one perceived the ad as terrible. Therefore, the two items were computed into a new irritation variable (M = 3.74, SD = 1.40). The second analysis showed that there was a weak positive association between the two items of factor 2 N = 207, r = .29, p < .001. The more one perceived the ad as unreasonable, the more one perceived the ad as less authentic. The two items were computed into a new irritation variable (M = 4.19, SD = 1.12).

Manipulation checks.

Advertisement length. The manipulation of advertisement length was controlled with one question via a movable scale: "How long do you think the advertisement before the video was?"; 0 = Zero seconds, 30 = Thirty seconds. Respondents could move the bar and indicate how many seconds they thought that the advertisement before their video lasted for the 3-second ad (M = 4.27, SD = 3.40), 5-second ad (M = 6.60, SD = 3.89), and 15-second ad (M = 13.08, SD = 4.54).

Level of congruity. Advertisement and video (in-)congruence was measured via one item on a 7-point semantic scale as done by Li and Lo (2015) by asking: "To what degree do you consider the advertisement prior to the video to be relevant to the video that followed?"; 1 = Strongly irrelevant, 7 = Strongly relevant (M = 4.03, SD = 2.16).

Advertisement. The manipulation check for advertisement served as a control

question as to whether the participant had seen an advertisement. This was measured via one item: "Do you remember seeing an advertisement before the video?"; Yes, Not sure, No (Mo = Yes. 87% answered yes, SD = .34).

(22)

Multitasking. Participants’ focus on the stimuli was measured via one item: "Were you simultaneously doing something else when the advertisement played before the video?"; Yes, Don’t remember, No. The variable was recoded as 0 = No, 1 = Yes (Mo = No. 72.9% answered no, SD = .41).

YouTube usage. Participants’ YouTube usage frequency was measured with one item: "How often do you use YouTube?"; Daily, Weekly, Monthly, Annually, Never. The variable was reverse coded as 1 = Never, 5 = Daily (Mo = Daily. 48.8% uses YouTube daily).

Ad familiarity. Advertisement familiarity was measured with one item on a 7-point semantic scale as done by Li and Lo (2015): "How familiar are you with the shown Coca-Cola advertisement?"; Unfamiliar/Familiar (1= Strongly unfamiliar, 6 = Strongly familiar, 0 = Don’t know it at all; M = 3.34, SD = 1.67).

Brand familiarity. Brand familiarity was measured with one item on a semantic 7-point scale as done by Rifon et al. (2004): "How familiar are you with the brand Coca-Cola?"; Unfamiliar/Familiar (1= Strongly unfamiliar, 6 = Strongly familiar, 0 = Don’t know

it at all; M = 5.57, SD = .93).

Brand usage. Brand usage was measured with one item: "How often do you consume Coca-Cola?"; Daily, Weekly, Monthly, Annually, Never. The variable was reverse coded as 1 = Never, 5 = Daily (Mo = Monthly. 33% consumes Coca-Cola monthly).

Advertisement avoidance. Advertisement avoidance consisted of affective, cognitive and behavioural avoidance and were measured as done by Cho and Cheon (2004). Affective advertisement avoidance was measured with two items on a 7-point Likert scale: e.g. "I hate online video advertisements"; 1 = Strongly disagree, 7 = Strongly agree. A correlation analysis between the two items showed there was a strong positive association N = 216, r = .74, p < .001. The more one disliked ads, the more one thought it would be better if there were no online ads at all. Therefore, the two items were computed into a new affective

(23)

variable (M = 5.33, SD = 1.41).

Cognitive advertisement avoidance was measured with two items on a 7-point Likert scale: e.g. "I intentionally ignore online video advertisements"; 1 = Strongly disagree, 7 =

Strongly agree. A correlation analysis between the two items showed there was a moderate positive association N = 216, r = .39, p < .001. The more one ignored video ads intentionally, the more one was likely to ignore ads in general. Therefore, the two items were computed into a new cognitive variable (M = 5.57, SD = 1.12).

Behavioural advertisement avoidance was measured with three items on a 7-point Likert scale: e.g. "I scroll down Web pages to avoid online video advertisements"; 1 =

Strongly disagree, 7 = Strongly agree. All items loaded on one factor that proved to be reliable (EV = 1.77, R2 = .59, α = .64, M = 4.72, SD = 1.27).

