Computers in Human Behavior 96 (2019) 46–55

Contents lists available at ScienceDirect

Computers in Human Behavior

journal homepage: www.elsevier.com/locate/comphumbeh

Content is king – But who is the king of kings? The effect of content T
marketing, sponsored content & user-generated content on brand responses
Johannes Müllerb, Fabian Christandl∗,a
a
Hochschule Fresenius - University of Applied Sciences, Im MediaPark 4c, 50670, Cologne, Germany
b
GMK Markenberatung, Agrippinawerft 30, 50678 Cologne, Germany

A R T I C LE I N FO A B S T R A C T

Keywords: Despite a growing focus on content marketing as a modern marketing tool, research on it is sparse. Missing
Content marketing completely is a comparison of content marketing with other forms of content, namely sponsored and user-
Sponsored content generated content, although these content types are of high relevance for the marketing strategy of a company.
User-generated content To fill this research gap, the present study examines how different content types are perceived and how they
Persuasion knowledge
influence brand responses through persuasion knowledge. A serial mediation model is developed, which posits
Brand attitude
Serial mediation
that different content types lead to a varying conceptual persuasion knowledge, which then influences the ac-
tivation of attitudinal persuasion knowledge and in turn results in different brand attitudes. The corresponding
model is tested in an experimental study, using different content types in the context of the video game industry.
The findings indicate that, through the proposed serial mediation, sponsored content leads to a more negative
brand attitude than user-generated content and content marketing. These results suggest that, although coming
directly from a company, content marketing seemingly is perceived in a similar way as user-generated content.
The implications for marketing managers concerning content marketing strategies are discussed.

1. Introduction understanding of how content is perceived, it also helps marketing

professionals to decide which content type best suits their needs. The
“Content is King” is the title of an essay which Microsoft founder Bill development of YouTube from a purely user-generated content plat-
Gates originally presented in 1996. In this essay, Gates talks about the form to a platform for professionally produced content shows that for
future of the internet, stating in his opening sentence that “Content is many companies and marketing professionals, content is already an
where I expect much of the real money will be made on the Internet important marketing tool (Kee & Yazdanifard, 2015; Kim, 2012).
(…)” (Gates, 1996, para. 1). By now this essay is more than 20 years This de facto omnipresence of content makes the question whether
old, but it seems that Gates hit the nail right on the head. When doing a content is king obsolete and instead lays the focus on the question of
quick Google search for “Content is King”, one can see that the phrase is what content is the best content, or in other words, who is the king of
still as popular as ever. While Gates (1996) originally described that the kings? Looking at this question, three types of content that are promi-
term “content” can mean many things on the internet, the title of his nent in the marketing literature are compared in an experimental de-
essay is especially popular in two connected fields, namely search en- sign: content marketing, sponsored content and user-generated content.
gine optimization and content marketing (textbroker, n.d.). Conse- A serial mediation model is developed and tested to examine whether
quently, there are many articles and blog entries discussing why con- these content types lead to differentiating brand responses and whether
tent supposedly is (see for instance Chef n.d.; Jefferson n.d.; West, such differences can be explained through persuasion knowledge. We
2015) or is not king (see for instance Tobak, 2016; White, 2016). hypothesized that these three content types are perceived differently by
Taking a different point of view, the question if content is king might consumers because they activate the conceptual and attitudinal per-
not be the right one to ask. Rather, it should be of interest which un- suasion knowledge of consumers to a different extent. By showing that
derlying mechanisms decide whether content is an effective commu- this is the case, our research goes beyond previous evidence by com-
nication tool or not. In addition it should be determined which type of paring three types of content that so far have only been investigated on
content is accepted more by recipients because of these mechanisms. their own. This provides an interesting contribution to the literature on
Not only does this approach further extend the theoretical content marketing and leads to explicit managerial implications for

∗
Corresponding author.
E-mail addresses: j.mueller@gmk-markenberatung.de (J. Müller), fabian.christandl@hs-fresenius.de (F. Christandl).

https://doi.org/10.1016/j.chb.2019.02.006
Received 15 November 2018; Received in revised form 3 January 2019; Accepted 7 February 2019
Available online 11 February 2019
0747-5632/ © 2019 Elsevier Ltd. All rights reserved.
J. Müller and F. Christandl Computers in Human Behavior 96 (2019) 46–55

