Perfectly Imperfect: A study on the relation between Brand Prominence in Native Advertising and Word of Mouth, mediated by Brand Attitude and moderated by Message Sidedness

Perfectly Imperfect:

Perfectly Imperfect:

A study on the relation between Brand Prominence in Native Advertising and Word of Mouth, mediated by Brand Attitude and moderated by Message Sidedness

By

ROSANNE NOORDA

University of Groningen Faculty of Economics and Business

Department of Marketing

Master Marketing Thesis

Supervisor: dr. J.C. Hoekstra Second Supervisor: dr. J.A. Voerman

Onderduikersstraat 72 9727 CX GRONINGEN

TABLE OF CONTENTS

ABSTRACT 3

PREFACE 4

1. INTRODUCTION 5

2. CONCEPTUAL MODEL AND HYPOTHESES 8

2.1 Conceptual Model 8

2.2 Hypotheses 8

3. METHODOLOGY 13

3.1 Research Setting, Pretest and Sample 13

3.2 Measurement scales 15

3.3 Factor analysis and Reliability analysis 17

3.4 Analyses 18

4. RESULTS 19

4.1 Control variables and manipulation check 19

4.2 Hypotheses 19

4.3 Additional analysis 23

5. DISCUSSION 24

5.1 Conclusions 24

5.2 Managerial implications 26

5.3 Limitations and future research 27

LITERATURE 28

ABSTRACT

This study investigates native advertisement effectiveness using different levels of brand prominence and message sidedness, regarding brand attitude and Word of Mouth. Data retrieved from 160 respondents in an online 2x2 in between subjects experiment show that, not in line with the hypotheses, high brand prominence combined with a two-sided message does not result in the highest ad effectiveness in terms of brand attitude and Word of Mouth. Contrary, high brand prominence combined with a one-sided message in native advertisements create the highest brand attitude and Word of Mouth. One-sided messages in native advertising significantly elicit more favorable brand attitudes than two-sided messages. However, the most effective level of brand prominence is not certain, although clear brand prominence is essential to create marketing effects. These results provide insights for managers and ad agencies that could increase the effectiveness of online native advertising.

PREFACE

In front of you is the thesis “Perfectly Imperfect: A study on the relation between Brand Prominence in Native Advertising and Word of Mouth, mediated by Brand Attitude and moderated by Message Sidedness”. With this thesis I hope to successfully complete my master in marketing management at the University of Groningen. The main topic of this thesis is digital advertising, which is an important marketing tool for business’ nowadays. However, it is important that business’ implement this correctly, since image damage and unsatisfied customers is not desired. Therefore, I thought it would be interesting to study what the antecedents are for the ‘perfect’ advertisement. In this study I focused on discovering the right amount of brand prominence, and the right type of message to achieve the highest possible brand attitude and customer engagement in the form of Word of Mouth. The message type I investigated is messages sidedness, where a two-sided message reveals disclaimers or negative attributes about a brand or product in the advertisement. This explains the title: creating the

perfect advertisement, by admitting that your company is imperfect. I hope that the insights of

this thesis are helpful and inspire other researchers and students to continue studying this subject.

I would like to show my gratitude to dr. J.C. Hoekstra, who was my first supervisor. Through her critical feedback and willingness to help, I developed my research skills and am I able to say that I am proud of this thesis. In addition, I would like to thank my thesis group, who were always attainable to answer my questions. Lastly, I would like to thank my family, boyfriend and friends, for the moral support they gave me and for always keeping me motivated.

I hope you enjoy reading my thesis. Rosanne Noorda

1. INTRODUCTION

Trends, developments, statistics, charts: nowadays, everything can be found about digital advertising online. Every year the numbers grow. Only in 2019, digital ad spending rose by 17.6% to $333.25 billion worldwide. With this number, digital advertising is the fastest growing advertising type (Enberg, 2019). However, consumers don’t seem to like this expansion of, for example, banner ads, because privacy concerns are rising (Cleff, 2007) as well as the amount of ad blocking users (Statista, 2019).

One way that advertisers use to get around the ad blockers is native advertising (Krouwer, 2019). Native advertising is a type of advertising, mostly online, that matches the form and function of the platform upon which it appears (Wojdynski & Evans, 2016). Native advertising is growing substantially, and research estimates that native advertising revenues will reach almost 75% of the U.S. ad market by 2021 (Fullerton, 2018). The fast growth of native advertising can be explained through its high level of integration with the content of its host publisher (Wojdynski, 2016). The thought of current literature regarding native advertising is that low brand prominence is the power of native advertising: no logos in the ads, so it is not really clear that it is an ad. It feels like ‘normal’ content to consumers, who will evaluate the ads as less annoying and share it more often, which creates bigger reach than traditional digital advertising (Fullerton, 2018). This results in positive effects on consumer evaluation towards the brand, because it is evaluated as less irritating and annoying (Reijmersdal, 2009). Through this advantage, advertisers acknowledge digital native advertising as an effective advertising tool and predict that they will continue to assign more of their budgets to this type of advertising (Harms, Bijmolt & Hoekstra, 2017).

Besides, low brand prominence in native ads could even arise negative consumer brand evaluation, since consumers might feel that companies ‘secretly’ try to convince them of something, while banners, on the other hand, are honest about advertising (Campbell & Marks, 2015).

To face these problems surrounding native advertising, the phenomenon brand prominence was introduced to the world of native advertising. High brand prominence indicates high brand visibility in ads whereas low brand prominence indicates low brand visibility in ads. The level of brand prominence depends on the exhibition and position of the brand name, logo, or URL (Wojdynski & Evans, 2016). High brand prominence, so less secretive native advertising, increases the effectiveness of the ad in the long run. The ad is easier to recognize, so persuasion is likely (Campbell & Marks, 2015). This seems the solution for the counter arguments of native advertising. However, research also found negative results by stating that high brand prominence could lead to activation of the consumer’s persuasion knowledge, which could lead to negative consumer evaluations (Reijmersdal, 2009).