Demographics. Gender, age, level of education, nationality and residence were each measured with one multiple question item at a nominal or ordinal level.

Results Manipulation checks

Advertisement length. A one-way analysis of variance (ANOVA) with ad length as the independent variable and the perception of advertisement length of participants as the dependent variable was conducted. The analysis showed a significant overall difference between the length conditions F(2,189) = 81.77, p < .001. A Bonferroni post hoc test showed that participants in the 3-second advertisement condition (M = 4.27, SD = 3.40) reported seeing a significantly shorter advertisement than participants in the 5-second condition (M = 6.60, SD = 3.89) with p = .004. The participants in the 3-second advertisement condition also evaluated the ad significantly shorter than participants in the 15-second advertisement

(24)

shorter than the 15-second advertisement (Mdifference = 6.48, p < .001). These results indicate that the manipulation for advertisement length was successful.

Congruity level. An independent samples t-test with congruity as independent variable and participant’s perception of congruity between the advertisement and ensuing video as dependent variable indicated that the manipulation was successful t(190) = 10.56, p < .001, 95% CI [2.14, 3.12]. Participants in the congruent condition (M = 5.42, SD = 1.08) significantly rated the advertisement and ensuing video as more relevant than participants in the incongruent video condition (M = 2.79, SD = 1.67). These results illustrate the

manipulation for congruity level was also successful. Randomization checks

A series of Chi-square tests were conducted to control if the random assignment to the conditions was successful. Ad length and congruity were used as independent variables and combined with the following control variables: gender, age, nationality, education, whether the participant was multitasking during the experiment, brand usage and YouTube usage. The results of the Chi-square tests displayed in Tables 1 and 2 (see Appendix A) show there were no differences between the conditions in the number of men and women, nor in the age category groups. The conditions did also not differ in nationality, educational level, multitasking activity or in brand usage.

The Chi-square test did show there were significant differences between the ad length conditions in YouTube usage, χ²(8, N = 207) = 21.72, p = .005. There were no differences between the congruity conditions χ²(4, N = 207) = 3.33, p = .504 in YouTube usage. Therefore YouTube usage was controlled for in the main analyses.

To check if the random assignment to the conditions was also successful for interval/ratio level control variables, a series of two-way analysis of variances (two-way ANOVA) were conducted. Ad length and congruity were used as independent variables and

(25)

combined with the following control variables: brand familiarity, ad familiarity, affective ad avoidance, cognitive ad avoidance, behavioural ad avoidance, ad awareness, ad attitude, ad irritation 1, ad irritation 2 and intrusiveness.

The results of the two-way ANOVA’s are displayed in Table 3 (see Appendix B). The results show there were no significant differences between the conditions in the familiarity of the brand or ad. The groups also did not differ in affective, cognitive and behavioural

avoidance towards ads or in ad awareness and ad attitude. The analyses did show there was an interaction between the ad length and level of congruity conditions for ad irritation 1 F(2, 201) = 4.12, p = .018, η² = .04 and for intrusiveness F(2, 201) = 3.96, p = .020, η² = .04. Therefore irritation towards the ad 1 and perceived intrusiveness of the ad were controlled in the main analyses.

Covariates

Table 4 shows the output of the correlations between the dependent variables brand recall, brand recognition and brand attitude and the control variables ad awareness, ad attitude, intrusiveness, ad irritation 1, ad irritation 2, ad familiarity and brand familiarity (see Appendix C).

The correlation output in Table 4 shows there was a significantly weak positive correlation between the dependent variable brand recognition and the control variable ad awareness. The higher one’s brand recognition, the higher one’s ad awareness N = 212, r = .21, p = .002. The analysis also shows there were significant correlations between the dependent variable brand attitude and the control variables. There was a weak negative correlation between brand attitude and ad irritation 1. It revealed that the more positive one’s brand attitude, the less one considered ads to be irritating N = 207, r = -.26, p < .001. There was a weak negative correlation between brand attitude and intrusiveness. The more positive one’s brand attitude, the less one perceived ads as intrusive N = 207, r = -.25, p < .001.

(26)

The correlation output also shows there was a moderate positive correlation between brand attitude and ad attitude. The more positive one’s brand attitude, the more positive one’s ad attitude N = 212, r = .40, p < .001. Brand attitude and control variable ad irritation 2 correlated weakly negative. The more positive one’s brand attitude, the less one considered ads to be irritating N = 207, r = -.25, p < .001. Lastly, the correlation output in Table 4 shows there was a weak positive correlation between brand attitude and ad familiarity. The more familiar the ad, the more positive one’s brand attitude N = 192, r = .19, p = .009.