marketing professionals. By investigating how different content types higher activation of persuasion knowledge while showing mixed results
possibly influence peoples' perception through conceptual and attitu- concerning the effect on brand attitude. A follow-up study did show
dinal persuasion knowledge in serial mediation, this study furthermore that a sponsorship disclosure primes viewers and as a consequence
contributes to the understanding of how people process information of activates resistance against persuasion, which then leads to a more
persuasive nature, therefore bearing importance for the literature on negative brand attitude compared to viewers that did not see a spon-
persuasion and consumer behavior. sorship disclosure (Boerman et al., 2014). Van Reijmersdal and col-
leagues (2016) found similar results when looking at the effect of
2. Previous work on different content types sponsorship disclosure in the context of blogging, with persuasion
knowledge mediating the effect of disclosure on brand attitude (van
The term user-generated content is used broadly and can be applied to Reijmersdal et al., 2016). While the studies mentioned so far did not
very different media types, which can be a problem when trying to explicitly state that sponsored content leads to a more negative brand
compare different user-generated content studies with each other. One attitude but only that the disclosure has a negative effect on it, it is
form of user-generated content that has been researched are reviews, important to remember that there are legal requirements for the dis-
with studies showing that positive reviews influence the number of closure of a sponsorship. Therefore, the disclosure is inextricably linked
bookings on a travel website (Ye, Law, Gu, & Chen, 2011) and that to sponsored content, meaning that the studies de facto propose that
people are more likely to use reviews if they perceive the credibility of sponsored content leads to a more negative attitude than non-sponsored
the source to be high (Ayeh, Au, & Law, 2013). A study by Zhu and content. Although there are many more studies on the effect of spon-
Zhang (2010) investigated the effect of user-generated content in the sorship disclosure (see for instance Hwang & Jeong, 2016; Janssen,
video game industry, finding that reviews were especially influential for Fransen, Wulff, & van Reijmersdal, 2016; Wojdynski & Evans, 2016),
lesser known games and that even one negative review can possibly the line of research presented thus far is of special importance for the
damage the success of such a game (Zhu & Zhang, 2010). Not all user- theoretical framework of this study and therefore suffices to give a
generated content, however, is primarily evaluative in its nature, which picture of the current research. It suggests that sponsored content is
could finally result in effects on the recipients differing from the effects able to influence the attitude towards the sponsoring brand, although
for reviews reported above. Tang, Fang, and Wang (2014) investigated its influence is lower compared to content in which the brand is in-
effects of mixed and indifferent user-generated content, both types cluded without sponsorship. For the current study this implies that
being neutral in essence, but mixed content including positive and sponsored content should have a more negative influence on brand
negative remarks while indifferent content included neither. The au- attitude than other forms of content that involve the brand without
thors found that mixed content led to higher sales numbers because it being sponsored, like user-generated content. To complete the overview
motivates the consumer to process the available information and en- of the three content types relevant for this study, a look at content
gage with the content while indifferent content led to lower sales marketing is required.
numbers because it is of no interest for the consumer (Tang et al., Companies using content marketing as a marketing tool are “(…)
2014). Integrating the notion of neutral user-generated content adds an creating, distributing and sharing relevant, compelling and timely
interesting aspect to the literature on user-generated content. Overall, content to engage customers at the appropriate point in their buying
literature supports the idea that there seems to be some sort of link consideration processes, such that it encourages them to convert to a
between user-generated content and purchase behavior, with positive business building outcome” (Holliman & Rowley, 2014, p. 285). It
user-generated content leading to higher sales figures. Still, there is a differentiates from other content types in that it is created as well as
gap in current literature concerning other forms of user-generated shared by the company itself, e.g. through a social media channel
content, for example blogs or YouTube videos that are made for en- owned by the company and that from the point of view of the con-
tertaining the user. Also missing from the literature on user-generated sumer, no third parties are visible. One fundamental insight into con-
content is a comparison between user-generated content and other tent marketing practices from Holliman and Rowley (2014) shows that
forms of content. the content produced for content marketing purposes needs to be free of
The literature on sponsored content is relatively comprehensive. selling messages and instead needs to focus on the particular interests of
There are different understandings and definitions of sponsored content consumers. This is a key component of content marketing as it funda-
that often correspond to the topic of the respective study. This study mentally distinguishes it from classical advertising messages and ex-
follows the definition of Boerman, van Reijmersdal, and Neijens (2014, plains why it works as an inbound marketing tool that people consume
p. 215) describing sponsored content as “(…) the intentional in- voluntarily (Holliman & Rowley, 2014). Similar results were found in
corporation of brands, products, or persuasive messages into tradi- another study that interviewed content marketing practitioners
tionally noncommercial, editorial content”. There are several studies (Järvinen & Taiminen, 2016). Focusing on firm generated content in
examining the effect of sponsored content. van Reijmersdal, Neijens, social media, Kumar, Bezawada, Rishika, Janakiraman, and Kannan
and Smit (2007) found that the brand image of people that watched (2016) found a positive relationship between social media participation
episodes of editorial content with said brand integrated into it became of customers and their respective spending and cross-buying behavior.
more similar to the image of the program they were watching. Adding There are several other studies on social media content that could
to these results, Dens, De Pelsmacker, Wouters, and Purnawirawan possibly be applied to content marketing (see for instance Tafesse,
(2012) showed that the prominence of a brand placement in a movie as 2015; Kilgour, Sasser, & Larke, 2015), but the applicability of these
well as its connection to the plot influences how well viewers can re- results would be highly speculative because of their difference to our
cognize the brand and how positive their attitude towards the brand is. research. From the literature on content marketing, we can infer for our
Looking at the effects of sponsored content on the source of the content, current study that, compared to no communication, content marketing
