The influence of Age on synergetic and direct effects of Online and Offline video advertisements

(1)

The influence of Age on synergetic and

direct effects of Online and Offline video

In the last decade, a large shift has occurred in the media landscape. More and more people consume media via their computers or mobile devices. This has caused a shift from traditional television to online channels such as YouTube and Facebook. Especially video is a growing driver of online media consumption. This has major consequences for the role of video advertisement in the marketing strategy of companies.

The goal of this study was to get a more in-depth understanding of the effect of video advertisement and the synergetic effects between online and offline video advertisement. More importantly, this study gave a deeper understanding of the influence of age on these effects. The research was performed on a dataset that consisted of panel data of 5678 Dutch households. The data was collected by GfK in the first 13 weeks of 2014. Two multiple regression models were specified to measure the effects of video advertisement on the sales of a global FMCG company.

The findings of this research are of major importance to the topic of video advertisement and especially on the influence of age. Surprisingly, the effects of television advertising were found to have a negative effect on sales. No effects were found of online channels on sales. Neither did the outcomes show a synergetic relationship between online and offline video advertising, nor between the difference in skippable and non-skippable pre-rolls. Some interesting findings were found on the topic of Age. It was found that older generations, i.e. people of 45 years and older, react positive on online video advertisements, while younger generations tent to react negatively on video advertisements on YouTube and RTL. Furthermore, younger people react positive on television advertisement while older generations react negatively on this.

(4)

4

Preface

After finishing my Bachelor in History on the University of Groningen, I am happy that I was able to join the Pre-Msc Marketing and the MSc Marketing Intelligence. I came a long way, but this is definitely the road I want to follow. Writing my Master thesis was a great experience. When looking back on the journey of writing my thesis, I have learned a lot. It was really interesting to work with real panel data from GfK and dive into the topic of Optimizing Media Strategies.

(5)

5

1. Introduction

In March 2017, the CEO of Adidas, Kasper Rorsted, declared that the company stops with television advertising and will focus primarily on digital channels. “It is clear that the younger

consumer engages with us predominately over the mobile device”, he said (CNBC, 2017). And

Adidas is not the only company that moves from traditional, linear television towards digital advertising. In the last decade, a large shift has occurred in the media landscape. With the rise of the internet, more and more people consume video via their computers, laptops and mobile devices. This has caused a shift from the traditional media channels - Television, Radio, and Print - towards online channels such as Facebook, Twitter and YouTube. Especially videos play an important role and are expected to constitute more than 85 percent of the traffic of consumer internet in the upcoming years (Handayani and Hudrasyah, 2015). Moreover, the amount of time that consumers are spending on online channels is more than they did on offline channels (Kumar, 2015).

It is therefore not surprising that companies have started to invest heavily in online advertising. Online advertising budgets have been rising steadily in the last few years. In the US alone online advertising surpassed television in total advertising spending at the end of 2016 (Forbes, 2016). According to a report from the IAB - which accounts for 86% of online advertising in the US - the average video advertisement spending has almost doubled between 2013 and 2016. The expectation is that this trend will continue in the upcoming years (IAB, 2016). But despite the growth of online channels, traditional media channels still play an important role in advertising. The spending on television advertisement in 2016 still accounts for more than 70 billion dollars in the US alone (IAB, 2016).

(8)

8

online platforms. These channels have different audiences and a different viewing experience than television, especially because younger people tend to react differently to these advertising campaigns than older people (Williams and Page, 2011). Second, the rise of mobile devices has also affected consumer behavior. Consumers use multiple forms of media - both online and offline- at the same time and watch advertisements on many different channels before proceeding to buy a product (Lin, Venkataraman and Jap, 2013). This can have influence on the effect an advertisement has, because seeing an ad on multiple channels can cause saturation (Zantedeschi et al, 2013).

It is therefore not surprisingly that the effectiveness of media has received more attention in marketing research in recent years (Kumar et al, 2004; Sethuraman, Tellis, and Briesch, 2011; Ataman, Van Heerde, and Mela, 2010). Research in recent years has been focusing on the effectiveness of using both online and offline channels. Multiple studies show that synergy effects can arise, where advertisements on multiple channels can strengthen the overall effectiveness as bigger than the sum of its parts (Naik and Peters, 2009; Lin, Venkataraman, and Jap, 2013). However, little research has been conducted on the influence of age on the effects of online video advertising and the synergy effects with television advertisement. This is especially of interest because young people tend to drive the online trend by spending 35 percent more time on online channels than older generations (Nielsen, 2016). For companies, it is necessary to get a better understanding of how different age groups react on online and offline video advertisements. For example, for Adidas it would be beneficial to know which age groups react positively on online video advertisement. The goal of this research is to address this gap in research. This paper will provide a deeper understanding of differences in age of the effects of television and online video advertising on offline sales. Thus, the research question of this study is:

(9)

9

AdWords, Radio and Print advertising. Second, by combining media effectiveness with age, new differences might be found between the effectiveness in different age groups. Third, this research not only focusses on the direct effects of advertisement, but also on synergy effects between channels. Because video advertising is increasingly more popular, this research can contribute to more insights in this topic. From a managerial point of view this research can also be useful. The outcomes can help managers to determine a better marketing strategy for companies that are advertising both online and online, while targeting the right age groups on the right channel.