Consumers will again gain a negative attitude towards the brand, since the brand name is shown clearly, which means ‘it is still an ad’.

However, native advertising is different is such a way that there is space to explain more. First admitting a mistake and then explaining why, is the power of message sidedness, which increases positive brand attitudes.

This study tries to extend literature on brand prominence in native advertising. With this study, we want to find out whether the attitude and engagement towards brand prominence in native advertising can be improved by combining it with message sidedness, to create positive brand attitudes and increase WoM. Besides, this investigation creates knowledge for companies and brands that want to advertise digitally in the nearby future. This research could help when creating a native digital ad, since a company or brand will know the level of brand prominence it should add and what kind of message should be used in a native advertising.

2. CONCEPTUAL MODEL AND HYPOTHESES

2.1 Conceptual Model

The conceptual model depicted in figure 1.1 assumes that using message sidedness in native advertising as a moderator, turns the negative relation between brand prominence and brand attitude into a positive one, which results in WoM. Besides, the model assumes that message sidedness strengthens the relation between brand attitude and WoM.

FIGURE 1.1 CONCEPTUAL MODEL

In addition, the variables that will be controlled for, are: age, gender, process involvement, product involvement and source attitude.

2.2 Hypotheses

Brand Prominence in Native Advertising and Brand Attitude

Native advertising is a type of advertising, mostly online, that matches the form and function of the platform upon which it appears (Wojdynski & Evans, 2016). It is common for native advertising to include a low level of brand prominence, which means the brand is not or barely visible in the ad. The level of brand prominence depends on the exhibition and position of the brand name, logo, or URL. Therefore, high brand prominence indicates high brand visibility in ads whereas low brand prominence indicates low brand visibility in ads (Wojdynski & Evans, 2016).

Brand attitude is a necessary communication effect if brand purchase is to occur (Percy & Rossiter, 1992; Nicholls, 1999). It is critical to the understanding of an effective advertising strategy (Percy & Rossiter, 1992; Gardner, 1985).

Brand

Brand attitude is an ad goal which incorporates long-run reactions to ads, such as consumers’ perceptions of quality, value or satisfaction (Gardner, 1985). These factors will be used to measure brand attitude in this research (Du, Joo & Wilbur, 2019).

Expected is that brand prominence has a negative influence on brand attitude. This is expected, since high brand prominence could lead to activation of the consumer’s persuasion knowledge, which could lead to negative consumer evaluations (Reijmersdal, 2009). Whenever this happens, a consumer becomes aware of an ad trying to convince them, which results in a negative attitude. An example of this are banner ads, where consumers can clearly recognize a company in advertising. It is clear to the consumer which brand is advertising what, so the information is easily processed in the mind. This makes it easy to arise negative feelings towards that brand (Reijmersdal, 2009), since consumers nowadays cannot see or do anything without companies trying to convince them of something. The brand prominent ad could arise negative feelings, which extend to feelings towards the advertiser, since attitude toward the advertisement affects attitude toward the advertised brand (Gardner, 1985). This could be explained by the fact that brand prominence evokes cognitive responses (Reijmersdal, Rozendaal & Buijzen, 2012). This refers to beliefs, evaluations, and attitudes towards the brand.

H1: Brand Prominence negatively influences Brand Attitude

Brand Attitude and Word of Mouth

WoM is a metric for the customer-initiated customer engagement of a brand or company (Shah, Rust, Parasuraman, Staelin & Day, 2006). WoM refers to informal communications between private parties concerning evaluations of goods and services, rather than formal complaints to firms (Anderson, 1998). Examples of WoM are recommending products and/or services to friends and family, or sharing experiences online. The tone of this can be positive or negative. Positive WoM occurs when good news testimonials and endorsements desired by the company are uttered. It is worth noting that what is negative from a corporate viewpoint may be regarded as positive from a consumer viewpoint (Buttle, 1998).

Literature supports the proposition that there is a relation between brand attitude and WoM, but mostly the other way around. A lot of research has been done on the effect of WoM on brand attitude (Herold, Tarkiainen & Sundqvist, 2016; Kudeshia & Kumar, 2017; Pongjit & Beise-Zee, 2015; Kim, Kim & Lee, 2019; JunSang Lim, & Beatty, 2005). WoM from other’s influences one’s brand attitude. Therefore, this relation exists, which means the variables are related.

The relation the other way around, so brand attitude on WoM, is expected in this study (see figure 1.1). Research confirms that customer satisfaction, which is one of the antecedents of brand attitude, has a positive effect on WoM (Zhang, Ma & Wang. 2019; Tripathi, 2017; Wangenheim & Bayón, 2007). Customer satisfaction is an important predictor of customers behavioral intentions, so it can be inferred that (highly) satisfied customers as a result of their internal evaluation are more inclined to spread WoM than those who are less satisfied (Konuk, 2019).

Another construct that is related to customer engagement, is customer brand value (Pansari & Kumar, 2017). This is defined as the differential effect of, among others, a customer’s brand attitude (Kumar, Luo & Rao, 2015). This supports the proposition of brand attitude having a relation on WoM, since it is a metric for customer engagement.

H2: Brand Attitude positively influences Word of Mouth

Mediation of Brand Attitude

Proposed is that there is a relation between brand prominence and WoM, through the mediation of brand attitude. This can be explained through the three stages of consumer responses that a consumer, unconsciously, goes through after exposure to an ad: cognitive responses, affective responses, and behavioral responses (Fennis & Stroebe, 2015). This means that after exposure to an ad, one will evaluate and create an attitude towards the ad (cognitive response), followed by creating positive or negative emotions (affective response) towards the ad. Lastly, one will act (behavioral response), which could for example be buying or spreading WoM.