a study found that participants' attitudes towards influencers that pro- can have a positive influence on the attitude and behavior of consumers
duce sponsored content can also change depending on the level of towards the brand that produces the content. At the same time, it has to
disclosure of the sponsoring, with a tacitly disclosure leading to a lower be mentioned that more specific literature on content marketing fo-
perceived credibility of the influencer (Carr & Hayes, 2014). In addi- cusing on persuasive effects is somewhat lacking. While some studies
tion, a series of studies examined the influence of sponsorship dis- analyze content marketing on Facebook in detail (see for instance
closure and the effect persuasion knowledge has in this context. Tafesse, 2015), others try to grasp what practitioners see as good con-
Boerman, van Reijmersdal, and Neijens (2012) examined how the dis- tent (see for instance Holliman & Rowley, 2014) and then again another
closure that the content is sponsored influences brands responses in a study focusses on the influence of content marketing on sales figures
TV show and found that the sponsored content condition lead to a (see for instance Kumar et al., 2016). This leaves the literature on

47
J. Müller and F. Christandl Computers in Human Behavior 96 (2019) 46–55

content marketing in a very confusing and unclear state, in which it is could be possible that it leads to a lower level because it feels close to
nigh on impossible to compare one study to another. Additionally, there “normal” content and is completely honest about its source, similar to
seems to be an almost complete lack of research about the psychological the results of Carr and Hayes (2014) showing that sponsored content is
effect content marketing has on the consumer. This makes it extremely evaluated more favorably when clearly disclosed as such. In a similar
difficult to formulate any hypotheses on the effect of content marketing vein, Lee and Youn (2009) found that people were more likely to re-
concerning brand responses without guessing the direction of the effect. commend a product based on information found on a company home-
page than in a personal blog. On the other hand, it also could be pos-
sible that it leads to a higher level of conceptual persuasion knowledge
2.1. The role of persuasion knowledge
as the company as the source of the content is naturally biased. There is
evidence in literature that marketing content from a service provider is
A construct often used to describe how a possible difference in
perceived more negatively than user-generated or editorial content
brand responses to different content types could be explained is per-
(Dickinger, 2011) and that the influence of content on a seller website is
suasion knowledge. The Persuasion Knowledge Model was originally
weaker than on an independent site (Ha, Bae, & Son, 2015). Because of
described by Friestad and Wright (1994), aiming to explain how people
the ambivalence of a possible effect, content marketing will not be in-
react when they are the target of a persuasion attempt. Other re-
cluded in the hypotheses formulated here. Instead, its effect will be
searchers have proposed that persuasion knowledge includes two
analyzed on an explorative basis.
components, criticizing that previous research only focusses on con-
Once persuasion knowledge is conceptually activated, it should lead
ceptual knowledge of persuasion and advertising (Rozendaal, Lapierre,
to an attitudinal reaction. Naturally, one would assume that something
van Reijmersdal, & Buijzen, 2011). Following this argumentation,
that is perceived as advertising will be less trustworthy, which is in
Boerman et al. (2012) divide persuasion knowledge into two parts,
accordance with results showing that people in general see advertising
conceptual and attitudinal. Conceptual persuasion knowledge includes
as not trustworthy (Shavitt, Lowrey, & Haefner, 1998). This attitudinal
recognizing a persuasion attempt and understanding that it comes from
reaction was also demonstrated in a study by Boerman et al. (2012),
a certain source and uses specific tactics to target an audience. Attitu-
who found a negative effect of recognition of advertising on trust-
dinal persuasion knowledge describes how people react to the persua-
worthiness in their study. Therefore high conceptual persuasion
sion attempt, ranging from skepticism or disliking of the message to
knowledge should negatively influence trustworthiness, indicating a
other critical attitudes like assessing the message as less trustworthy
higher activation of attitudinal persuasion knowledge:
and honest (Boerman et al., 2012). Following this approach, this study
differentiates between conceptual and attitudinal persuasion knowl- Hypothesis 2. Sponsored content leads to a higher level of conceptual
edge. persuasion knowledge, which then leads to a higher activation of
attitudinal persuasion knowledge compared to user-generated content.
3. Hypotheses Attitudinal persuasion knowledge should in turn have an influence
on brand attitude. The effect of one attitudinal reaction, the assessment
Following the literature on user-generated content, sponsored con- of the sources trustworthiness, on attitudes has been supported by
tent and content marketing, a difference regarding brand responses is to several studies in the past (Ohanian, 1990). Ayeh et al. (2013) found
be expected. If different content types lead to a different level of per- trustworthiness to influence attitude towards user-generated content.
suasion knowledge, this should in turn influence participants brand Therefore attitudinal persuasion knowledge should influence brand
attitude as a reaction to the persuasion attempt. As described in the attitude in the following way:
chapter on persuasion knowledge, the first step in its activation is to
Hypothesis 3. Sponsored content, through conceptual persuasion
recognize a message as persuasive in nature. For many content forms on
knowledge leads to a higher activation of attitudinal persuasion
the web, the only perceivable difference between sponsored content
knowledge, which then leads to a more negative brand attitude
and user-generated content is the disclosure of sponsorship. Following
compared to user-generated content.
several research results (see for instance Boerman et al., 2012; 2014),
content in which a sponsorship is disclosed should lead to a higher These hypotheses do not assume a direct effect of the content type
perception of content as advertising compared to no disclosure. If the on attitudinal persuasion knowledge or brand attitude but rather an
sponsored content is identical to user-generated content, except for the indirect effect through conceptual persuasion knowledge and conse-
disclosure, the level of conceptual persuasion knowledge should be quently attitudinal persuasion knowledge. This assumption is backed
higher: up by several studies (see for instance Boerman et al., 2012; 2014; van
Reijmersdal et al., 2016). Conceptual and attitudinal persuasion
Hypothesis 1. Sponsored content leads to a higher level of conceptual
knowledge are mediating the effect of content type on brand attitude.
persuasion knowledge compared to user-generated content.
The resulting serial mediation model is depicted in Fig. 1.
As the literature on content marketing is very unclear on possible
effects on conceptual persuasion knowledge, it is impossible to assume
the direction of the effect content marketing could have on the level of
conceptual persuasion knowledge without guessing. On the one hand it