(10)

10

2. Literature review

This chapter will elaborate more on the concepts discussed in the introduction and the existing research that has been done. A literature review will be done on all the concepts covered by this research. First, online and offline advertising effectiveness will be discussed. Second, synergy effects between television advertising and online video advertising on sales will be elaborated on. Third, research on the differences between age groups on media effectiveness will be examined.

2.1 Offline and Online advertising

2.1.1 Offline advertising

Danaher and Dagger (2013) found that traditional media channels such as catalogs, television, radio, and direct mail are still the most effective channels when it comes to purchase outcomes. As a part of the traditional media channels, the traditional television is one of the most effective channels, and appears to become even more effective than before the rise of the internet (Rubinson, 2009). This is because television mainly affects brand awareness, which has a large impact on sales. Dinner, van Heerde and Neslin (2011) found that offline channels have short-term and long-term effects on offline sales. These findings were confirmed by a meta-analysis performed by Sethuraman et al (2011). They found that advertisement has a higher elasticity for long-term effectiveness than for short-term effectiveness. Thus, although we have seen that a shift arises in the media landscape and marketing managers are investing in online advertisement, research shows that television advertising still has a large influence on sales.

(11)

11

research on this topic is therefore always useful. Third, testing this relationship is necessary to measure possible synergy effects within this model and must therefore be included. Based on the finding’s in literature, we can derive the following hypothesis:

H1a: Offline video advertisement positively effects offline sales

2.1.2 Online video advertising 2.1.2.1 Online advertising

Multiple studies found that online advertising has a positive effect on offline sales. Dinner et al (2014) found that the impact of advertising exists across multiple purchase channels. These are called cross-channel effects. Their research found that especially the effects from online advertising on offline sales are large. Chan et al (2011) found that spill-over effects can occur from online advertising to offline sales. Thus, customers that are acquired through online advertising influence offline sales, and positively affect customer the customer lifetime value. Dinner, van Heerde and Neslin (2011) found that online advertising is more effective than advertising via traditional media channels. This was confirmed by Lewis and Reiley (2014). They conducted a controlled experiment on search engine Yahoo. Against their expectations, they found that online advertisement has a large influence on offline sales. Wiesel, Pauwel and Arts (2011) also found that online advertisement affects all stages of the purchase funnel, and that it has a large effect on offline sales.

2.1.2.2 Online video advertising

(12)

12

watch the full advertisement. The data used in this research distinguishes both kind of advertisements: advertisements on YouTube and RTL XL (a Dutch online video platform). More than 85 percent of all YouTube advertisements are skippable advertisements that can be skipped after five seconds (YouTube, 2015). RTL only supports non-skippable advertisements that last for at least fifteen seconds.

Although much research has been conducted on the effect of online advertisement on offline sales, less research has been conducted on the specific influence of online video advertisement. The research of Dehgani and Niaki (2016) is the only study that addresses this topic. They found that YouTube advertising has a positive effect on consumers purchase intentions. Li and Lo (2015) found that online video advertising has a positive effect on brand recognition. No research has been done on the influence of (non-)skippable ads on (offline) sales or purchase intention. However, several studies have been focusing on the difference in effectiveness of both types. Handayani and Hudrasyah (2015) found that skippable ads could decrease negative impact and increase the effectiveness of the ad because viewers are more involved on clicking, and thus create more brand awareness. They also found that non-skippable ads are less effective because viewers are forced to watch the whole ad, and therefore tend to avoid the ad by refreshing the page or divert their attention. Pashkevich et. al (2012) did a mayor study on (non-)skippable ads on YouTube. Although this study was done by employees of Google and therefore could cause bias because of conflicting interests, no other research was actually done on this scale on the YouTube platform. Because the article was also placed in a leading scientific journal, it can be safe to assume the results are unbiased. The researchers found that skippable advertisements were perceived eight times better than non-skippable advertisements. Furthermore, they found that skippable ads increased advertising effectiveness and user engagement. It is therefore interesting to make a distinction between both kinds of advertisements.

(13)

13

consists of skippable adds, it is expected that YouTube advertisement has a higher effect on sales than RTL. Based on this the following hypotheses can be derived:

H1b: Online video advertisement positively effects offline sales H1c: Skippable video advertisements positively effects offline sales H1d: Non-Skippable video advertisements positively effects offline sales

H1e: Skippable advertisements have a higher effect on offline sales than non-skippable advertisements

2.2. Synergy effects

Because media synergies are a relatively new phenomenon, much research has been done on this subject in recent years. As mentioned earlier, synergy effects are defined as the effect that emerges when the combined effect of two media exceeds their individual effects on the measure of outcome (Naik and Peters, 2009).

(14)

14

However, most of the studies mentioned above were performed on a wide variety of online channels, including Google AdWords, Google Banners and Email. Few studies have focused solely on synergy effects between online video advertising and television advertising.