Researchers that studied consumer responses of ads confirm that ads generated evaluative outcomes (brand attitude) followed by behaviour (WoM) (Eelen, Rauwers, Wottrich, Voorveld & van Noort, 2016; Olney, Holbrook & Batra, 1991), so a direct relation between brand prominence and WoM is not plausible. Brand prominence is expected to create the attitude, since evaluating advertisements is the first thing consumers, unconsciously, do in their brain when they are exposed to an ad (Reijmersdal, 2009). Brand attitude is expected to have a relation on WoM, since (highly) satisfied customers as a result of their internal evaluation are more inclined to spread WoM than those who are less satisfied (Konuk, 2019). This can be explained through the success factors for WoM, which are, for example, brand image (Rageh & Spinelli, 2012) and customer satisfaction (Zhang, Ma & Wang. 2019; Tripathi, 2017; Wangenheim & Bayón, 2007).

Furthermore, conversations that are “advertising influenced” are significantly more likely to involve recommendations to buy or try a brand, so to spread WoM, when compared with other WoM discussions about brands. However, this is after one has created an attitude towards the brand (Keller & Fay, 2009), since recommending a brand would mean a positive attitude towards the brand was created.

H3: The relation between Brand Prominence and Word of Mouth is mediated by Brand Attitude

Moderation of Message Sidedness

Phrasing in advertising can be the final push to persuade a customer to respond behavioral (Danescu-Niculescu-Mizil, et al., 2012). A marketing technique that involves a certain manner of phrasing is message sidedness. A one-sided message is characterized by favorable claims on all brand attributes mentioned in the message, whereas a two-sided message contains disclaimers or admissions of inferior performance on one or more relatively unimportant attributes (Hastak & Park, 1990).

It is expected that using a two-sided message in a brand prominent native advertising, will turn the negative relation between brand prominence and brand attitude into a positive relationship. Current literature in consumer research confirms the possibility of a moderator changing the sign of the relationship (Petty, Cacioppo & Schumann, 1983).

Using a two-sided message in the native ad, instead of a one-sided message, which are currently used in native advertisements, could turn this relation around. This is because two-sided messages in advertising elicit more favorable ad attitudes than one-sided messages (Kao, 2011). In addition,if consumers' prior attitude towards the brand is negative or neutral, two-sided messages are more effective in altering evaluations and changing attitudes, and in being more consistent with the receiver's attitudinal schema. They may therefore encounter less resistance (Crowley & Hoyer, 1994).

This attitude change could be explained through two theories. Firstly, the social judgement theory (Sherif, Sherif & Nebergall, 1961), which states that when the consumer’s attitude is negative, a message containing negative information may be more likely to fall into the receiver's "latitude of acceptance". This will increase the probability that the message will be processed and decrease the probability that the message will be actively rejected or counter-argued. In addition, the adaptation-level theory(Appley, 1971) states that a two-sided message may be more consistent with the receiver's attitudinal schema and may thus encounter less resistance.

H4: Message sidedness turns the negative relation between Brand Prominence and Brand Attitude into a positive relation

Besides, it is expected that a two-sided message also has an influence on the relation between brand attitude and WoM. This relation is expected to be positive, since customer satisfaction, which is one of the antecedents of brand attitude, has a positive relation on WoM (Zhang, Ma & Wang. 2019; Tripathi, 2017; Wangenheim & Bayón, 2007).

However, adding a two-sided message will strengthen this relation, because if the consumer held a favorable prior attitude while it was already aware of the negative information, message acceptance is likely (Crowley & Hoyer, 1994). This means that when a consumer already has a positive attitude before exposure to the two-sided ad, its positive attitude remains after exposure to the ‘negative statements’. This message acceptance, and thus the positive attitude, increase effectiveness of ads including two-sided messages regarding engagement and behavioural intentions, referring to WoM (Pierro, Giacomantonio, Pica, Giannini, Kruglanski & Higgins, 2013; Konuk, 2019).

3. METHODOLOGY

3.1 Research Setting, Pretest and Sample

In this research a 2 x 2 factorial design was used in a between-subject-experiment to study the influence of brand prominence and message sidedness. Four different native advertisements were designed to show various combinations: high or low brand prominent ads with a one-sided or two-one-sided message. The native advertisements were designed as native article style advertorials which were shown at a credible Dutch entertainment and news site (linda.nl) to ensure professional appearance. The website is credible since it uses credible sources in every article (examples: ANP; Max Havelaar). An existing advertorial about fairtrade chocolate on linda.nl was used (see appendix 1). The brand that was used in this experiment is a fictitious brand called Chofy. A fictitious brand was created to avoid biased responses through brand (dis)liking. The advertorial tells about the new fairtrade products of Chofy, and it makes clear that buying fairtrade is important for a fair life for African cacao farmers.

Brand prominence was manipulated by using different visuals in the advertorial. The low brand prominent ad contained a picture of chocolate in the header of the article, without any brand name or logo. No other visuals were used in this advertorial. The high brand prominent ad contained the same picture of the chocolate in the header, but the logo of Chofy was added to this picture. Besides, a photo of the ‘new’ fairtrade Chofy chocolate bars was added in the advertorial itself and the brand name of Chofy was bold in the text.

TABLE 3.1

STIMULI DESIGN MESSAGE SIDEDNESS

One-sided Two-sided

WHAT DOES A CACAO FARMER EARN FROM YOUR

(NON)FAIRTRADE CHOCOLATE You are not the only who enjoys their bar of chocolate without thinking of the African cacao farmer who made it. However, do you know how much of the value of your bar of chocolate goes to the those farmers?

40% of the value of fairtrade chocolate goes to the African cacao farmer,

which is three times as much as with non-fairtrade chocolate.

So, when you buy fairtrade chocolate, the African cacao farmer receives a higher salary! Because of this, from now on Chofy only produces fairtrade chocolate.

The new chocolate bars from Chofy are fair trade. As a result, Chofy helps farmers in developing countries to gain a better place in the trade chain, so that they can live off their work and invest in a sustainable future.