Conceptual Attitudinal
Persuasion Persuasion
Knowledge Knowledge

Content Type Brand Attitude

Fig. 1. Complete variable model with conceptual and attitudinal persuasion knowledge mediating the effect of content type on brand attitude.

48
J. Müller and F. Christandl Computers in Human Behavior 96 (2019) 46–55

4. Method YouTuber who got paid for the video by [game developer]. The video
can be found on the YouTube channel of [YouTuber], as seen in the
4.1. Participants picture below”. This time, a screenshot showing the channel of the
YouTuber was shown below this statement. This manipulation estab-
As the study dealt with video games, people with an interest in lished the YouTuber as an independent YouTuber with his own channel
video games were recruited as participants for this study. Recruitment who got sponsored by the developer for the video. Participants in the
platforms were popular online gaming communities as well as gaming sponsored content condition saw a slightly different version of the video
related social media channels. All of these platforms were exclusively shown in the content marketing condition. Only one detail was different
English speaking, so that it was ensured that participants were able to for the sponsored content condition: there were 3 s at the start of the
understand English. Participants were told beforehand that the study video where a sponsorship disclosure statement was displayed. The
was about gaming videos on YouTube and was specifically searching for statement was added in the sponsored content condition to keep the
gamers as participants as well as information about the time it would setting as realistic as possible. Content that is sponsored is required to
take to participate and that they should have audio enabled as they include some form of disclosure by law and YouTube videos are no
would see a video as part of the study. Additionally, participants were exception (Boerman et al., 2014). Consequently, many sponsored Let's
told that their answers would be treated confidentially and only used Play videos on YouTube include either a spoken or a written disclosure
for scientific purposes. at the start of the video. To keep the sponsored content condition as
Altogether 184 people completed the survey. From these partici- realistic as possible, a disclosure statement was included in the video.
pants, 28 were excluded because they had prior knowledge of the sti- The wording used the term “sponsored” to keep the manipulation as
mulus shown or skipped whole answer scales and therefore could not be strong as possible, following the results of Wojdynski and Evans (2016),
analyzed further. After these exclusions, 156 participants were left for who found that this wording leads to a higher recognition of adver-
further analysis. tising.
The average age among participants was 23.84 years (SD = 6.11). The user-generated content condition was very similar to the spon-
Of the participating people, 90% were male, 5% female, 1% indicated sored content condition, showing the same screenshot below a slightly
trans (standing for transgender, transsexual, genderless etc.) and 3% altered statement: “The following video is about [video game], devel-
did not answer the question. The most common nationality among oped by [game developer]. It was recorded by [YouTuber], an in-
participants was American (n = 45) with British (n = 14) and German dependent YouTuber, as content for his channel”. The same screenshot
(n = 13) in second and third place. Overall, the indicated nationalities as in the sponsored content condition, showing the YouTuber's channel
were diverse, ranging from countries as Vietnam to Norway and was shown again. This established the YouTuber as a self-motivated
Mexico, as was intended by the recruitment method. user of the game without any link to the game developer. The video in
the user-generated content condition was similar to the video shown in
4.2. Procedure the content marketing condition, meaning that there was no sponsorship
disclosure before the video.
The stimulus chosen was an excerpt from a so called “Let's Play” After watching the video, all participants were asked several ques-
video from YouTube, showing a YouTuber playing a video game. It tions measuring their attitudinal and conceptual persuasion knowledge
included favorable arguments for the product (the video game being as well as their brand attitude and asking for their social demographics.
played, e.g. the AI of the game being smart and features included in the The measures will be described in more detail in an own paragraph.
game being cool), started without telling the viewer too much about its Participants were also asked for their age, gender and their country of
origin, so that it was plausible for every content condition, and was origin. To make sure that participants' answers were not influenced by
easily understandable for somebody who had not seen the game being previous knowledge about the game or the YouTuber, participants were
played before. The channel of the chosen YouTuber was not too big but asked whether they knew the game, had seen the video or knew the
still big enough to appear professional, so that the video was not too YouTuber before participating. If they answered one of these questions
well known but still was a good representation of this form of content in with yes, they were excluded from further analysis as mentioned before
general. The YouTuber gave his approval for the video to be used in the this chapter.
context of this study. The game played in the video did also meet cer-
tain criteria, namely that it was not well known beforehand but still 4.3. Measures
offered some interesting elements so that the viewers would possibly be
interested. Overall, these criteria were chosen to assure that people Persuasion knowledge was captured by using two measures corre-
were new to the product shown and therefore had no stable attitude sponding to its conceptual and attitudinal dimensions. Conceptual
towards it prior to this study. persuasion knowledge was measured by asking participants for their
An online survey was used to conduct the study. Participants were recognition of advertising. Therefore, we adapted the procedure from
told that the video they will see is about gaming videos on YouTube. In Boerman et al. (2012) by asking participants to indicate on a scale from
the content marketing condition participants saw the following statement 1 (strongly disagree) to 7 (strongly agree) if they felt like the video they
including the actual names of the game, the developer and the saw was advertising. The original wording was slightly changed so that
YouTuber: “The following video is about [video game], developed by it fits the stimulus used in this study. According to Rossiter (2011), this
[game developer]. It was recorded by [YouTuber], who is an employee single item provides a sufficient measure for recognition of advertising.
of [game developer]. The video can be found on the YouTube channel In line with the procedure pursued by Boerman et al. (2012), we
of [game developer], as seen in the picture below”. Below this state- measured participants' attitudinal persuasion knowledge by adapting a
ment, a screenshot of the YouTube channel of the game developer was scale for measuring trustworthiness. The scale originally constructed
shown. In this condition, the developer of the game was clearly pre- and validated by Ohanian (1990) as a subscale for celebrity endorsers'
sented as the source of the video. Following this page, participants in credibility asked participants to rate how they feel about the video they
the content marketing condition saw the excerpt from the YouTube video saw using a total of five items, including pairs such as dependable –
without any changes to the excerpt. undependable or honest – dishonest. Participants should indicate on a
In the sponsored content condition, participants saw the following scale of 1–7 whether they agreed with one pole (e.g. honest) or the
statement, again including the actual names of the game, the developer other (e.g. dishonest). The results were recoded so that a high score
and the YouTuber: “The following video is about [video game], de- indicated a high trust and vice versa. The scale was slightly adapted in
veloped by [game developer]. It was recorded by [YouTuber], a such a way that the wording of the initial statement fits the stimulus of