The correlation analysis showed that ad awareness, ad attitude, ad irritation 1, ad irritation 2, ad familiarity and intrusiveness were significantly correlated with the dependent variables. However, the randomization was successful for gender, age, nationality, education, multitasking, brand usage, ad familiarity, ad avoidance, ad awareness, ad attitude and ad irritation 2. Consequently, ad awareness, ad attitude and ad irritation 2 were not included as covariates. Only YouTube usage, ad irritation 1 and intrusiveness were therefore used as covariates in the analyses.

Hypotheses testing

The hypotheses were tested via three separate two-way analysis of covariance‘s (ANCOVA’s). Each ANCOVA tested hypotheses H1 and H2 with a different dependent variable (brand recall, brand recognition or brand attitude).

Brand recall. A two-way ANCOVA with ad length and congruity as independent variable and moderator variable and brand recall as dependent variable was conducted. The variables YouTube usage, ad irritation 1 and intrusiveness were included as covariates in the analysis. The analysis showed there was no main effect of ad length F(2,198) = 2.31, p = .051 (one-tailed), although it did approach significance. There were also no main effects of

congruity F(1,198) = 1.28, p = .130 (one-tailed), nor a moderation effect of ad length and congruity F(2,198) = .39, p = .340 (one-tailed).

(27)

There was also no significant effect of the covariate YouTube usage F(1,198) = 2.39, p = .123 (two-tailed), nor of covariate ad irritation 1 F(1,198) = 2.21, p = .139 (two-tailed), nor an effect of covariate intrusiveness F(1,198) = 3.85, p = .051 (two-tailed) on brand recall. Therefore H1, H1a and H1b were rejected with regards to brand recall. Hypothesis H2a, H2b and H2c were also rejected with regard to brand recall, as there was no moderation effect.

Brand recognition. A two-way ANCOVA with ad length and congruity as independent variable and moderator variable and brand recognition as dependent variable was conducted. The variables YouTube usage, ad irritation 1 and intrusiveness were included as covariates in the analysis.

The analysis indicated there was a main effect of ad length F(2,198) = 2.50, p = .043, η² = .03 (one-tailed). A Bonferroni post hoc test showed that respondents in the 3-second advertisement condition (M = .84, SD = .04) had a significantly lower brand recognition than participants in the 15-second condition (M = .96, SD = .04) with p = .041 (one-tailed). However, there were no significant differences between the participants in the 3-second condition and the participants in the 5-second condition (M = .89, SD = .04) with p = .477 (one-tailed), nor between the participants in the 5-second condition and the 15-second condition (Mdifference = -.06) with p = .322 (one-tailed) for brand recognition (see Table 5, Appendix D).

There was no main effect of congruity F(1,198) = .84, p = .180 (one-tailed), nor a moderation effect of ad length and congruity F(2,198) = .61, p = .273 (one-tailed). The analysis also showed the covariates YouTube usage F(1,198) = 1.13, p = .289 (two-tailed) , ad irritation F(1,198) = .32, p = .574 (two-tailed), and intrusiveness F(1,198) = .19, p = .666 (two-tailed) were not significantly related to brand recognition. The two-way ANCOVA showed significant differences between the 3-second condition and 15-second condition and therefore H1a was partially supported with regard to brand recognition. H1b and H2a, H2b

(28)

and H2c were rejected with regard to brand recognition, as there were no significant

differences between the 3-second and 5-second condition and there was no moderation effect. Brand attitude. A two-way ANCOVA with ad length and congruity as independent variable and moderator variable and brand attitude as dependent variable was conducted. The variables YouTube usage, ad irritation 1, and intrusiveness were included as covariates in the analysis. The analysis indicated there were no main effects of ad length F(2,198) = 1.03, p = .180 (one-tailed), nor a main effect of congruity F(1,198) = .25, p = .308 (one-tailed), nor a moderation effect of ad length and congruity F(2,198) = .74, p = .239 (one-tailed) on brand attitude.