WHAT DOES A CACAO FARMER EARN FROM YOUR

(NON)FAIRTRADE CHOCOLATE You are not the only who enjoys their bar of chocolate without thinking of the African cacao farmer who made it. However, do you know how much of the value of your bar of chocolate goes to the those farmers?

Only 7% of the value of non-fairtrade chocolate goes to the cocoa farmer,

A supermarket receives much more, namely 44%.

For a long time, Chofy did not produce fairtrade chocolate, so the supermarkets’ profit increased instead of the cacao farmers’ one”.

The new chocolate bars from Chofy are fair trade. As a result, Chofy helps farmers in developing countries to gain a better place in the trade chain, so that they can live off their work and invest in a sustainable future.

To collect data, the survey was spread online via social media (Facebook, Instagram, LinkedIn, Whatsapp) among Dutch citizens. Through these media, data was collected from 160 participants, of which 121 were female (75,6%) and 39 were male (24,4%). The age of the sample was between 15 and 61 (M = 28.21 years, SD = 10.93)

The participants were randomly assigned to four groups (groups of 36, 40, 41 and 43), and exposed to the webpage linda.nl with a brand prominent advertorial including a one- or two-sided message, or a non-brand prominent advertorial including a one- or two-sided message. The survey started with a brief introduction and explanation of the procedure. After exposure to the advertorial, the participants continued to the survey where they answered a series of questions.

3.2 Measurement scales

The items and scales are displayed in table 3.2. As prior research suggests, brand attitude was measured using three items; perceived quality; perceived value; and brand satisfaction (Du, et al., 2019). To measure WoM, two items were used constructed by Oliver and Swan (1989), and Singh (1988); desire to recommend the brand to other people and willingness to say good things about the brand to other people.One item from a scale by Valentini, Romenti, Murtarelli and Pizzetti (2018) was added, regarding online sharing, to make this measurement scale complete, since it revolves around online content. The reason for only adding the item of online sharing and not the whole scale by Valentini, et al. (2018), is that with the kind of advertorials used in this experiment, it is only possible to share it online. Clicking a ‘like-button’ or leaving a comment is not possible.

In this research is controlled for process involvement, product involvement and source attitude. To ensure that respondents were in an enhanced processing condition when exposed to the advertorial, they were asked for their involvement with the advertorial on a three item 7-point involvement scale which was adopted from previous studies (Zaichkowsky, 1994). Besides, there is controlled for product involvement, by including a three item scale by Chandrashekaran (2004). Lastly, in this research is controlled for source attitude by using the Reysen Likeability scale (Reysen, 2005).

In addition, a Cronbach’s alpha consistently above 0.6, indicated good internal consistency (Hair, Hult, Ringle & Sarstedt, 2014). The scales of the items were averaged to measure the constructs.

TABLE 3.2

CONSTRUCTS AND MEASURES

Construct Item FL CA

Brand prominence (Han, et al., 2010)

● How prominently does this advertorial display its trademark? (Brand name, logo, etc.)

Not prominent at all/very prominent

● To what extent would this advertorial be recognizable as a Chofy advertisement?

Very unrecognizable/very recognizable

.848 .871

.844

Message

sidedness Strongly disagree/strongly agree ● Chofy used to harm African cacao farmers Brand attitude

(Du, et al., 2019) ● I associate Chofy with good quality ● I associate Chofy with good value-for-money

● I could identify myself as a satisfied customer of Chofy Strongly disagree/strongly agree

.640 .666 .645

.708

Word of Mouth (Oliver & Swan, 1989; Singh, 1988)

● I would recommend Chofy to friends/family ● I am willing to tell positive things about Chofy to

friends/family

● I am willing to share this article online (Valentini, et al., 2018)** Strongly disagree/strongly agree

.810 .791 .894 Process involvement (Zaichkowsky, 1994)

Reading this advertorial was [...] ● boring/interesting ● irrelevant/relevant ● exciting/unexciting .797 .793 .835 .871 Product involvement (Chandrashekara n, 2004) ● I am interested in chocolate

● Given my personal interests, chocolate is not very relevant to me*

● I am involved when I am purchasing chocolate for personal use**

Strongly disagree/strongly agree

.886

.880 .808

Source attitude

(Reysen, 2005) ● I like Linda.nl ● Linda.nl is a friendly website

● Linda.nl is a physically attractive website Strongly disagree/strongly agree

.864 .894 .894

.923

Notes: All are 7-point scales.

Cronbach Alpha scores presented are the scores after removing the uncorrelated items. * Reversed coded variable.

3.3 Factor analysis and Reliability analysis

Table 3.2 presents the outcomes of the factor analysis and reliability analysis. The scores of the items of brand prominence were good (KMO =.500, Bartlett <.001, communalities >.8), since KMO must be .5 or higher, the Bartlett’s test must be significant (<.05), and the communalities must .4 or higher (Malhotra, 2009). Therefore, these items are averaged into brand prominence.

Next is brand attitude which consists of three items. Again, the KMO (=.666), Bartlett’s test (<.001), and communalities (<.6) are good. The factor loadings correlate, so these three items were averaged into brand attitude.

WoM also consists of three items. The KMO (=.640) and Bartlett’s test (<.001) were good, but the factor loadings were not correlating. In addition, the communalities of two items were above .8, but the third item was only .398. Therefore, one item was removed. This was the item that was added to an original WoM-model, as mentioned in chapter 3. It turned out that this item did not match the original WoM-model. The two remaining items were averaged into WoM.

The scores of the three items of process involvement were good (KMO=.733, Bartlett <.001, communalities >.7). In addition, the factor loadings and Cronbach's Alpha were high, so these three items were averaged into process involvement.

Product involvement consisted of three items of which the scores were good (KMO=.551, Bartlett <.001), but the communalities (=.358) and factor loading of one item was very low in comparing to the other two items. This item was removed. The remaining two items were averaged into product involvement.