Logan (2013) performed a study on consumer attitude towards advertising on Online Streaming Television (OST), which includes platforms like YouTube and RTLXL. He found that the higher entertainment value on OST does not positively consumers attitude towards advertisement. However, as YouTube is increasingly developing itself as an original content platform, we can also view it as a form of Social Media. Kumar et al (2016) showed that advertising on Social Media shows synergy with other media channels, including television, as long as the message is the same across all channels. Lim et al (2015) found strong synergetic effects between television and online video advertisement. Their research showed that respondents exposed to repetitive video advertisements on television and online channels perceived better message, ad and brand credibility than those exposed to a single medium. Their research also showed that consumers attitude towards buying increased heavily when perceiving advertisements through multiple channels.

Synergy effects can also occur within media channels, which are called within-media synergy effects. These occur between two types in the same channel. Naik and Peeters (2009) found that within-media synergies also occur between online channels or offline channels. But their research focused mainly on all sorts of advertising on online channels, including. No research has been done on the synergy effects between online video advertisement. Furthermore, no research has focused on the effects between skippable and non-skippable advertisements. However, Varan et al (2013) found strong synergetic effects on video advertisement between multiple devices. Their results show that video advertisements viewed on mobile, laptop or pc strengthen the effect on advertising effectiveness and brand recognition. In combination with the findings in the research of Naik and Peeters, this could be an indication that synergy effects between RTL and YouTube could occur.

(15)

15

difference between skippable and non-skippable advertisements, this research will elaborate more on this subject. Based on literature, it is expected that there is a positive synergetic effect between YouTube and RTL. The following hypotheses were derived:

H2a: Exposure to Online- and Offline video advertisements has a positive synergetic effect on offline sales.

H2b: Exposure to skippable and non-skippable advertisements has a positive synergetic effect on offline sales.

2.3 Age

Advertisements and media usage vary highly between age groups. Williams and Page (2011) distinguish six types of generations that all react differently on advertisements: Pre-Depression Generation, Pre-Depression Generation, Baby Boomers, Generation X, Generation Y (Millennials), and Generation Z. They emphasize that marketing managers need to understand that these demographics influence the way consumers perceive an advertisement. These differences could also influence the way in which consumers perceive video advertisements. Furthermore, there are big differences in media usage across generations. According to a report from Nielsen (2016), Millennials and Generation Z are driving the trend in online video while older generations still mainly use television. The content that is being watched also differences greatly among generations.

(16)

16

more goal oriented – i.e. seeking more entertainment - than younger generations. Advertisements is therefore more disturbing. For younger generations, it is the other way around. They are more disturbed by online advertisements because they are more goal oriented on the internet and therefore perceive online ads more negative. These conclusions were also drawn by a large report from market research company Kantar Milward-Brown (2016). They found that this holds for both skippable and non-skippable pre-roll advertisements.

According to the literature above, it can be concluded that there is a difference between age groups. Younger generations are likely to be more negative towards online video advertisements than older generations. It is assumed that this holds for both skippable and non-skippable advertisements. Furthermore, no research has been done on the effect on cross-channel or within synergy effects. Mckay-Nesbitt, Manchanda, Smith and Huhmann (2009) found that there are significant differences in the way that younger people process their emotions while watching ads compared to older people. According to their research, older people are easier to persuade and have a more positive stance toward advertising than younger people. Therefore, the assumption is made that age will have a positive influence on the synergy effects between online and offline advertisements. Thus, based on the findings of the literature, the following hypotheses can be derived:

H3a: Age has a positive influence on the effect of Skippable advertisements to offline sales.

H3b: Age has a positive influence on the effect of Non-Skippable advertisements to offline sales.

H3c: Age has a positive influence on the effect of Online advertisements to offline sales. H3d: Age has a negative influence on the effect on offline video advertisement to offline

sales.

(17)

17

An overview of all hypotheses in this research can be found in table 1.

Hypothesis

H1a: Offline video advertisement positively effects offline sales

H1b: Online video advertisement positively effects offline sales

H1c: Non-Skippable video advertisements positively effects offline sales

H1d: Skippable video advertisements positively effects offline sales

H1e: Skippable advertisements have a higher effect on offline sales than non-skippable

H2a: Exposure to Online- and Offline video advertisements has a positive synergetic effect on offline sales

H2b: Exposure to skippable and non-skippable advertisements has a positive synergetic effect on offline sales

H3a: Age has a positive influence on the effect of Skippable advertisements to offline sales

H3b: Age has a positive influence on the effect of Non-Skippable advertisements to offline sales

H3c: Age has a positive influence on the effect of Online advertisements to offline sales

H3d: Age has a negative influence on the effect on offline video advertisement to offline sales

H3e: Age has a positive Influence on the synergetic effects of online and offline video advertisements on offline sales

(18)

18

3. Conceptual model

The hypotheses that were derived in Chapter 2 are visualized in the Conceptual model of figure 1. The model encompasses the effect of RTL (Non-Skippable ads), YouTube (Skippable ads) and television advertisement on offline sales. It also accounts for the combined effect of RTL and YouTube, measured as Online Video Advertising. Furthermore, synergy effects between YouTube and RTL as well as between online and offline video advertising are also present. Age is included as a moderation variable. Furthermore, Income and Household size serve as control variables.