The analysis also showed that the covariates YouTube usage F(1,198) = .07, p = .787 (two-tailed), ad irritation 1 F(1,198) = 2.32, p = .129 (two-tailed), and ad intrusiveness F(1,198) = 1.70, p = .194 (two-tailed) were not significantly related to a participant’s brand attitude. Therefore H1, H1a and H1b were rejected with regard to brand attitude. Hypotheses H2a, H2b and H2c were also not supported and therefore rejected with regard to brand attitude, as there was no moderation effect.

Conclusion

This research examined the effect of three lengths of forced advertisements in an online video environment on viewers’ brand awareness and brand attitude. The results showed that advertisement length did have an effect on brand recognition but not on brand recall. The brand Coca-Cola was better recognized in the case of a 15-second advertisement than when 3-second advertisements were used. However, there were no differences in brand recognition when comparing the 3-second advertisements and the 5-second advertisements or between the 5-second advertisements and the 15-second advertisements. Contrary to

(29)

shorter ads of 5- or 3-seconds, nor vice versa: the diverse advertisement lengths did not lead to different attitudes towards the brand.

Additionally, the findings indicated that the level of thematic congruity between advertisement and the ensuing video had no moderating effect on brand awareness or on brand attitude. Participants did not recognize or recall the brand better when the

advertisement was combined with a congruent or moderately incongruent ensuing video. The effect of advertisement length on brand attitude was also not moderated by the level of thematic congruity between the ad and ensuing video: none of the effects of 15-, 5- or 3-second advertisements on brand attitude differed significantly from one another regardless of its level of congruity.

Discussion

In accordance with the ELM it was hypothesized that advertisement length was positively related to brand awareness, as longer ads gave viewers more time to process information. The findings of this study only partially confirmed these predictions:

participants exposed to the 15-second advertisements recognized the brand better than those in the 3-second advertisement conditions. This outcome shows moderate similarities with Li and Lo’s (2015) findings, which showed that longer online advertisements of 30-seconds resulted in greater brand recognition than 15-second advertisements. However, in this study there were no differences between the 3- and second advertisement conditions or in the 5-second and 15-5-second advertisement conditions. Noteworthy is that the difference in ad length in Li and Lo’s (2015) study was 15-seconds, while the differences in ad length were shorter in this analysis. It could be that the effects of ad length differ when the difference between the ad length conditions is greater. It could also be that the differences between shorter ad lengths do not result in greater or lower brand recognition in general.

(30)

Unlike previous findings by Goodrich et al. (2015) no effects of ad length were found on brand recall, which indicates that ad length does not influence viewers’ brand recall. An explanation for this could be that a well-known brand such as Coca-Cola requires few cues to trigger awareness, as people already have stored pieces of memory of a brand. It could have therefore been that any cue, short or long, already triggered these existing schemas, which lead to an equal brand recall despite ad length.

It was also expected that longer advertisements interfere with a viewer’s goal to watch a short online video and as a result, viewers’ feelings of interference would lead to a lower brand attitude than short ads (Edwards et al., 2002). In this study no evidence was found to support this notion. Diverse advertisement lengths did not lead to different attitudes towards the brand. It is debatable whether these findings are the result of a type of automatic "anti-ad mechanism" of viewers. According to reactance theory from psychology literature, people respond to every action with an opposite reaction when their freedom or goal is threatened (Edwards et al., 2002; Quick, Kam, Morgan, Montero Liberona, & Smith, 2015). In this research, viewers did not have the option to skip the ad. In this manner, their freedom to view the video was threatened by the ad and it could have been that they automatically responded in a negative way to the advertisement. For example, Cho and Cheon (2004) found that a threat to consumers’ goals led to greater ad avoidance. It is possible that once viewers learn that they do not have any control over the advertisement whatsoever, an automatic line of thinking of "here is yet another ad" emerges resulting in overlooking all ad distinctions in general. As people respond adversely to ads, perhaps when they are given the option to interact with the advertisement by being able to skip it, other brand attitudes for different ad lengths emerge.

Additionally, it was researched whether the level of thematic congruity between the advertisement and ensuing video moderated the effect of ad length on brand awareness and

(31)

brand attitude. It was proposed that the incongruence between the ad and ensuing video would enhance recall and recognition due to the surprising effect of the deviation from the expected norm (Dahlén et al., 2005). Congruency between the ad and ensuing video would reinforce viewers’ existing perceptions, which would generate positive attitudes (Hsieh et al., 2016; Noseworthy et al., 2014). However the results show these assumptions were not met. There were no effects of the level of congruity between the advertisement and ensuing video on brand attitude and/ or brand awareness.