3.4 Analyses

Table 3.3 presents the estimated models for the hypotheses. TABLE 3.3

ESTIMATED MODELS FOR THE HYPOTHESES

Model 0: WoM = WoM = β₀ + β₁BP + ε WoM = Word of Mouth BP = Brand Prominence (1 = Low brand prominence, 7 = High brand prominence) BA = Brand Attitude MS = Message Sidedness (1 = One-sidedness, 7 = Two- sidedness) ε = Error Term Model 1 (H1): BA = β₀ + β₁BP + ε

Model 2 (H2): WoM = β₀ + β₁BA + ε Model 3 (H3): (1) WoM = β₀ + β₁BP + ε (2) BA = β₀ + β₁BP + ε (3) WoM = β₀ + β₁BA + ε (4) WoM = β₀ + β₁BP + β₂BA + ε Model 4 (H4): BA = β₀ + β₁BP + β₂MS + β₃BP*MS + ε Model 5 (H5): WoM = β₀ + β₁BA + β₂MS + β₃BA*MS + ε

4. RESULTS

4.1 Control variables and manipulation check

The means of an independent samples T-test showed that all participants are at least medium involved with the advertorial (M=4.54, SD=1.31) and with the product (M=5.31, SD=1.55). Involvement is required for a two-sided message to succeed (Hastak & Park, 1990).

In addition, a linear regression shows that process involvement (B=.308, t=3.85, p<.001), source attitude (B=.250, t=2.90, p=.004) and age (B=.186, t=2.33, p=.021) have a significant relation on WoM. Therefore, these variables were included in the tests to control for. Gender (B=-.026, t=-.331, p=.741) and product involvement (B=.082, t=1.049, p=.296) do not have a significant relation on WoM, therefore these variables will not be included in the tests.

As mentioned in chapter 3, the manipulations were successfully pretested. To ensure that the manipulation was successful in the experiment itself as well, the manipulation was checked again. The results of a One-way ANOVA showed significant difference (F=6.28, p<.001) between high brand prominence (M= 5.20) and low brand prominence (M=2.80). This supports a successful manipulation of brand prominence. Besides, the One-way ANOVA showed that the difference between one-sidedness (M=2.19) and two-sidedness (M=4.90) was significant (F=.213, p<.001). This supports a successful manipulation of message sidedness.

4.2 Hypotheses

Table 4.1 summarizes the results of the tests. To test H1 (model 1 in table 3.3 and 4.1), a regression was performed which showed that there is not a significant relation between brand prominence and brand attitude (B=.045, p=.921). This means that we cannot accept H1.

Another regression was performed to test H2 (model 2 in table 3.3 and 4.1), which resulted in a significant relation between brand attitude and WoM (B=.620, p<.001). This relation is positive, which means that people who have a positive brand attitude, are more likely to spread WoM. Therefore, we can confirm H2.

A regression showed (model 0 in table 3.3 and 4.1) that the relation between brand prominence and WoM is not significant (B=.002, p =.481), which means a partial mediation effect is not possible. The second step would be to show that the causal variable (brand prominence) would be related with the mediator (M = brand attitude). It was shown with a regression (model 1 in table 3.3 and 4.1) that the relation between brand prominence and brand attitude was not significant (B=.045, p=.921). The next step would be to show the relation between the mediator and the outcome variable, so, the relation between brand attitude and WoM (model 2 in table 3.3 and 4.1). In this case, this is the only significant relation (B=.620, p<.001). The last step would be to show that the M mediates the relation between X and Y. This indirect relation is not significant, since it includes zero. The relation should not include zero for it to be significant. In addition, if from table 3.3 and 4.1 models 0, 1 and 2 were significant, partial mediation would exist. If model 0, 1, 2 and 3 were significant, full mediation would exist. Lastly, if models 1 and 2 were significant, mediation would also exist. However, just model 2 is significant. Therefore, these scenarios are not applicable. Thus, we can conclude that we cannot accept H3.

Therefore, we cannot confirm H4.Figure 4.1 shows the expectation of the moderation on the left. The figure on the right are the actual means, where is becomes clear that the combination of high brand prominence and two-sidedness does not create the highest brand attitude, which was proposed. Overall, one-sidedness scores higher than two-sidedness.

FIGURE 4.1

MEANS USING INDEPENDENT SAMPLES T-TEST

TABLE 4.1

4.3 Additional analysis

To check for the results, an additional analysis was performed. The full model was tested using Process model 58 (Hayes, 2013). Figure 4.2 shows the results, from which becomes clear that there is only a relation between brand attitude and WoM. This confirms the previous analysis and can we only accept H2.

FIGURE 4.2

MODERATED MEDIATION MODEL

*p≤.005, **p<.001

Besides, as an additional analysis, a linear regression of message sidedness on brand attitude and on WoM was done. The linear regression of message sidedness on brand attitude shows a significant relation (B=.165, p=.038). Plotting the means show that one-sidedness positively influences brand attitude. Moreover, when one understands the one-sided message, one is more likely to have a positive attitude towards the advertising, and thus the brand.

The result of message sidedness on WoM also turned out significant (B=.200, p=.011). Plotting the means show that one-sidedness positively influences WoM. Therefore, when one understands the one-sided message, one is more likely to spread WoM.

Brand

5. DISCUSSION

5.1 Conclusions

This research provides insights on the effectiveness of brand prominence and message sidedness on brand attitude and WoM. In addition, this research contributes to the academic relevance of native advertising by exploring the effectiveness of new variables and combinations of variables that were not studied before. The goal was to find the right balance of brand prominence and message sidedness to create the highest possible brand attitude and WoM. However, the effects of brand prominence and message sidedness were not significant. The result did not show that high brand prominence combining with two-sidedness creates a higher brand attitude, and higher WoM.Another contrasting, result is that two-sided messages in general did not perform better than one-sided messages.

