WHETHER
OR NOT
CONTENT
IS KING
An examination of content marketing
and the effects of advertising
recognition, persuasion knowledge
and resistance towards persuasion on
brand attitude and purchase intention.
Whether or not content really is king
An examination of content marketing and the effects of advertising recognition, persuasion knowledge and resistance
towards persuasion on brand attitude and purchase intention. Vera A. Wijermars
10165290 Master’s Thesis
Graduate School of Communication Master’s programme Communication Science
Supervision by E. Peelen February 1, 2019
Abstract
Content marketing, creating valuable content to attract and retain consumers (Pulizzi, 2014), developed from a trend to a marketing practice. However, this practice is currently academically understudied and not yet fully established in the Dutch market. This research aims at providing insight in the persuasive effect of content marketing. Based on the
Persuasion Knowledge Model (PKM) (Friestad & Wright, 1994) it is expected that people do not recognise content marketing as advertising. Therefore, their knowledge about the persuasion attempt is not activated, they will resist the content less and consequently have a more positive attitude towards the communicating brand and a higher intention to
purchase. A preliminary qualitative study is conducted to develop experimental materials and to explore how content marketing is currently experienced and practiced in the field. Expert interviews (N = 4) indicate that it is important to create relevant and valuable content that foremost serves the consumers’ needs, but eventually results in sales and is also search engine proof. The brand does not need to be prominent in the content, because content marketing is distributed via owned media channels. In an online experiment (N = 77) the effects of a content marketing website are compared to a website with a more traditional marketing approach focused on selling. Contrary to the hypotheses, the results only show an effect of content marketing on persuasion knowledge (PK), mediated by whether or not the participants recognised the advertising. The recognition of advertising negatively affects brand attitude, mediated by a credibility dimension of PK. Activated credibility PK results in more affective resistance, mediated by cognitive resistance, and in a lower intention to purchase, mediated by brand attitude. This study thus suggests that the PKM holds in a content marketing context. Interestingly, resistance does not result in more negative
attitudes and intentions. It seems that consumers do not mind persuasive attempts, as long as the content is credible enough. This implicates that a healthy balance between credible content and commerciality might be crucial for successfully applying content marketing. Keywords: content marketing, advertising recognition, persuasion knowledge, persuasion, resistance.
Table of contents
Whether or not content really is king 4
Theoretical background 6
Antecedents of content marketing 6
Defining content marketing 7
Resistance towards persuasion and content marketing 9
The role of persuasion knowledge 10
Effects of content marketing 11
Presenting a conceptual model 12
Preliminary study: Expert interviews 13
Method 13
Results 14
Conclusion and discussion 19
Main study: Online content experiment 20
Method 21
Results 26
Conclusion and discussion 32
General discussion 34
Limitations and future research 35
References 36
Appendix A.: Interview guide 40
Appendix B.: Codelist 41
Whether or not content really is king
‘Content is king’ is a common sentence in the marketing and advertising industry. This phrase is often associated with content marketing. Nowadays content marketing is no longer just a hype or buzz word. The Dutch search-trends of the past ten years show a peak popularity around 2017, after which the trend line is going down the slope (Google Trends, n.d.; Figure 1.). Content marketing is applied by more and more marketers, organisations and brands (Dubbeld, 2018). It is about time, because content marketing is not a new technique at all. One of the first known examples comes from 1895, when an agricultural company decided to publish its own magazine: The Furrow (Pulizzi, 2014). This magazine informed farmers on relevant topics and educated them as well. The company did not try to sell their products directly through the content of the magazine, which differentiated it from the typical distribution of catalogues. That is the core of content marketing: creating and distributing valuable content that customers want to consume, while communicating without selling to attract and retain loyal consumers (Pulizzi, 2014).
Figure 1. Trends for content marketing in the Netherlands.
Note: Google trends for the topic and keyword content marketing over past 10 years in the Netherlands.
Modern day examples of content marketing are often found in large brands, such as Coca Cola, that focusses on telling stories; IKEA, which wants to inspire consumers on its dedicated platform; or Red Bull, that turned into a complete media company to produce extreme sports content. Those big players might be in the luxury position of being able to step away from a selling oriented marketing strategy while still selling their products.
However, even for smaller brands and companies, content marketing can be a great solution for the declining effectiveness of traditional marketing techniques (Holliman & Rowley, 2014). It gets harder each day to stand out in a cluttered media landscape. Consumers are increasingly getting used to or are avoiding traditional advertising and marketing techniques. Content marketing can be a way for advertisers to cope with this resistance, because this type of marketing is not directly focussed on trying to sell something or being recognised as persuasive by consumers. Rather, the brand focusses on becoming an expert in a specific niche. This results in consumers actively looking up the brand’s content marketing instead of advertisements being forced upon them.