It is attainable that there will be an effect when viewers have chosen a specific video themselves. In that case they would select a video of their own interest with a specific goal (e.g. entertainment, education) and are likely to be more involved and motivated to watch the video. As discussed above, people will respond to advertisements when it threatens their freedom or goal to watch an online video (Edwards et al., 2002). Thus as a consequence of the goal interference, it could be that the level of congruity will be better observed by viewers when they have selected a video themselves.

Theoretical implications

Interestingly, the findings of this study provide several implications for theory and practice. This study shows that the effect of ad length differs for brand recall and brand recognition, indicating that recall and recognition measure two separate constructs of brand awareness. Lang (2002) explains this through the Limited Capacity Model in which memory consists of several distinctive components. However, previous research measuring brand awareness has often failed to recognize a clear distinction between brand recall and brand recognition or has even used the two terms interchangeably (e.g. Jeong, Sanders, & Zhao, 2011).

In this study, the effect of length on the separated constructs brand recall (retrieval) and brand recognition (storage) showed different outcomes; ad length did have an effect on

(32)

brand recognition but not on brand recall, indicating different effects of length on two constructs which jointly measure brand awareness. These effects of ad length can also be translated to Lang’s (2002) Limited Capacity Model. Ad length did effect how new pieces of information were stored and linked into an individual’s associative memory network, while ad length did not have an effect on how previously stored information was retrieved from the memory (Lang, 2002). It can therefore be argued that ad length differently influences distinct pieces of the memory as to recall/retrieval and recognition/storage.

More recent research on ad cross-media synergy and ad congruity did measure both brand recall and brand recognition as separate constructs (e.g. Voorveld & Valkenburg, 2015; Kononova & Yuan, 2015). This research therefore contributes on a broader level to theory and knowledge on the effects of ad length specifically on brand awareness. Other studies on the effects of ad length have either only considered aided recall (e.g. Singh & Cole, 1998; Goodrich et al., 2015) or merely evaluated brand recognition (e.g. Li & Lo, 2015) in their measurements of brand awareness.The findings of this research therefore complement Jeong’s (2015) research results on the length of television ads on brand recall and brand recognition.

Limitations

However, it should be noted that this study also has some limitations, which in turn provide stepping-stones for future research. In this study the worldwide well-known brand Coca-Cola was used. The reasoning behind this choice was that it is likely that viewers encounter online advertisements from brands that they are already familiar with. However, it may have been that the brand Coca-Cola was too well-known for viewers. People who are familiar with a brand often already have unyielding established attitudes about it.

In retrospect, when studying the effect of ad length and thematic congruity on brand awareness and brand attitude, advertising frequency should therefore be taken into account.

(33)

People’s established awareness and/or attitudes of a familiar brand are less likely to be directly influenced by a single exposure; it could instead take time and/ or an event or incident before a person changes his or her attitude towards a familiar brand. Research by Jeong, Sanders and Zhao (2011) showed that ad appearance frequency had an advantage over ad length in improving viewers’ brand recall and advertising appreciation. Thus, future research should take ad frequency into account when measuring the effects for ad length and thematic congruity of the ad and ensuing video to uncover the optimal conditions for ad effectiveness.

Practical implications

This study found an effect of ad length on brand recognition, in which the brand in a 15-second advertisement was better recognized than in a 3-second advertisement. However, one can argue on what the non-findings tell can tell practitioners. The ad length did not influence viewers’ brand attitudes or brand recall. So, if ad length does not influence viewers’ brand recall or brand attitude, it could be argued that it would be more profitable for

practitioners to merely produce short video advertisements of 3- or 5-seconds instead of 15-second or longer advertisements.

For example, a research report by MediaScience and Adfo Groep (n.d.) showed that the second most important reason for Dutch marketers to not include online video advertising in their marketing mix was the high costs of production of video material (MediaScience & Adfo Groep, n.d.). However, as viewers’ brand recall and brand attitude do not differ when the ad length is 3-, 5- or 15-seconds, it marks an opportunity for practitioners. First,

marketers can save money by creating shorter online video ads, which limits the high

production costs. Second, shorter ads can simultaneously generate the same brand recall and brand attitude as expensive 15-second ads. The challenge for practitioners is no longer how to cope with high production costs for long ads, but instead how they can successfully

(34)

incorporate the brand’s marketing messages in shorter online video advertisements.