Proposed was that high brand prominence would lead to a low brand attitude, since high brand prominence leads to activation of the consumer’s persuasion knowledge, which leads to negative consumer evaluations (Reijmersdal, 2009). Which is why subtle advertising leads to better consumer evaluations than prominent advertising (Tutaj & Reijmersdal, 2012). The relation was not tested significant, but the result was the other way around. Therefore, high brand prominence led to slightly higher brand attitude than low brand prominence. Previous research confirms this finding (Harms, Bijmolt & Hoekstra, 2019), where prominent brand advertising was evaluated higher than low brand prominent advertising. This could be explained through the lower level of ‘secretive’ in ads, which would make the ad more successful in the long run (Campbell & Marks, 2015). Besides, high brand prominence previously resulted in generating awareness and building brands (Lobschat, Osinga, & Reinartz, 2017), which is missing with low brand prominence.

Although this relation is significant, a mediation effect was not found, which is not in line with the hypothesis. The relation of brand prominence on brand attitude was not significant, which makes mediation impossible. This contrasts current literature, since exciting theories show the stages a consumer goes through (Fennis & Stroebe, 2015), which corresponds with the proposed mediation. We account for this significance in the limitations section.

Another proposition was that two-sided messages would increase brand attitude, and even turn the negative relation of brand prominence on brand attitude into a positive relation. Message sidedness did not significantly moderate the relation between brand prominence and brand attitude, which means the combination of high brand prominence and two-sidedness is not the optimal combination to create an ad with high brand prominence. However, it did became clear that one-sided messages resulted in significant higher brand attitude than two-sided messages. This contrasts current literature (Allen, Hale, Mongeau, Berkowitz‐Stafford, Stafford, Shanahan, Agee, Dillon, Jackson & Ray, 1990) which propose that two-sided messages are more persuasive than one-sided messages. Besides, the one-sided ad including high brand prominence scored slightly higher on brand attitude in comparison to the low brand prominent ad. Previous research confirms that although clear brand prominence is essential to create marketing effects, the most effective level of brand prominence is not certain (Harms, et al., 2017).

Lastly, proposed was that the significant relation of brand attitude on WoM was moderated by message sidedness. The moderator was not significant, which means an interaction effect does not exist. This means that there is not a difference between one-sided or two-sided message regarding the effect on WoM. This is not in line with current literature, which states that including two-sided messages are more effective regarding engagement and behavioural intentions, such as WoM (Pierro, Giacomantonio, Pica, Giannini, Kruglanski & Higgins, 2013; Konuk, 2019).

This key determinant, when a positive prior attitude exists, is likely to be whether the consumer has prior awareness of the negative information contained in the two-sided message. If the consumer was not aware of the negative information, the message will be counter attitudinal and the consumer may be motivated to reject and/or counter-argue the message (Sawyer, 1973). By creating a fictitious chocolate brand, an attempt was made to prevent prior positive attitudes. Still, product involvement of the sample was very high. If the majority of the sample was unaware of the poor living situation of the African cacao farmers, this perceived novelty leads to higher motivation to process the message. Still, this leads to more unfavorable attitude changes since more counter-arguments are processed (Crowley & Hoyer, 1994).

Surprisingly, an additional analysis showed that message sidedness has a significant relation on brand attitude and on WoM. The additional analysis shows contrary results on what would logically follow from this research. Namely, one-sidedness showed to be more successful, since the relation is positive.

5.2 Managerial implications

For practitioners, these findings imply that creating high brand attitude is important if you want people to spread Word of Mouth for your company. If a practitioner decides to go with article style native advertising to create high brand attitude, the combination of high brand attitude and one-sidedness should be considered, since this combination gave the highest effectiveness. One-sided message had a significant positive influence on both brand attitude and WoM. Although, the difference between high and low brand prominence was not significantly proven, advised would still be to go for high brand prominence, since previous research showed the important advantages of creating awareness and success in the long run, which was clearly missing with low brand prominence.

5.3 Limitations and future research

We acknowledge some limitations for this research. Firstly, regarding the measurement of brand prominence. It was proposed that brand prominence, without the influence of message sidedness, would negatively influence brand attitude. However, through this research design, it was not possible to measure brand prominence independently from message sidedness, since they were shown in the stimuli design simultaneously. One did not get the opportunity to get influenced by just the amount of brand prominence they were exposed to. It is interesting to correct for this in future research, to find out whether a mediation effect exists. The theory from current literature was convincing to believe a mediation effect could exist.

Secondly, the manipulation check for message sidedness was significant, but the means were not ideal. People found it hard to correctly answer the question, especially regarding two-sidedness, causing that a third of the sample answered this question ‘neutrally’, instead of either agree or disagree. Besides, involvement is necessary for message sidedness to succeed and persuade (Hastak & Park, 1990), but the sample did not show high involvement, just medium. Even, the two-sided advertorial showed lower involvement than the one-sided advertorial. These could be reasons for the insignificant results, which is why we propose not to ignore message sidedness right away, but let it open for future researchers to explore and account for these limitations.

In addition, the sample consisted of 75% female participants. Therefore, our findings cannot generalize for the whole population, since men and women have different consumer emotions and have different brand connections (Moore & Wurster, 2007), which could result in different brand attitudes and WoM. Future research should account for this limitation.

LITERATURE

Allen, M., Hale, J., Mongeau, P., Berkowitz‐Stafford, S., Stafford, S., Shanahan, W., Agee, P., Dillon, K., Jackson, R. & Ray, C. (1990). Testing a model of message sidedness: Three replications. Communication Monographs, 57(4), 275-291.

Anderson, E. W. (1998). Customer Satisfaction and Word of Mouth. Journal of Service

Research, 1(1), 5–17.

Appley, M. H. (1971). Adaptation-level Theory. Academic Press.

Baron, R. M. & Kenny, D. A. (1968). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of

Personality and Social Psychology. 51, 1173-1182.