(19)

19

4. Methodology and Data Exploration

This chapter elaborates on the methodology that will be used. First the data will be explored and the dependent and independent variables will be described. After that the models will be specified and further explained.

4.1 Data exploration

4.1.1 General overview

The dataset was provided by market research company Gfk. The data consists of household data on the sales and advertising of a large brand in the beverage industry. The panel was composed in order to form an accurate representation of Dutch households. The data has been provided on the condition that the brand will not be named because of competitor sensitive information. Therefore, the brand will be called ‘Brand A’ from now on. The dataset consists of daily household data of circa 10.000 households in The Netherlands over the course of 90 days, from the 30th_{of December 2013 until the 29}th_{of March 2014. The dataset}

(20)

20

The dataset was cleaned by removing all households that did not participate in the online panel or the television panel. Furthermore, only households that participated through the whole period were included. There were no missing’s in the dataset, except for some households that did not contain a value for age. Because there were only missing’s for 13 households and because age is an important variable in this research, these were also deleted from the dataset. Therefore, the dataset consisted of 5678 households.

4.1.2 Sales

Sales, the dependent variable, consists of purchase data in the period of 13 weeks. The sales is based on offline purchases in supermarkets. The products of Brand A are sold in different types of packages with different prices. Therefore, the total sales in Euro’s is calculated by multiplying the price per unit by the number of units bought. As expected for an FMCG product, the sales in the period was relatively stable but there was variance over the weeks. The sales per household per week varied from € 0, - to € 47,98.

Figure 2: Sales for Brand A

(21)

21

4.1.3 Media exposure

The dataset contained the exposures of three different channels: Television, YouTube and RTL. The latter two were calculated based on the amount of times a panel member came in to contact with the pre-roll (advertisement). As mentioned earlier, television contacts were measured actively via a survey. In order to make the data more reliable, the TV variable was rounded to zero if a value was lower than 0.5. This was done under the assumption that the chance that an ad was seen if it was watched. Furthermore, in order to perform a correct analysis on all online contacts, a new variable named ‘Online Contacts’ was constructed by adding the number of contacts with YouTube and RTL advertisements per customer. As can be seen in table 2, advertisements on YouTube and RTL started in week 6. All numbers are the total amount of contacts with the medium. Television advertisement was active during the whole period, except in week 7, week 8 and week 13. RTL advertisement was active during week 6 until week 12. YouTube advertisement was active during week 6 until week 13. Furthermore, YouTube advertisements were seen on average 0.009 times per week, with a minimum of 0 and a maximum of 3.08 per household. Households saw RTL advertisements on average 0.005 times per week, with a minimum of 0 and a maximum of 2.54. Households saw television advertisements the most - on average 0.78 times per week – with 0 as a minimum but up to 12.38 times. The reason these numbers are not integers is because the dataset contained data on a daily basis, but was aggregated to a weekly basis. All descriptives can be found in Table 3.

Week 1 2 3 4 5 6 7 8 9 10 11 12 13

YouTube 0 0 0 0 0 125 134 101 137 127 125 123 112

RTL 0 0 0 0 0 115 43 38 85 114 56 2 0

TV 5500 4985 3320 3738 5538 3411 0 0 7468 4285 5039 1350 0

(22)

22

N Min. 1st Qu. Median Mean 3rd Qu. Max. St. Dev.

YouTube 73814 .00 .00 .00 0.0094 .00 3.08 0.102

RTL 73814 .00 .00 .00 0.0057 .00 2.54 0.074

Online 73814 .00 .00 .00 0.0151 .00 3.62 0.129

TV 73814 .00 .00 .00 0.7809 1.2 12.38 0.974

Table 3: Descriptives of exposure on Media Channels

4.1.4 Age

(23)

23

Figure 3: Amount of panel members per age group

4.2 Model specification

(24)

24

advertisement. In model 2, age is a moderator on the direct effects of online and offline advertisement, and on the interaction effects of these two variables. Furthermore, two control variables are added to both models. These are Household Income and Household Size.

Model 1

Sales

t

= α + β

1

Age

i +

β

2

YT

it

+

β

3

RTL

it

+

β

4

TV

it

+ β

5(

TVit *

Age

i

)

+

β

6(

RTL

it

*

YT

it) +

β

7(

RTL

it

* Age

i)

+ β

8(

YT

it

* Age

i

) + β

9(

Income

i

) + β

10(

HHSize

i

) + 𝜀

_𝑖𝑡

Model 2

Sales

t

=

α + β

1Agei + β2ONLINEit + β3OFFLINEit + β4(OFFLINEit * ONLINEit) + β5(OFFLINEit * Agei) + β6(ONLINEit * Agei) + β7(OFFLINEit * ONLINEit * Agei) + β8 (Incomei) + β9 (HHSizei) + 𝜀_𝑖𝑡

Where:

•

α =

Constant (Intercept); •

β =

Parameter;

•

t =

The week number ranging from Week 1 – Week 13; •

i =

The Household ID;

• Salest = Sales in Euro’s at time t;

• TVit

=

Exposure to television advertisement for household i at time t;

• RTLit

=

Exposure to RTL online video advertisement for household iat time t; • YTit

=

Exposure to YouTube online video advertisement for household iat time t; • ONLINEit = Exposure to an online video advertisement for household i at time t; • OFFLINEit = Exposure to an offline video advertisement for household i at time t; • Incomei = Income of Household i;

(25)

25

5. Results

In this chapter, the results of the analysis will be presented. First, the outcomes of the models will be evaluated. Next, the model quality will be discussed and the model will be validated. Third, the outcomes of the analysis will be discussed and the hypotheses will be answered.