49
J. Müller and F. Christandl Computers in Human Behavior 96 (2019) 46–55

Table 1
Descriptives and correlation coefficients for the dependent variables (N = 156).
M SD 1 2 3

∗∗
1. Conceptual Persuasion Knowledge [Recognition of Advertising] 3.94 1.74 – -.32 -.20∗
2. Attitudinal Persuasion Knowledge [Trustworthiness] 4.85 1.15 – .35∗∗
3. Brand Attitude 4.87 1.26 –

Note: All variables were rated on a 7-point Likert scale (1 = lowest value, 7 = highest value).
∗
p < .05 ∗∗p < .01.

this study. Its internal consistency was excellent with α = 0.92. The were only marginally significant (p < .10).
mean of the five items was used to create a single score for attitudinal To analyze the serial mediation model statistically, the procedure
persuasion knowledge, with people with a high score trusting the video, described by Hayes (2013) and Hayes and Preacher (2014) was applied,
indicating a low activation of attitudinal persuasion knowledge. including the use of PROCESS for SPSS. Fig. 2 shows the serial med-
Brand attitude was assessed by adapting a validated scale from iation model with labels for every single effect. As sponsored content as
Spears and Singh (2004), asking participants to describe their overall well as content marketing are compared to user-generated content,
feelings towards the game showcased in the video. They could then rate their coefficients are differentiated in Fig. 2 (a1, d1 and e1 for sponsored
their feelings in a semantic differential on a scale from 1 to 7. Examples content, a2, d2 and e2 for content marketing). The indirect effect is
for the pair of poles are unappealing – appealing or unpleasant – described through the path a1bc or a2bc, which shows how content type
pleasant. Participants who got a high score on the scale had a positive influences brand attitude indirectly through conceptual and attitudinal
attitude towards the game. The only change from the original scale persuasion knowledge.
included changing the wording of the initial statement so that it fit to As the independent variable, content type, is a multicategorical
the stimulus of this study. The internal consistency of the scale was variable, two dummy variables were created for the content marketing
excellent with α = 0.92. The mean value of this scale was calculated and sponsored content conditions and in turn used in two serial multiple
and used as a single score for the brand attitude of participants. mediator analyses using the user-generated content group as reference,
as recommended by Hayes and Preacher (2014).
Overall, we conduct our analysis in three steps. First, we compare
5. Results
the sponsored content condition to the user-generated content condition
as reference group, focusing on the hypotheses formulated beforehand.
The descriptive statistics and correlations are summarized in
Second, we analyze the indirect effect of content marketing on brand
Table 1. It is important to note, that a high level of attitudinal per-
attitude compared to user-generated content on an exploratory basis. The
suasion knowledge is indicated by a low trustworthiness. As con-
results of these analyses are summarized in Table 3. In the third step,
ceptually assumed, the results show significant correlations between
we change the reference group from user-generated content to spon-
conceptual persuasion knowledge and attitudinal persuasion knowl-
sored content, so that we can directly compare content marketing to
edge as well as attitudinal persuasion knowledge and brand attitude.
sponsored content. The results of this analysis are summarized in
While no hypothesis concerning content marketing was formulated,
Table 4.
the content marketing condition will still be included into the tests on an
To test the hypotheses, only a comparison between sponsored con-
explorative basis. The means for the different experimental conditions
tent and user-generated content is necessary. Therefore, the mediation
can be found in Table 2. When looking at Table 2 as well as the further
analysis using the dummy variable for sponsored content as in-
results it is important to remember that attitudinal persuasion knowl-
dependent variable is reported first. In a serial mediation analysis,
edge was measured through trustworthiness, therefore a lower score
several regression analyses are calculated. While the hypotheses state
indicates a lower trustworthiness, meaning a higher level of attitudinal
that conceptual and attitudinal persuasion knowledge are mediators, it
persuasion knowledge. The content type had a significant effect on
is also possible that only one or two of these variables are a mediator.
conceptual persuasion knowledge as well as brand attitude, with no
This possibility is included in Fig. 2 with the effects e1, e2 and f showing
significant effect on attitudinal persuasion knowledge. Using one-way
that content type and the mediators could directly influence brand at-
between subjects ANOVAs, a significant effect of content type on con-
titude or each other. Because of this and the fact that there are two
ceptual persuasion knowledge (F (2, 153) = 3.12, p < .05, ηp2 = 0.04)
types of content (a1 and a2, d1 and d2, e1 and e2), there are 9 different
and brand attitude (F (2, 153) = 3.12, p < . 05, ηp2 = 0.04) was found
coefficients that result from this serial mediation analysis. These coef-
while there was no significant effect on attitudinal persuasion knowl-
ficients can be found in Table 3. It summarizes the coefficients, stan-
edge (F (2, 153) = 0.22, p = .80, ηp2 = 0.003). A post hoc mean com-
dard errors and the significance for each variable that could be a pre-
parison between the conditions using the LSD test did find a significant
dictor in the model as well as R2 and F-tests for each of the three
difference between sponsored (M = 4.41, SD = 1.91) and user gener-
regressions. As content type is a multicategorical variable, each content
ated content (M = 3.62, SD = 1.64) regarding conceptual persuasion
type is separately compared to user-generated content. For this reason
knowledge (p < .05) and between the content marketing (M = 5.16,
each content type is listed in Table 3.
SD = 1.13) and sponsored content (M = 4.56, SD = 1.28) condition
Hypothesis 1 stated that sponsored content should lead to a higher
concerning brand attitude (p < .05). The difference between spon-
level of conceptual persuasion knowledge than user-generated content.
sored content (M = 4.41, SD = 1.91) and content marketing
A comparison between both conditions confirms that people in the
(M = 3.77, SD = 1.58) regarding conceptual persuasion knowledge