Buttle, F. A. (1998) Word of mouth: understanding and managing referral marketing. Journal

of Strategic Marketing, 6(3), 241-254.

Campbell, C. & Marks, L. J. (2015). Good native advertising isn’t secret. Business Horizons

58(6), 599-606.

Chandrashekaran, R. (2004). The influence of redundant comparison prices and other price presentation formats on consumers’ evaluations and purchase intentions. Journal of Retailing. 80(1), 53-66.

Cleff, E. (2007). Privacy issues in mobile advertising. International Review of Law

Computers and Technology, 21(3), 225-236.

Crowley, A. E., & Hoyer, W. D. (1994). An integrative framework for understanding two-sided persuasion. Journal of Consumer Research, 20, 561−574.

Danescu-Niculescu-Mizil, C., Cheng, J., Kleinberg, J., and Lee, L. (2012). You had me at hello: how phrasing affects memorability. Association for Computational Linguistics, 1, 892-901. DeCarlo, T. E., Laczniak, R. N., Motley, C. M. & Ramaswami, S. (2007). Influence of image and familiarity on consumer response to negative Word-of-Mouth communication about retail entities. Journal of Marketing Theory and Practice, 15(1), 41-51.

Du, R. Y. Joo, M., Wilbur, K. C. (2019). Advertising and brand attitudes: Evidence from 575 brands over five years. Quantitative Marketing and Economics, 17(3), 257.

Eelen, J., Rauwers, R., Wottrich, V. M., Voorveld, H. A. M., & van Noort, G. (2016). Consumer responses to creative media advertising: A literature review. Advertising in new

formats and media: Current research and implications for marketers Bingley.

Fennis, B. M. & Stroebe, W. S. (2015). Psychology of Advertising.

Hair, J. F., Hult, T. M., Ringle, C. M., & Sarstedt, M. (2014). A primer on partial least squares structural equation modelling (PLS-SEM).

Han, Y. J., Nunes, J. C., & Drèze, X. (2010). Signaling Status with Luxury Goods: The Role of Brand Prominence. Journal of Marketing, 74(4), 15–30.

Harms, B., Bijmolt T. H. A. & Hoekstra, J. C. (2017). Digital Native Advertising:

Practitioner perspectives and a research agenda. Journal of Interactive Advertising, 17(2), 80-91.

Harms, B., Bijmolt T. H. A. & Hoekstra, J. C. (2019). You don’t fool me! Consumer perceptions of digital native advertising and banner advertising. Journal of Media Business

Studies, 16(4), 275-294.

Hastak, M. & Park, J. W. (1990). Mediators of message sidedness effects on cognitive structure for involved and uninvolved audiences. Advances in Consumer Research. 329–336. Herold, K., Tarkiainen, A., & Sundqvist, S. (2016). How the source of word-of-mouth influences information processing in the formation of brand attitudes. Journal of Marketing for

Higher Education, 26(1), 64–85.

Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: a regression-based approach. Journal of Education Measurement. 51, 335-337.

JunSang Lim, & Beatty, S. E. (2005). The Impact of Inconsistent Word of Mouth on Brand Attitude. Advances in Consumer Research - Asia-Pacific Conference Proceedings, 6, 262–270. Kao, D. T. (2011). Message sidedness in advertising: The moderating roles of need for

cognition and time pressure in persuasion. Scandinavian Journal of Psychology 52, 329–340. Keller, E. & Fay, B. (2009). The role of advertising in word of mouth. Journal of Advertising

Research. 49(2), 154-158.

Kim, E.-J., Kim, S.-H., & Lee, Y.-K. (2019). The effects of brand hearsay on brand trust and brand attitudes. Journal of Hospitality Marketing & Management, 28(7), 765–784.

Konuk, F. A. (2019). The influence of perceived food quality, price fairness, perceived value and satisfaction on customers’ revisit and word-of-mouth intentions towards organic food restaurants. Journal of Retailing & Consumer Services, 50, 103–110.

Kudeshia, C., & Kumar, A. (2017). Social eWOM: does it affect the brand attitude and purchase intention of brands? Management Research Review, 40(3), 310–330.

Kumar, V., Luo, A. & Rao V. (2015). Linking customer brand value to customer lifetime value. Working Paper.

Malhotra, N. K. (2009). Marketing research: An applied orientation.

McAlister, L., Srinivasan, R., Jindal, N., & Cannella, A. A. (2016). Advertising Effectiveness: The Moderating Effect of Firm Strategy. Journal of Marketing Research, 53(2), 207–224. McGuire, W J. (1968). Personality and susceptibility to social influence. Handbook of

personality theory and research. 1130-1187.

Moore, D. & Wurster, D. (2007). Self-Brand Connections and Brand Resonance: the Role of Gender and Consumer Emotions. Advances in Consumer Research, 34, 64-66.

Nicholls, J.A.F. (1999). Brand recall and brand preference at sponsored golf and tennis tournaments. European Journal of Marketing, 33(3/4), 365.

Oliver, R. L., & Swan, J. E. (1989). Consumer perceptions of interpersonal equity and satisfaction in transaction: A field survey approach. Journal of Marketing, 53(2), 21-35. Olney, T. J., Holbrook, M. B. & Batra, R. (1991). Consumer responses to advertising: The effects of ad content, emotions, and attitude toward the ad on viewing time. Journal of

Consumer Research, 17(4), 440-453.

Pansari, A. & Kumar, V. (2017). Customer engagement: the construct, antecedents, and consequences. Journal of the Academy of Marketing Science, 45(3), 294–311.

Percy, L & Rossiter, J. R. (1992). A model of brand awareness and brand attitude advertising strategies. Journal of Marketing & Psychology, 9(4), 263-274.

Petty, R. E., Cacioppo, J. T., & Schumann, D. (1983). Central and peripheral routes to advertising effectiveness: The moderating role of involvement. Journal of Consumer Research,

10(2), 135-146.