5.1 Model estimation

The first model that was estimated was the model as described in section 4.2, with sales as dependent variable and YouTube, RTL and TV as independent variables, with Age as mediator, and Household Size and Income as control variables. The second model was estimated the same, but with YouTube and RTL combined as Online Advertisement. Both model 1 and 2 were significant (p < .01) with the same Adjusted R-square of 0.0032. However, none of the variables were significant except for the demographic variables Age, Income and Household Size. All outcomes can be found in Appendix 3 and 4. Furthermore, the VIF scores of some variables in this model had high VIF scores. This is a strong indication of multicollinearity.

In order to increase the strength of the models, the advertisement variables (YouTube, RTL and TV contacts) were transformed to Adstock variables. This way, the model also accounts for advertisements in the weeks previous to a sale. This is not uncommon in marketing research, because advertising can have a long-term effect on sales (Ataman, van Heerde and Mela, 2010). A decay factor of 0.54 was chosen because this is a common used value in marketing research, and is based on Leone’s half time value (Leone, 1995). This means that the Adstock value of a previous week has a carryover effect of 54%. For the purpose of readability, the variable names in the models have not been changed.

(26)

26

variables and the advertising variables with VIF scores of 15 and higher. To solve this, the age variable was mean centered. After that, the VIF scores did not surpass 1.9, meaning that none of the variables had to be removed. The VIF scores for the variables in both models can be found in table 4 and 5.

Model 1 B Std. Error t p-value VIF

(Intercept) 0,219 0,020 10,729 0,000 YouTube 0,120 0,081 1,471 0,141 1,41 RTL -0,109 0,107 -1,017 0,309 1,28 TV -0,020 0,007 -2,838 0,005 1,01 Age -0,019 0,004 -4,579 0,000 1,61 Income 0,013 0,001 8,600 0,000 1,02 HH Size 0,044 0,006 7,300 0,000 1,11 YouTube * RTL 0,086 0,307 0,280 0,779 1,20 YouTube * Age 0,083 0,038 2,180 0,029 1,43 RTL * Age 0,013 0,042 0,301 0,763 1,24 TV * Age -0,003 0,004 -0,800 0,424 1,51

Table 4: Summary of model 1, Confidence level 95%.

Model 2 B Std. Er. t p-value VIF

(Intercept) 0,219 0,020 10,688 0,000 Online 0,101 0,075 1,344 0,179 1,940 Offline -0,019 0,007 -2,677 0,007 1,030 Age -0,019 0,004 -4,563 0,000 1,620 Income 0,013 0,001 8,597 0,000 1,020 HH Size 0,044 0,006 7,300 0,000 1,110 Online * Offline -0,090 0,057 -1,569 0,117 1,640 Online * Age 0,065 0,033 1,979 0,048 1,930 Offline * Age -0,003 0,004 -0,807 0,420 1,530

Online * Offline * Age -0,018 0,030 -0,593 0,553 1,620

Table 5: Summary of model 2, Confidence level 95%.

(27)

27

models, the test outcomes showed that there is autocorrelation (Model 1: DW = 0.99, p < .00; Model 2: DW = 0.99, p <.00). To resolve this, the models were re-estimated by generalized leased squares (GLS). Applying this on model 1 increased the Akaike Information Criterion (AIC) from 30,3821.8 to 303749.8, meaning that the quality of the GLS-model is better. The AIC of model 2 also dropped from 30,3821.8 to 30,3748.5, which also means that the quality of the GLS-model was improved.

Furthermore, both models were checked for nonnormality. If normality is not satisfied, the p-values can’t be trusted. In order to detect nonnormality, a Kolmogorov-Smirnov test was performed. The outcome showed for both models a significant result (Model 1: p < .01; Model 2: p < .01), indicating that there is nonnormality. In order to solve this, bootstrapping was done. This does not affect the interpretation of the Beta’s in the model, but gives more reliable confidence intervals. Bootstrapping had impact on the confidence interval of some variables in both the models. There were no variables that were significant before the bootstrap, but insignificant after the bootstrap. However, there were insignificant variables that became significant after bootstrapping. For Model 1, the interaction effect between Online advertisement and Offline advertisement was significant (p < .05) as well as the moderation effect of Age on Online advertisement (p <.05). For Model 2, the moderation effect of age on RTL and TV also became significant at a 5% level. The re-estimated confidence intervals can be found in Appendix 4 and 5.