Table 2
Means and standard deviations in the experimental conditions.
Groups Conceptual Persuasion Knowledge [Recognition of Advertising] Attitudinal Persuasion Knowledge [Trustworthiness] Brand Attitude

Content Marketing (n = 52) 3.77 (1.58) 4.93 (1.17) 5.16 (1.13)

Sponsored Content (n = 54) 4.41 (1.91) 4.80 (1.12) 4.56 (1.28)
User-Generated Content (n = 50) 3.62 (1.64) 4.82 (1.18) 4.91 (1.30)

50
J. Müller and F. Christandl Computers in Human Behavior 96 (2019) 46–55

Conceptual b Attitudinal
Persuasion Persuasion
Knowledge Knowledge

a1 e1
c
a2 e2 f

Content Type
(1 = Sponsored
Content; Brand Attitude
2 = Content
Marketing) d1
d2

Fig. 2. Hypothesized serial mediation model with labeled effects differentiating the sponsored content (subscript 1) and content marketing (subscript 2) conditions.

sponsored content condition had a higher activation of conceptual Hypothesis 2. Sponsored content indirectly leads to a higher activation
persuasion knowledge than participants in the user-generated content of attitudinal persuasion knowledge because of a higher level of con-
condition. The regression analysis shows that sponsored content sig- ceptual persuasion knowledge when compared to user-generated con-
nificantly differs in its influence on conceptual persuasion knowledge, tent.
as indicated by recognition of advertising, compared to user-generated Hypothesis 3 states that sponsored content leads to a more negative
content (a1 = 0.79, p < .05). This result supports Hypothesis 1. brand attitude through conceptual persuasion knowledge and attitu-
Hypothesis 2 states that content type influences attitudinal persua- dinal persuasion knowledge. This hypothesis is corroborated by a serial
sion knowledge through conceptual persuasion knowledge, with spon- mediation analysis. Its results show that conceptual and attitudinal
sored content leading to a higher level of conceptual persuasion persuasion knowledge act as serial mediators between the content
knowledge and in turn to a stronger activation of attitudinal persuasion condition and the resulting brand attitude, with participants in the
knowledge. Therefore, a mediation analysis with attitudinal persuasion sponsored content condition having a more negative brand attitude than
knowledge as dependent variable was conducted, which corroborates their counterparts in the user-generated content condition through a
Hypothesis 2. The coefficients describing the influence of content type higher level of conceptual and a resulting higher activation of attitu-
on conceptual persuasion knowledge (a, e.g. the influence of content dinal persuasion knowledge, indicated by a lower trustworthiness
marketing on conceptual persuasion knowledge) and the influence of score. A bias-corrected bootstrap confidence interval with 10,000 re-
conceptual persuasion knowledge on attitudinal persuasion knowledge samples displaying the indirect effect (a1bc = −0.06) did not include
(b) are equal to the ones in the serial mediation analysis that can be zero (−0.17 to −0.01). Sponsored content leads to higher conceptual
found in Table 3 under coefficients. The results show that the effect of persuasion knowledge (a1 = 0.79) which in turn leads to a higher level
the content condition on the attitudinal persuasion knowledge was of attitudinal persuasion knowledge, indicated by a lower trustworthi-
mediated by conceptual persuasion knowledge. People in the sponsored ness (b = −0.22) and then to a more negative brand attitude
content condition had a higher level of conceptual persuasion knowl- (c = 0.36). The direct effect of sponsored content on brand attitude was
edge (a1 = 0.79, p < .05) and in turn reacted with a higher activation not significant (d1 = −0.31, p = .18). Furthermore, the bias-corrected
of attitudinal persuasion knowledge, as indicated by a lower trust- bootstrap confidence intervals for the indirect effect including only
worthiness assessment (b = −0.22, p < .01). A bias-corrected boot- conceptual persuasion knowledge (a1e2 = −0.04) or only attitudinal
strap confidence interval with 10,000 resamples for the indirect effect persuasion knowledge (e1c = 0.05) both did include zero (−0.19 to
(ab = −0.17) did not include zero (−0.40 to −0.03). Additionally, 0.04 and −0.08 to 0.27) and therefore do not support a mediation
there was no significant direct effect of sponsored content on attitudinal model with only a single mediator.
persuasion knowledge (e1 = 0.15, p = .49). These findings support The overall serial mediation model with the corresponding

Table 3
Model Coefficients for the Serial Mediation Model using the User-Generated Content Condition as Reference Group.
Antecedent Conceptual Persuasion Knowledge [Recognition of Attitudinal Persuasion Knowledge Brand Attitude
Advertising] [Trustworthiness]

Coeff. SE p Coeff. SE p Coeff. SE p

Sponsored Content .79 .34 .02 .15 .22 .49 -.31 .23 .18
Content Marketing .15 .34 .66 .15 .22 .49 .21 .23 .36
Conceptual Persuasion – – – -.22 .05 < .01 -.05 .06 .39
Knowledge
Attitudinal Persuasion – – – – – – .36 .09 < .01
Knowledge
Constant 3.62 .24 < .01 5.61 .24 < .01 3.38 .55 < .01
R2 = 0.04 R2 = 0.11 R2 = 0.16
F (2, 153) = 3.12, p < .05 F (3, 152) = 6.15, p < .01 F (4, 151) = 7.44, p < .01

51
J. Müller and F. Christandl Computers in Human Behavior 96 (2019) 46–55

Table 4
Model Coefficients for the Serial Mediation Model using the Sponsored Content Condition as Reference Group.
Antecedent Conceptual Persuasion Knowledge [Recognition of Attitudinal Persuasion Knowledge Brand Attitude
Advertising] [Trustworthiness]

Coeff. SE p Coeff. SE p Coeff. SE p

Content Marketing -.64 .33 .06 -.001 .22 .99 .52 .23 < .05
User-Generated Content -.79 .34 < .05 -.15 .22 .49 .31 .23 .18
Conceptual Persuasion – – – -.22 .05 < .01 -.05 .06 .39
Knowledge
Attitudinal Persuasion – – – – – – .36 .09 < .01
Knowledge
Constant 4.41 .23 < .01 5.76 .27 < .01 3.07 .57 < .01
R2 = 0.04 R2 = 0.11 R2 = 0.16
F (2, 153) = 3.12, p < .05 F (3, 152) = 6.15, p < .01 F (4, 151) = 7.44, p < .01