References

Barreda, A. A., Bilgihan, A., Nusair, K., & Okumus, F. (2015). Generating brand awareness in online social networks. Computers in Human Behavior, 50, 600-609.

doi:10.1016/j.chb.2015.03.023

Blattberg, E. (2015, April 24). The demographics of YouTube, in 5 charts [Platform post]. Retrieved from http://digiday.com/platforms/demographics-youtube-5-charts/

Chaffey, D. (2016, April 21). Global social media research summary 2016 [Research report]. Retrieved from http://www.smartinsights.com/social-media-marketing/social-media- strategy/new-global-social-media-research/

Cho, C. H., & Cheon, H. J. (2004). Why do people avoid advertising on the internet? Journal of Advertising, 33(4), 89-97. doi:10.1080/00913367.2004.10639175

Coca-Cola. (2016, January 19). Anthem [Video file]. Retrieved from https://www.

youtube.com/watch?v=-AmKP9VE2Ms

Coursaris, C. K., Sung, J., & Swierenga, S. J. (2010, June). Effects of message characteristics, age, and gender on perceptions of mobile advertising. In 2010 ninth international conference on mobile business and 2010 ninth global mobility roundtable (pp. 198-205). doi: 10.1109/ICMB-GMR.2010.45

Dahlén, M., Lange, F., Sjödin, H., & Törn, F. (2005). Effects of ad-brand incongruence. Journal of Current Issues & Research in Advertising, 27(2), 1-12.

doi:10.1080/10641734.2005.10505178

Dahlén, M., Rosengren, S., Törn, F., & Öhman, N. (2008). Could placing ads wrong be right? Advertising effects of thematic incongruence. Journal of Advertising, 37(3), 57-67. doi: 10.2753joa0091-3367370305

(35)

Davidson, J., Liebald, B., Liu, J., Nandy, P., Van Vleet, T., Gargi, U., ... & Sampath,

D. (2010). The YouTube video recommendation system. In Proceedings of the fourth ACM conference on recommender systems (pp. 293-296).

doi: 10.1145/1864708.1864770

Dehghani, M., Niaki, M. K., Ramezani, I., & Sali, R. (2016). Evaluating the influence of YouTube advertising for attraction of young customers. Computers in Human

Behavior, 59, 165-172. doi:10.1016/j.chb.2016.01.037

Edwards, S. M., Li, H., & Lee, J. H. (2002). Forced exposure and psychological reactance: Antecedents and consequences of the perceived intrusiveness of pop-up ads. Journal of Advertising, 31(3), 83-95. doi:10.1080/00913367.2002.10673678

Frequency at which internet users access YouTube in Germany in February 2015, by age

group [Graph]. (n.d.). Retrieved from http://www.statista.com/statistics/429521/

youtube-usage-frequency-in-germany-by-age-group/

Goldstein, D. G., McAfee, R. P., & Suri, S. (2015). Improving the effectiveness of time- based display advertising. ACM Transactions on Economics and Computation, 3(2), 1–20. doi:10.1145/2716323

Goodrich, K., Schiller, S. Z., & Galletta, D. (2015). Consumer reactions to intrusiveness of online-video advertisements. Journal of Advertising Research, 55(1), 37-50.

doi:10.2501/jar-55-1-037-050

How many countries sell Coca-Cola? And is there anywhere in the world that doesn't sell it? [FAQ]. (n.d.). Retrieved from http://www.coca-cola.co.uk/faq/company/how-many- countries-sell-coca-cola-is-there-anywhere-in-the-world-that-doesnt/

Hsieh, A. Y., Lo, S. K., & Chiu, Y. P. (2016). Where to place online advertisements? The commercialization congruence between online advertising and web site context. Journal of Electronic Commerce Research, 17(1), 36-46.

(36)

Jagre, E., Watson, J. J., & Watson, J. G. (2001). Sponsorship and congruity theory: A theoretical framework for explaining consumer attitude and recall or event sponsorship. Advances in Consumer Research, 28(1), 439-445.