5.2 Model quality and hypothesis testing

(28)

28

5.2.1 Direct effects

The direct effects of Offline video advertisement and Online video advertisement can be seen in Model 2, while the direct effects of Skippable and Non-Skippable advertisements can be found in Model 1. It must be noted that ‘direct effects’ in this section must be interpreted as Adstock variables, thus including lagged effects. First, Offline video advertisement is significant (p <.05) with a Beta of -0.019. This means that an increase of one contact with an offline advertisement, in this case TV advertisement, decreases the sales of Brand A with almost € 0,02. This is against the expectation that offline video advertisement has a positive effect on sales. Therefore, hypothesis 1a must be rejected. Although the Beta of Online Video Advertisement was positive, the effect was insignificant. This means that no evidence was found that there is a relationship with the sales of Brand A. The expectation was that online video advertisement would have a positive effect on sales. Therefore, hypothesis 1b cannot be accepted. Furthermore, the same applies for Skippable and Non-Skippable advertisements. Both were found to be insignificant. The expectation was that both would have a positive effect on sales, and that Skippable advertisements would have a higher effect than Non-Skippable advertisements. However, since the effects are insignificant, hypothesis 1c, 1d and 1e are not accepted.

5.2.2 Synergetic effects

(29)

29

hypothesis 2b was that there would be a positive synergy effect, but this hypothesis cannot be accepted.

5.2.3 Influence of age on direct and synergetic effects

In order to answer hypotheses 3, the moderation effects of age on the independent variables must be significant. The moderation effects of age on Skippable and Non-Skippable advertisements can be seen in Model 1, while the moderation effects on Online, Offline and on the synergetic effects between them can found in Model 2. First it is important to point out that the age variable was mean centered around 7.6. The variable thus has a mean of zero, a minimum of -6.63 and a maximum of 3.37. That means that when the value of age is 0 this is actually age group 7 which consists of people between 45 and 49, while for example an age variable of 3.37 means it is age group 11, which consists of people older than 75. An overview of values per age group can be found in Appendix 1.

(30)

30

When looking to the results of Model 2, the influence of age on Offline advertisement was found to be insignificant. This is surprising since the effect of age on TV in Model 1 is significant (p < .05) and both are composed of the same values. Therefore, the outcomes of age on TV advertisement in the first model will be interpreted. The Beta of Age on TV advertisement was found to be -0.003, which means that for people between 45 and 49 one contact with a TV advertisement would lead to a decrease of € 0,003 in sales. This is less than one Eurocent and would therefore not have a big impact on sales for this age group. However, the outcome is still an indication of the effect of television advertisement because for older age groups this effect would increase. Furthermore, for younger people this would lead to an increase in additional sales. For example, for the age group between the age of 20 and 24 this would lead to an increase of € 0,02 in additional sales. The expectation was that age would have a negative influence on the effect of offline video advertisement. The findings confirm this expectation and therefore hypothesis 3d is accepted. Furthermore, the influence of age on Online advertisements was found to be significant (p < .05) with a Beta of 0.065. This again indicates that for people between 45 and 49 who come into contact with an online advertisement would create an additional € 0,07 in sales. For younger people a contact with an online advertisement would create an increasingly negative value in additional sales. This confirms the expectation that age would have a positive influence on sales. Therefore, hypothesis 3c is accepted.

(31)

31

5.2.4 Summary of results

The following table gives an overview of the results:

Hypothesis Supported?

H1a: Offline video advertisement positively effects offline sales Rejected

H1b: Online video advertisement positively effects offline sales Not Accepted

H1c: Non-Skippable video advertisements positively effects offline sales Not Accepted

H1d: Skippable video advertisements positively effects offline sales Not Accepted

H1e: Skippable advertisements have a higher effect on offline sales than non-skippable Not Accepted H2a: Exposure to Online- and Offline video advertisements has a positive synergetic effect on

offline sales.

Rejected H2b: Exposure to Skippable- and Non-Skippable video advertisements has a positive

synergetic effect on offline sales

Not Accepted h3a: Age has a positive influence on the effect of Skippable advertisements to offline sales. Accepted H3b: Age has a positive influence on the effect of Non-Skippable advertisements to offline

sales.

Accepted H3c: Age has a positive influence on the effect of Online advertisements to offline sales. Accepted H3d: Age has a negative influence on the effect on offline video advertisement to offline

sales.

Accepted H3e: Age has a positive Influence on the synergetic effects of online and offline video

advertisements on offline sales.

Not Accepted

(32)

32

6. Discussion

The goal of this research was to provide new insights into the topic of video advertisement. With the rise of Social Media, online video advertisement has become increasingly important, while television advertising is still proving to be an effective channel. With the fragmentation of online channels, each with its own audience, the question for managers is how to form an effective marketing strategy for all age groups. The question that this research set out to answer was “What influence does age have on the effect of online and

offline video advertising on sales?”. The question was answered by looking at the direct

effects of advertising channels on Sales of ‘Brand A’, possible synergy effects – between and in-between channels – and the moderating effect of Age on these effects.

(33)

33

television advertising have changed because of the shift in media usage. Further research on this topic is therefore needed.

No evidence was found that synergy effects occur between online and offline video advertisements or between Skippable and Non-Skippable advertisements. There could be several reasons for this. First, using multiple media channels does not mean that a synergy effect occurs. It is also important that an integrated marketing communication campaign is created in which all channels cooperate (Winer, 2009). Furthermore, as mentioned earlier, the amount of contacts with online advertisements was quite low. More contact points could have made the results more significant.