-0.22
(0.05)**
Conceptual Attitudinal
Persuasion Persuasion
Knowledge Knowledge
(Recognition (Trustworthin
of ess)
Advertising)

a1 = 0.79 0.36
(0.34)* e1 = 0.15 -0.05 (0.09)**
a2 = 0.15 (0.22) (0.06)
(0.34) e2 = 0.15
(0.22)
Content Type Brand
(1 = Sponsored Attitude
Content;
2 = Content d1 = -0.31
Marketing) (0.23)
d2= 0.21
(0.23)
Fig. 3. Hypothesized serial mediation model comparing the sponsored content and content marketing to the user generated content condition with coefficients and
standard errors (Coefficients significantly different from 0 are indicated by asterisks: ∗p < .05, ∗∗p < .01).

coefficients can be found in Fig. 3. Sponsored content seems to lead to a structured similarly to the three hypotheses examined before, so that it
more critical response from viewers, which leads to a more negative becomes possible to pin down how content marketing performs com-
brand attitude when compared to user-generated content. As content pared to sponsored and user-generated content. Overall, similarly to
marketing is also a deliberately used marketing activity, the question user-generated content, participants in the content marketing condition
arises whether the same phenomenon applies to content marketing. showed a significantly more positive brand attitude when compared to
Different from the sponsored content condition, participants in the participants from the sponsored content condition. This effect was se-
content marketing condition showed no significant difference in their rially mediated through a lower level of conceptual and a lower acti-
brand attitude when compared to participants from the user-generated vation of attitudinal persuasion knowledge, indicated through a higher
content condition. The corresponding bias-corrected bootstrap con- trustworthiness score. Participants in the content marketing condition
fidence interval for the indirect effect of content marketing as in- (aspc = −0.64) did marginally significantly differ in their conceptual
dependent variable (a2bc = −0.01) did include zero (−0.18 to 0.03). persuasion knowledge, as indicated by recognition of advertising, from
Confidence intervals for a mediation model omitting one of the two participants in the sponsored content condition (p = .06), i.e., they
mediators included zero, thus not supporting such a mediation model. displayed lower conceptual persuasion knowledge. In turn, they
This pattern of results indicates that in the perception of the viewers, showed a lower activation of attitudinal persuasion knowledge by
there seemingly is no difference between content marketing and user- rating the video as more trustworthy (bspc = −0.22, p < .01) than
generated content. people in the sponsored content condition and a more positive brand
This result begs the question of whether viewers perceive content attitude (cspc = 0.36, p < .01). A bias-corrected bootstrap confidence
marketing and sponsored content in different ways. We investigated interval with 10,000 resamples for the indirect effect (aspcbspc = 0.14)
this question on an exploratory basis, as we had no clear hypothesis a did not include zero (0.01–0.34), showing that content marketing did
priori. So far, only user-generated content was used as reference group, significantly differ from sponsored content in its brand attitude through
without a direct comparison between content marketing and sponsored the serial mediation of conceptual and attitudinal persuasion knowl-
content. To make this comparison, the same procedure for the serial edge. Additionally, there was no significant direct effect of content
mediation analysis was used, but this time sponsored content is in- marketing on attitudinal persuasion knowledge (dspc = −0.001,
cluded as reference group. To avoid confusion regarding the annota- p = .99). Looking at the complete serial mediation model, participants
tions of coefficients, all coefficients concerning the sponsored content in the content marketing condition did show a significantly more positive
condition as reference group are marked with a subscript “spc”. The brand attitude compared to the sponsored content condition through the
corresponding coefficients can be found in Table 4. The analysis is serial mediation of conceptual and attitudinal persuasion knowledge. A

52
J. Müller and F. Christandl Computers in Human Behavior 96 (2019) 46–55

bias-corrected bootstrap confidence interval with 10,000 resamples Future research should further broaden the theoretical understanding of
testing the indirect effect through conceptual and attitudinal persuasion content marketing by investigating it in a more general way instead of
knowledge on brand attitude (aspcbspccspc = 0.05) did not include zero focusing on different distribution channels and measurements in each
(0.01–0.15). Content marketing lead to a lower conceptual persuasion study. While few studies on content marketing can be compared to our
knowledge (aspc = −0.64) which in turn lead to a lower attitudinal current research, at least some comparisons can be drawn. In addition
persuasion knowledge, as indicated by a higher trustworthiness to research showing a positive relationship between content marketing
(bspc = −0.22) and then to a more positive brand attitude (cspc = 0.36). and buying behavior (Kumar et al., 2016), another study found that a
Unlike in the previous analysis using user-generated content as re- blog by a company employee is considered as more trustworthy by
ference group, the direct effect of content marketing on brand attitude consumers because there is full disclosure of the relationship as em-
was significant (f3spc = 0.52, p < .05). ployee of the company (Carr & Hayes, 2014). This could be one reason
Overall, the results related to content marketing suggest that it in- as to why participants did not rate content marketing worse than user-
fluences participants similarly to user-generated content while showing generated content although it clearly is more biased towards being
significant differences compared to sponsored content. People in the positive about the product or brand. As a direction for future research it
content marketing condition had a more positive brand attitude than should be further investigated why consumers seem to perceive content
people in the sponsored content condition, explained through the serial marketing in a more positive way. The fact that persuasive commu-
mediation of a lower conceptual persuasion knowledge and a lower nication is often intentionally disguised may lead consumers to ap-
activation of attitudinal persuasion knowledge. preciate honest and transparent communication that clearly states its
source. If this is the case, consumers might be open to the message
6. Discussion behind persuasive communication but dislike not being clearly told
about a persuasion attempt.
Overall, the pattern of results supported hypotheses 1 to 3. People in Regarding the theoretical foundation on persuasion knowledge, our
the sponsored content condition, compared to the user-generated content results support the differentiation between conceptual and attitudinal
condition, had a higher level of conceptual persuasion knowledge, persuasion knowledge found in the literature (see for instance
leading to a stronger activation of attitudinal persuasion knowledge and Rozendaal et al., 2011; Boerman et al., 2012). The fact that there was
finally resulting in a more negative brand attitude. Additionally, no no significant direct influence of conceptual persuasion knowledge on
difference between the content marketing condition and the user-gener- brand attitude backs the assumption that an attitudinal component has
ated content condition was found, whereas content marketing lead to a to be considered.
more positive brand attitude through the serial mediators than spon- Overall, the current study helps provide a link between different
sored content. In addition to these differences, the results further sup- types of content in a marketing setting. Through a serial mediation
port a serial mediation model using conceptual and attitudinal per- model, our results do not only show how content types differ from each
suasion knowledge as mediators to explain how people react to other but also through which underlying mechanisms they are per-
persuasion attempts. ceived differently. To the authors' knowledge, this study is the first to
The results of this study are consistent with the literature on spon- directly compare sponsored and user-generated content as well as
sored content, which repeatedly found disclosed sponsored content to content marketing to each other by looking at their respective psy-
be perceived less favorably (see for instance Boerman et al., 2012; chological impact. Future research should also try to apply the findings
2014; van Reijmersdal et al., 2016), as is the case in our study for a of this study to other forms of media, as some differences could possibly
comparison with user-generated content and content marketing. Our be expected. Blogs come to mind as most obvious form of media that are
findings are in line with results of other studies that found that user- often used in the context of content marketing and should be compared
generated content can positively influence consumers' opinion towards to Let's Plays. The focus of a Let's Play is very clearly on the product (the
a product and their buying behavior (see for instance Ye et al., 2011; game being played). Other forms of content incorporate a product in a
Zhu & Zhang, 2010). Additionally, this study further advances the re- more indirect way, which could lead to an even more negative response
search in this field as it does not only focus on the disclosure of spon- to sponsored content as the sponsorship may feel more alien to the
sorship but makes the next logical step by comparing content that does original focus of the content, which in turn would irritate consumers. At
not have to be disclosed (user-generated content) to content that has to the same time, content marketing could be less influential for other
be disclosed (sponsored content). In doing this, our results reflect a forms of media, as other products can only be incorporated into the
current reality concerning content. An interesting direction for future content indirectly, so that the content does not become advertising.
research could be the difference in legally allowed kinds of disclosures This may weaken the effect on the brand attitude of consumers, as the
for sponsored content on the internet, e.g. disclosure through text in the product is less prominently placed in the content. However, we can only
actual video vs. in the video description or a verbal disclosure vs. a speculate on whether there are differences between media types and
written disclosure and a combination of both. While several of the how these differences look like.
studies mentioned above investigated how different factors of dis-
closure influence persuasion knowledge and the attitude of the con- 6.1. Managerial implications
sumer, there has been no study looking at these factors in an online
setting like YouTube, where a written disclosure in front of the video The focus of this study lay on the question “Which content type is
but also auditory disclosures are possible. the best?” This study presents one possible answer that is of high re-
The findings further revealed that content marketing did not sig- levance for the marketing strategy of a company. The analysis showed
nificantly differ from user-generated content, which is very interesting, that user-generated content led to a more positive brand attitude than
as the employee that produces the content could be considered to be sponsored content. For a company this means that it should try to
even more biased than the content creator in the sponsored content motivate users to generate brand focused content by themselves, as it is
condition. As described in more detail in an earlier chapter, the lit- less likely to be perceived as advertising and therefore is considered
erature on content marketing is still sparse. Studies are often difficult to more trustworthy. The problem with this recommendation is that a
compare to each other because they focus on very specific content company does not have direct control over the creation of user-gener-
scenarios and rarely use psychological constructs to investigate the ef- ated content. This means that they cannot control whether consumers
fect of content marketing. This study closes this gap by focusing on will produce such content and additionally what this content will in-
content marketing at a broader scale and uses psychological constructs clude. The video in this study was rather favorable towards the game
to investigate the effect content marketing is having on consumers. played. Not all user-generated content is that favorable and a company