Jassies, K. (2015, December 17). Zo gebruiken jongeren Instagram, Snapchat & YouTube in 2016 [Trend report]. Retrieved from http://www.frankwatching.com/archive/2015/ 12/17/zo-gebruiken-jongeren-instagram-snapchat-youtube-in-2016/

Jeong, Y. (2015). The impact of the length of preceding and succeeding ads on television advertising effectiveness. Journal of Marketing Communications, 1-15.

doi:10.1080/13527266.2015.1048272

Jeong, Y., Sanders, M., & Zhao, X. (2011). Bridging the gap between time and space: Examining the impact of commercial length and frequency on advertising effectiveness. Journal of Marketing Communications, 17(4), 263–279. doi:10.1080/13527261003590259

Jpriestley. (2011, September 6). Social media usage in the UK: The findings [Blog post]. Retrieved from http://www.umpf.co.uk/blog/social-media/social-media-usage-in-the- uk-the-findings/

Kim, S., & Choi, S. M. (2012). An examination of effects of credibility and congruency on consumer responses to banner advertisements. Journal of Internet Commerce, 11(2), 139–160. doi:10.1080/15332861.2012.689565

Kinder- en Jeugdreclamecode (KJC) [Advertising code]. (n.d.). Retrieved from https:// www.reclamecode.nl/nrc/pagina.asp?paginaID=279%20&deel=2

Kitchen, P. J., Kerr, G., Schultz, D. E., McColl, R., & Pals, H. (2014). The elaboration likelihood model: Review, critique and research agenda. European Journal of

Marketing, 48(11/12), 2033–2050. doi:10.1108/ejm-12-2011-0776

(37)

environmental PSAs with thematically congruent advertisements in different formats affects memory and attitudes. Journal of Interactive Advertising, 15(1), 2-15. doi:10.1080/15252019.2015.1009524

Lang, A. (2000). The limited capacity model of mediated message processing. Journal of

Communication, 50(1), 46–70. doi:10.1093/joc/50.1.46

Lee, E. J., & Schumann, D. W. (2004). Explaining the special case of incongruity in

advertising: Combining classic theoretical approaches. Marketing Theory, 4(1/2), 59- 90. doi:10.1177/1470593104044087

Lee, J. G., & Thorson, E. (2009). Cognitive and emotional processes in individuals and commercial web sites. Journal of Business and Psychology, 24(1), 105-115. doi:10.1007/s10869-008-9087-8

Li, H., & Lo, H. Y. (2015). Do you recognize its brand? The effectiveness of

online in-stream video advertisements. Journal of Advertising, 44(3), 208-218. doi:10.1080/00913367.2014.956376

MediaScience & Adfo Groep. (n.d.). Video marketing monitor 2015. De opmars van online video advertising in de marketingmix. Adverteerders over doelstellingen, inzet en bestedingen. Retrieved from http://www.adformatie.nl/sites/default/files/

WhitePaperOnlineVideo_infographic_DEF.pdf

Moorman, M., Neijens, P. C., & Smit, E. G. (2002). The effects of magazine-induced psychological responses and thematic congruence on memory and attitude toward the ad in a real-life setting. Journal of Advertising, 31(4), 27-40. doi. 10.1080/00913367.2002.10673683

Newell, S. J., & Henderson, K. V. (1998). Super Bowl advertising: Field testing the importance of advertisement frequency, length and placement on recall. Journal of

Referenties

GERELATEERDE DOCUMENTEN

gemanipuleerde filmpje dat zij op dat moment zien, en niet de officiële brand promotion die begin vorig jaar op televisie te zien was. Omdat alle vijf de filmpjes op dezelfde

An opportunity exists, and will be shown in this study, to increase the average AFT of the coal fed to the Sasol-Lurgi FBDB gasifiers by adding AFT increasing minerals

It also presupposes some agreement on how these disciplines are or should be (distinguished and then) grouped. This article, therefore, 1) supplies a demarcation criterion

The executional cues that were used to measure advertising effectiveness were based on theory and consisted of nine different variables: celebrity, real people in real

research can be used to identify the impact of a humorous or irritating commercial, whether loyal customers respond differently to an advertisement and if they change their attitude

Hence, the main question examined in this study is: how does brand repositioning influence customer loyalty and how do exposure and self-congruity affect

I have investigated whether the effect of celebrity endorsement and producttype is different among the various reported personality traits, and if higher levels of brand

there was either no significant moderating effect of self-consistency or functional congruity on the relationships between inconsistent brand positioning, with reference to showing