This research did find interesting results on the influence of age on video adverting. First, people older than 45 tend to react increasingly more positive on online advertisements than younger generations do. This also holds when a distinction is made between Skippable and Non-Skippable online video advertisements. This is in line with recent research of Logan (2013). Younger people are more likely to be influenced by television advertisement than online advertisement. Furthermore, age had especially a high impact on the direct effects of skippable advertisements on YouTube. This could indicate that skippable pre-roll advertisement works very effective for people older than 45. These findings would suggest that in order to form an effective video advertisement strategy on young adults, it would be better to focus on television advertisement than on online video advertisement. For older people, it would be better for companies to focus more on online advertising.

(34)

34

7. Managerial Implications

Nowadays, online advertising can no longer be ignored by large FMCG companies. With the rise of multiple online channels in which video advertisement becomes increasingly popular, manager need to know how to spend their advertisement money with the highest ROI. The outcomes of this research assume that television advertisement might no longer be the most effective video advertising platform there is. Especially because elder people tend to react much more positive towards online video advertising on platforms like YouTube. Especially skippable pre-roll advertisements seem to have a high effectiveness. Managers should also consider to rethink their strategies towards younger generations. The outcomes of this research show that younger people are negatively influenced by online advertisements, because they find them more disturbing.

This could help managers make better decisions, since many companies logically assume that younger people have more experience with online video, and more prone to be positively influenced by online advertisements. For these companies, it would be wise to invest more into television advertisement. With older generations shifting more and more towards online channels, this could also give managers an incentive to retract budget for television advertisement and invest more into online. This can even further be confirmed by the findings of this thesis, because against the findings of other research, the effects of television advertisement seem to decline.

8. Limitations

(35)

35

(36)

36

Literature

Articles

Ataman, M. B., Van Heerde, H. J., & Mela, C. F. 2010. The long-term effect of marketing strategy on brand sales. Journal of Marketing Research, 47(5): 866-882.

Campbell, C., Pitt, L. F., Parent, M., & Berthon, P. R. 2011. Understanding consumer conversations around ads in a Web 2.0 world. Journal of Advertising, 40(1): 87-102.

Chan, T. Y., Wu, C., and Xie, Y. (2011). Measuring the lifetime value of customers acquired from Google search advertising. Marketing Science, 30(5), 837-850.

Cheben, J. 2014. Effectiveness of TV Advertising when Targeting Generations Y and X. Metropolitan University Prague.

Danaher, P. J., and Dagger, T. S. (2013). Comparing the relative effectiveness of advertising channels: A case study of a multimedia blitz campaign. Journal of Marketing Research, 50(4), 517-534.

Dinner, I.M., Van Heerde, H.J. and Neslin, S., 2011. Driving online and offline sales: The cross-channel effects of digital versus traditional advertising. Tuck School of Business Working Paper: 527-545.

Hanayani, T.A. and Hudrasyah, H., 2015. Comparative Analysis of Skippable and Non-Skippable Pre-Roll Advertising on YouTube in Bandung. Journal of Business and Management, 4(6): 694-701.

Jagpal, H. 1981. Measuring Joint Advertising Effects in Multiproduct Firms. Journal of Advertising Research, 21(1): 65-69.

(37)

37

Kumar, A., Bezawada, B., Rishika, R., Janakiraman, R., and Kanna, P.K. 2015. From Social to Sale: The Effects of Firm-Generated Content in Social Media on Customer Behavior. Journal Of Marketing, 2: 1-19.

Pashkevich, M., Dorai-Raj, S., Kellar, M. and Zigmond, D. 2012. Empowering online Advertisements by Empowering Viewers with the Right to Choose. The Relative Effectiveness of Skippable Video Advertisements on YouTube. Journal of Advertising Research, 4: 451-457. Lewis, R.A. and Reiley, D.H. 2014. Online Ads and Online Sales: Measuring the Effects of Retail Advertising via a Controlled Experiment on Yahoo! Quantitative Marketing and Economics, 12(3): 235-266.

Li, H., Lo, L. 2015. Do You Recognize Its Brand? The Effectiveness of Online In-Stream Video Advertisements. Journal of Advertising, 44(3), 208-218.

Lin, C., Venkataraman, S. & Jap S.D. 2013. Media Multiplexing Behavior: Implications for Targeting and Media Planning. Marketing Science, 32(2): 310-324.

Lim, J.S., Ri, S.Y., Egan, B.D., Biocca, F.A. 2015. The cross-platform synergies of digital video advertising: Implications for cross-media campaigns in television, Internet and mobile TV. Computers in Human Behaviour, 48: 463-472.

Logan, K. 2013. And now a word from our sponsor: Do consumers perceive advertising on traditional television and online streaming video differently? Journal of Marketing Communications, 19(4): 258-276.

Naik, P.A. & Peters, K. 2009. A Hierarchical Marketing Communications Model of Online and Offline Media Synergies. Journal of Interactive Marketing, 23: 288–299.

(38)

38

Raju, S.T. 1992. The Effect of Price Promotions on Variability in Product Category Sales. Marketing Science, 11(3): 207-220.