53
J. Müller and F. Christandl Computers in Human Behavior 96 (2019) 46–55

can quickly lose control over the public appearance of a brand if there is opinion leader's credibility and electronic word of mouth in two-step flow. Journal of
too much critical user-generated content. For this reason, content Interactive Advertising, 14, 38–50. https://doi.org/10.1080/15252019.2014.909296.
Chef, T.(n.d.). The Content Monarchy: Who Says "Content is King" and Why?. Retrieved
marketing seems to be a commendable alternative. This study found no from https://cognitiveseo.com/blog/216/the-content-monarchy-who-says-content-
significant difference in brand attitude between content marketing and is-king-and-why/(11 November 2018).
user-generated content. At the same time, content marketing does po- Dens, N., De Pelsmacker, P., Wouters, M., & Purnawirawan, N. (2012). DO YOU LIKE
WHAT YOU RECOGNIZE? The effects of brand placement prominence and movie plot
sitively differ from sponsored content. This seemingly makes content connection on brand attitude as mediated by recognition. Journal of Advertising, 41,
marketing the best alternative to user-generated content, as it leads to 35–53. https://doi.org/10.2753/JOA0091-3367410303.
favorable attitudes towards the product and the company has complete Dickinger, A. (2011). The trustworthiness of online channles for experience- and goal-
directed search tasks. Journal of Travel Research, 50, 378–391. https://doi.org/10.
control over the content produced. Its only downside might be that the 1177%2F0047287510371694.
effort for a company to produce its own content is higher than to pay Friestad, M., & Wright, P. (1994). The persuasion knowledge model: How people cope
someone else in a sponsored format. Managers and marketing practi- with persuasion attempts. Journal of Consumer Research, 21, 1–31. https://doi.org/10.
1086/209380.
tioners have to weigh whether a more positive brand attitude and more
Gates, B. (1996). CONTENT IS KING BY BILL GATES. Retrieved from: https://www.
control is worth the extra effort compared to sponsored and user-gen- craigbailey.net/content-is-king-by-bill-gates/, Accessed date: 11 November 2018.
erated content. Ha, S. H., Bae, S.yong, & Son, L. K. (2015). Impact of online consumer reviews on product
scales: Quantitative analysis of the source effect. Applied Mathematics & Information
Sciences, 9, 373–387. https://doi.org/10.1016/j.dss.2018.05.006.
7. Conclusion Hayes, A. (2013). Introduction to mediation, moderation, and conditional process analysis - a
regression-based approach. New York: The Guilford Press.
The research on different content types is still in its early stages. Hayes, A., & Preacher, K. (2014). Statistical mediation analysis with a multicategorical
independent variable. British Journal of Mathematical and Statistical Psychology, 67,
While there is a good foundation concerning sponsored content and 451–470. https://doi.org/10.1111/bmsp.12028.
user-generated content, the research on content marketing is still Holliman, G., & Rowley, J. (2014). Business to business digital content marketing:
lacking and comparisons between content types are virtually non-ex- Marketers' perceptions of best practice. The Journal of Research in Indian Medicine, 8,
269–293. https://doi.org/10.1108/JRIM-02-2014-0013.
istent. This study provides some insights on the differences between Hwang, Y., & Jeong, S.-H. (2016). This is a sponsored blog post, but all opinions are my
content types and how they affect brand responses. Compared to user- own": The effects of sponsorship disclosure on responses to sponsored blog posts.
generated content, sponsored content leads to a higher conceptual Computers in Human Behavior, 62, 528–535. https://doi.org/10.1016/j.chb.2016.04.
026.
persuasion knowledge, resulting in a higher attitudinal persuasion Janssen, L., Fransen, M., Wulff, R., & van Reijmersdal, E. (2016). Brand placement dis-
knowledge and a more negative brand attitude. Interestingly, content closure effects on persuasion. The moderating role of consumer self-control. Journal
marketing does not differ in this way from user-generated content but of Consumer Behaviour, 15, 503–515. https://doi.org/10.1002/cb.1587.
Järvinen, J., & Taiminen, H. (2016). Harnessing marketing automation for B2B content
does so from sponsored content, even though content marketing could
marketing. Industrial Marketing Management, 54, 164–175. https://doi.org/10.1016/j.
be seen as even more biased, as the source of the content is the com- indmarman.2015.07.002.
pany behind the brand. Jefferson, S.(n.d.). Why content is king in today's marketing. Retrieved from http://www.
The results of this study indicate that companies should consider marketingdonut.co.uk/online-marketing/content-marketing/why-content-is-king-in-
today-s-marketing (11 November 2018).
creating their own content as marketing method, as it seemingly is Kee, A., & Yazdanifard, R. (2015). The review of content marketing as a new trend in
perceived more favorably than sponsored content but is still more marketing practices. International Journal of Management, Accounting and Economics,
controllable than user-generated content. Future research should focus 2, 1055–1064.
Kilgour, M., Sasser, S., & Larke, R. (2015). The social media transformation process:
on replicating the effects found in this study for other forms of media Curating content into strategy. Corporate Communications: An International Journal,
and different brands and product categories to determine their applic- 20, 326–343. https://doi.org/10.1108/CCIJ-07-2014-0046.
ability for different industry sectors. The theoretical foundation ex- Kim, J. (2012). The institutionalization of YouTube: From user-generated content to
professionally generated content. Media, Culture & Society, 34, 53–67. https://doi.
plaining why content marketing is perceived as well as it seemingly is, org/10.1177%2F0163443711427199.
is still lacking and has to be expanded through further research. In this Kumar, A., Bezawada, R., Rishika, R., Janakiraman, R., & Kannan, P. K. (2016). From
context, it is advisable to focus more on the psychological effects among social to sale: The effects of firm-generated content in social media on customer
behavior. Journal of Marketing, 80, 7–25. https://doi.org/10.1509/jm.14.0249.
consumers than on the point of view of marketing practitioners, as is
Lee, M., & Youn, S. (2009). Electronic word of mouth (eWOM) How eWOM platforms
currently the case in content marketing literature. The present study is a influence consumer product judgement. International Journal of Advertising, 28,
first step in that direction. 473–499. https://doi.org/10.2501/S0265048709200709.
Ohanian, R. (1990). Construction and validation of a scale to measure celebrity endorses'
perceived expertise, trustworthiness, and attractiveness. Journal of Advertising, 19,
Financial disclosure 39–52. http://psycnet.apa.org/doi/10.1080/00913367.1990.10673191.
Rossiter, J. (2011). Marketing measurement revolution - the C-OAR-SE method and why it
There are no financial conflicts of interest to disclose. must replace psychometrics. European Journal of Marketing, 45, 1561–1588. https://
doi.org/10.1108/03090561111167298.
Rozendaal, E., Lapierre, M., van Reijmersdal, E., & Buijzen, M. (2011). Reconsidering
Acknowledgements advertising literacy as a defense against advertising effects. Media Psychology, 14,
333–354. https://doi.org/10.1080/15213269.2011.620540.
Shavitt, S., Lowrey, P., & Haefner, J. (1998). Public attitudes toward advertising: More
We thank the YouTuber Aavak for the permission to use his video as favorable than you might think. Journal of Advertising Research, 38, 7–22.
a part of this study. We also thank the developer Lion Shield for the Spears, N., & Singh, S. (2004). Measuring attitude toward the brand and pruchase in-
permission to use their game as a part of this study. tentions. Journal of Current Issues and Research in Advertising, 26, 53–66. https://doi.
org/10.1080/10641734.2004.10505164.
Tafesse, W. (2015). Content strategies and audience response on Facebook brand pages.
References Marketing Intelligence & Planning, 33, 927–943. https://doi.org/10.1108/MIP-07-
2014-0135.
Tang, T., Fang, E., & Wang, F. (2014). Is neutral really neutral? The effects of neutral user-
Ayeh, J., Au, N., & Law, R. (2013). Do we believe in TripAdvisor?" examining credibility
generated content on product sales. Journal of Marketing, 78, 41–58. https://doi.org/
perceptions and online travelers' attitude toward using user-generated content.
10.1509/jm.13.0301.
Journal of Travel Research, 52, 437–452. https://doi.org/10.1177%
textbroker. (n.d.). Content is King. Retrieved from https://www.textbroker.com/content-
2F0047287512475217.
king (11 November 2018).
Boerman, S., van Reijmersdal, E., & Neijens, P. (2012). Sponsorship disclosure: Effects of
Tobak, S. (2016). Why 'content is king' is a myth. Retrieved from http://www.foxbusiness.
duration on persuasion knowledge and brand responses. Journal of Communication,
com/features/2016/08/01/why-content-is-king-is-myth.html, Accessed date: 11
62, 1047–1064. https://doi.org/10.1111/j.1460-2466.2012.01677.x.
November 2018.
Boerman, S., van Reijmersdal, E., & Neijens, P. (2014). Effects of sponsorship disclosure
van Reijmersdal, E., Neijens, P., & Smit, E. (2007). Effects of television brand placement
timing on the processing of sponsored content: A study on the effectiveness of eur-
on brand image. Psychology and Marketing, 24, 403–420. https://doi.org/10.1002/
opean disclosure regulations. Psychology and Marketing, 31, 214–224. https://doi.
mar.20166.
org/10.1002/mar.20688.
van Reijmersdal, E., Fransen, M., van Noort, G., Opree, S., Vandeberg, L., Reusch, S., et al.
Carr, C., & Hayes, R. (2014). The effect of disclosure of third-party influence on an
(2016). Effects of disclosing sponsored content in blogs: How the use of resistance

54
J. Müller and F. Christandl Computers in Human Behavior 96 (2019) 46–55

strategies mediates effects on persuasion. American Behavioral Scientist, 60, Advertising, 45, 157–168. https://doi.org/10.1080/00913367.2015.1115380.
1458–1474. https://doi.org/10.1177/0002764216660141. Ye, Q., Law, R., Gu, B., & Chen, W. (2011). The influence of user-generated content on
West, T. (2015). 5 reasons why content is king. Retrieved from https://scrunch.com/blog/ traveler behavior: An empricial investigation on the effects of e-word-of-mouth to
5-reasons-why-content-is-king/, Accessed date: 11 November 2018. hotel online bookings. Computers in Human Behavior, 27, 634–639. https://doi.org/
White, A. (2016). Why "content is king" is the biggest myth in SEO. Retrieved from https:// 10.1016/j.chb.2010.04.014.
www.semrush.com/blog/why-content-is-king-is-the-biggest-myth-in-seo/, Accessed Zhu, F., & Zhang, X. (2010). Impact of online consumer reviews on sales: The moderating
date: 11 November 2018. role of product and consumer charactersistics. Journal of Marketing, 74, 133–148.
Wojdynski, B., & Evans, N. (2016). Going native: Effects of disclosure position and lan- https://doi.org/10.1509/jmkg.74.2.133.
guage on the recognition and evaluation of online native advertising. Journal of

55