Pierro, A., Giacomantonio, M., Pica, G., Giannini, A. M., Kruglanski, A. W., & Higgins, E. T. (2013). Persuading drivers to refrain from speeding: Effects of message sidedness and regulatory fit. Accident Analysis & Prevention, 50, 917–925.

Pongjit, C., & Beise-Zee, R. (2015). The effects of word-of-mouth incentivization on consumer brand attitude. Journal of Product & Brand Management, 24(7), 720–735.

Rageh, I. A. & Spinelli, G. (2012). Effects of brand love, personality and image on word of mouth: The case of fashion brands among young consumers. Journal of Fashion Marketing

and Management, 16(4), 386-398.

Reijmersdal, E. (2009). Brand placement prominence: Good for memory! Bad for attitudes.

Journal of Advertising 49, 151-153.

Reijmersdal, E., Rozendaal, E. & Buijzen, M. (2012). Effects of Prominence, Involvement, and Persuasion Knowledge on Children's Cognitive and Affective Responses to Advergames.

Reysen, S. (2005). Construction of a new scale: The Reysen likeability scale. Social Behavior

and Personality, 33, 201– 208.

Sawyer, A. G. (1973). The Effects of Repetition of Refutational and Supportive Advertising Appeals. Journal of Marketing Research, 10, 23-33.

Shah, D., Rust, R. T., Parasuraman, A., Staelin, R., & Day, G. S. (2006). The Path to Customer Centricity. Journal of Service Research, 9(2), 113–124.

Sherif, C. W., Sherif, M. & Nebergall, R. E. (1961). Attitude and Attitude Change. Yale

University Press.

Singh, J. (1988). Consumer complaint intentions and behavior: Definitional and taxonomical issues. Journal of Marketing, 52(1), 93-107.

Tripathi, G. (2017). Customer Satisfaction and Word of Mouth Intentions: Testing the Mediating Effect of Customer Loyalty. Journal of Services Research, 17(2), 1–16.

Tutaj, K. & Reijmersdal, van, E. A. (2012). Effects of online advertising format and

persuasion knowledge on audience reactions. Journal of Marketing Communications, 18(1), 5-18.

Valentini, C., Romenti, S., Murtarelli, G. and Pizzetti, M. (2018). Digital visual engagement: influencing purchase intentions on Instagram. Journal of Communication Management, 22(4), 362-381.

Wangenheim, F. & Bayón, T. (2007). The chain from customer satisfaction via word-of-mouth referrals to new customer acquisition. Journal of the Academy of Marketing Science. 35(233). Wojdynski, B. W. (2016). The Deceptiveness of Sponsored News Articles: How Readers Recognize and Perceive Native Advertising. American Behavioral Scientist, 60(12), 1475– 1491.

Wojdynski, B. W. & Evans, N. J. (2016). Going Native: Effects of Disclosure Position and Language on the Recognition and Evaluation of Online Native Advertising. Journal of

Advertising, 45(2), 157-168.

Zaichkowsky, J.L. 1994. The personal involvement inventory: Reduction, revision, and application to advertising. Journal of Advertising 23(4), 5970.

Zhang, X., Ma, L., & Wang, G.-S. (2019). Investigating consumer word-of-mouth behaviour in a Chinese context. Total Quality Management & Business Excellence, 30(5/6), 579–593.

Electronic documents

Enberg, J.; (2019); Emarketer.com; Digital ad spending global; Accessed at 19 September, 2019

Krouwer, S.; (2019); Frankwatching.com; Waarom Native Advertising in 2019 blijft groeien; Accessed at 19 September, 2019

APPENDIX

1. Design of stimuli

Online Advertising is growing

★ Matches the form and function of

the platform upon which it appears

★ Power = Low brand prominence

★ Positive attitude

Disadvantages

of

N ative

Advertising

Low recognition

Simple solution?

Increase Brand Prominence

Message Sidedness

One-Sided message

Favorable claims on all brand attributes

Contains disclaimers or admissions of

Two-Sided message

inferior performance on one or more

M odel

★

★

★ H1: Activation of the consumer’s persuasion knowledge → Negative consumer

evaluations (Reijmersdal, 2009).

★ H2: Customer satisfaction → Positive effect on WoM (Zhang, Ma & Wang. 2019;

Tripathi, 2017; Wangenheim & Bayón, 2007).

★ H3:

Stages of consumer responses: cognitive responses, affective responses,

and behavioral responses (Fennis & Stroebe, 2015).

★ H4:

Two-sided messages in advertising elicit more favorable ad attitudes than

one-sided messages (Kao, 2011).

★ H5:

Favorable prior attitude → Message acceptance (Crowley & Hoyer, 1994) →

Increases effectiveness of ads regarding engagement and behavioural intentions

(Pierro, Giacomantonio, Pica, Giannini, Kruglanski & Higgins, 2013; Konuk, 2019)

★ Between-subjects

experimental design

★ Sample of 160

(121 female

(75,6%) and 39 were (24,4%; M =

28.21 years, SD = 10.93)

★ 4 conditions

○ High Brand Prominence +

One-sided message

○ Low Brand Prominence +

One-sided Message

○

High Brand Prominence +

Two-sided Message

Mediation effect: Process model

4 (Hayes, 2013) → Barron and

Kenny method (1968)

Linear Regressions

Full Model: Process model

58 (Hayes, 2013)

★ H1 = Not Significant

★ H2 = Significant

★ H3 = Not Significant

★ H4 = Significant difference,

but relation is not significant

★ H5 = Not Significant

14

★ The right amount of Brand Prominence is not clear ★ Brand Attitude + WoM ★ No mediation (limitations) ★ One-sided Messages perform

Managerial Implications

Limitations & Future Research

Measurement of

Brand Prominence

One-sided Message

+

High Brand Prominence

Age

+

Source

Thank you!

Questions?