Rethans, A.J., Swasy, J.L. and Marks, L.J., Effects of Television Commercial Repetition, Receiver Knowledge, and Commercial Length: A Test of the Two-Factor Model. Journal of Marketing Research, 23(1): 50-61.

Rubinson, J. 2009. Empirical Evidence of TV Advertising Effectiveness. Journal of Advertising Research, 49(2): 220-226.

Sethuraman, R., Tellis, G.J. and Briesch, R.A. 2011. How well does advertising work? Generalizations from meta-analysis of brand advertising elasticities. Journal of Marketing Research, 48(3): 457-471.

Smith, K.T. 2011. Digital marketing strategies that Millennials find appealing, motivating, or just annoying. Journal of Strategic Marketing, 19(6): 489-499.

Varan, D., Murphy, J.,Hofacker, C.F., Robinson, J.A., Potter, R.F. and Bellman, S. 2013. What Works Best When Combining Television Sets, PCs, Tablets, or Mobile Phones? How Synergies Across Devices Result From Cross-Device Effects and Cross-Format Synergies. Journal of Advertising Research, 53(2): 212-220.

Winer, R.S. 2009. New Communications Approaches in Marketing: Issues and Research Directions. Journal of Interactive Marketing, 23(2): 108-117.

Williams, K.C. and Page, R.A. 2011. Marketing to the generations. Journal of Behavioral Studies in Business, 3: 1-16.

(39)

39

Wiesel, T., Pauwels, K. and Arts, J., 2011. Practice Prize Paper Marketing's Profit Impact: Quantifying Online and Offline Funnel Progression. Marketing Science , 30 (4), pp.604611. Zantedeschi, D., Feit, E. and Bradlow, E., 2013. Measuring Multi-Channel Advertising Effectiveness Using Consumer-Level Advertising. University of Pennsylvania.

Websites CBS, 2016 (visited 07-04-2017) http://statline.cbs.nl/StatWeb/publication/?DM=SLNL&PA=60060ned CNBC, 2017 (visited 06-04-2017) http://www.cnbc.com/2017/03/15/adidas-steps-away-from-tv-advertising-as-it-targets-4-billion-growth.html Forbes, 2016 (visited 15-03-2017) https://www.forbes.com/sites/brandonkatz/2016/09/14/digital-ad-spending-will-surpass-tv-spending-for-the-first-time-in-u-s-history/#4b27223c4207 IAB, 2016 (visited 15-03-2017) https://www.iab.com/wp-content/uploads/2016/04/2016-IAB-Video-Ad-Spend-Study.pdf Nielsen, 2016 (visited 17-3-2017) http://www.nielsen.com/nz/en/insights/news/2016/millennials-and-generation-z-lead-the-future-of-media.html Nielsen, 2016 (visited 01-04-2017) http://www.nielsen.com/us/en/insights/news/2016/their-generation-from-location-to-listening-habits-a-media-divide-exists.html

Kantar Milward-Brown, 2016 (visited 05-06-2017) http://www.millwardbrown.com/adreaction/genxyz/ Variety, 2016 (visited 01-04-2017)

(40)

40

YuMe, 2016 (visited 05-06-2017)

(41)

41

Appendix

1. Overview of age groups

Age Group Age Mean centered

1 12-19 year 3.370 2 20-24 year 2.370 3 25-29 year 1.370 4 30-34 year 0.370 5 35-39 year -0.629 6 40-44 year -1.629 7 45-49 year -2.629 8 50-54 year -3.629 9 55-64 year -4.629 10 65-74 year -5.629 11 75 year or older -6.629

2. Model 1 without Adstock Variables

(42)

42

3. Model 2 without Adstock Variables

B Std. Error t value p (Intercept) 0,338 0,040 8,544 0,000 Online contacts -0,270 0,250 -1,079 0,281 Offline contacts 0,056 0,036 1,563 0,118 Age -0,018 0,004 -4,720 0,000 Income 0,013 0,001 8,696 0,000 HH Size 0,044 0,006 7,326 0,000

Online Contacts * Offline contacts 0,072 0,366 0,197 0,844

Online Contacts * Age 0,041 0,037 1,108 0,268

Offline contacts * Age -0,006 0,004 -1,291 0,197

Online Contacts * Offline contacts * Age -0,017 0,048 -0,363 0,717

4. Model 1 Confidence intervals after bootstrapping

B Std.Error p-bootstrap p-value

(Intercept) 0,219 0,021 0,000 0,000 YouTube 0,120 0,094 0,182 0,141 RTL -0,109 0,073 0,126 0,309 TV -0,020 0,006 0,002 0,005 Age -0,019 0,004 0,000 0,000 Income 0,013 0,001 0,000 0,000 HH Size 0,044 0,006 0,000 0,000 YouTube * RTL 0,086 0,216 0,364 0,779 YouTube * Age 0,083 0,035 0,026 0,029 RTL * Age 0,013 0,019 0,032 0,763 TV * Age -0,003 0,003 0,025 0,424

5. Model 2 Confidence intervals after bootstrapping

B Std.Error p-bootstrap p-value

(43)

43

Offline * Age -0,003 0,003 0,270 0,420

The influence of Age on synergetic and direct effects of Online and Offline video advertisements