Although, many people still do not know how to apply it in practice despite the successful cases and the fact that content marketing has been around for years now. Especially in the Dutch market, content marketing is not yet established (Dubbeld, 2018). Moreover, the topic of content marketing is presumably still almost non-existent in academic literature. That is why this study aims to contribute to the current academic knowledge.
Is content really king? To explore this question and gain more insight, expert
interviews will be conducted to explore how content marketing is applied in the advertising and marketing industry. This preliminary study will be followed by an experiment. The goal of the experiment is to provide insight in the underlying mechanism that causes effect-differences between content marketing and traditional marketing and aims to answer the following research question:
RQ: What is the persuasive effect of content marketing (B2C), compared to traditional marketing, on brand attitude and purchase intention and to what extent does resistance towards persuasion play a role in this effect?
This study is relevant for society because the findings can provide insight in a possible solution for the current problems related to clutter in advertising. The study has a big
scientific value as well, because Kerkhof orated that there is a clear gap between practice and science (2012). The topic of content marketing is still academically under-studied at the moment. Although there is some exploratory research available, most articles are focussed on business to business (B2B) practices only and not on business to consumer (B2C) applications of content marketing (e.g. Holliman & Rowley, 2014; Järvinen & Taiminen, 2016).
Theoretical background
This literature study provides an overview of what is known in the academic literature about content marketing, its relationship with resistance towards persuasion and its effects. Content marketing is a specific marketing practice and it is therefore important to start with an overview of content marketing in general.
Antecedents of content marketing
The professorship in content marketing at the University of Amsterdam took over from the chair in customer media. This indicates that customer media might be the basis of the current practice. When Peter Kerkhof was named professor by special appointment of customer media, he describes customer media as a means to persuade consumers into voluntary consumption of content with marketing oriented goals, whereby brands and organisations utilise owned media channels (Kerkhof, 2012). This practice used to be executed via the spread of customer magazines. Van Reijmersdal, Neijens and Smit (2010) define customer magazines as content that is created by advertisers with no clear
boundaries between the commercial message and editorial content. They also point out that customer magazines’ content may vary in the brand’s prominence. A study by Bronner and Suèr (2001) provides an overview of 20 years of research into customer magazines. The authors present a model with the five A’s of customer magazines: audience, attention, attitude towards the brand, adherence and active experience. They conclude that customer magazines are able to approach a target group with tailored information and may therefore better reach the consumer and grab their attention. Besides that, customer magazines can
create and positively impact brand associations and contribute to consumer loyalty and adherence to the brand. In order to achieve this it is important that the magazine informs and entertains.
Not only customer magazines are part of customer media. According to the oration of Kerkhof (2012), newsletters, blogs or branded content distributed via owned media are also examples of this category. Other concepts related to content marketing are brand- and product-placement, native- and covert-advertising. Brand- and product-placements are sponsored content, where the commercial message is integrated into the editorial content (Tutaj & Van Reijmersdal, 2012). Wei, Fischer and Main (2008) see sponsored content as a covert form of marketing and advertising. Covert advertising embeds brands into contexts that are not commonly used for advertising, for example integrating advertising and entertainment (Evans & Park, 2015). Native advertising is another term for sponsored content, which is paid advertising that looks like editorial content and the format of the publisher (e.g. Couldry & Turow, 2014; Wojdynski & Evans, 2016).
Defining content marketing
Content marketing in academic literature. One of the first academics to use the term content marketing was Kerkhof. In an article about customer media, the author describes content marketing as a counterpart of traditional advertising – which is focused on the
promotion of a brand’s products or service. The characteristics of content marketing however, are relationship building with (future) customers and providing them with relevant content, mostly in an online environment (2010). Another article about content marketing, describes both a literature review and semi-structured interviews with professional key informants. Some interviewees indicated at the time that the topic was relatively new and the definitions were poorly understood under marketers or not consistent (Holliman & Rowley, 2014). Some key characteristics of content marketing deriving from this research are attracting and
retaining customers with relevant content that they want to receive from brands that become publishers, by telling good and engaging stories. Holliman and Rowley (2014) conclude their article with the first conceptual definition of content marketing based on the interviews and grounded in empirical research:
“B2B digital content marketing involves creating, distributing and sharing relevant, compelling and timely content to engage customers at the appropriate point in their buying consideration processes, such that it encourages them to convert to a business building outcome.” (p. 23).
Peelen later also orates that content marketing includes delivering relevant and valuable content to a specific audience via owned media. To reach this audience, a brand needs to act as a publisher (2017).
Content marketing in industry literature. In practice, however, the definition of content marketing by the Content Marketing Institute (CMI) is popular. The founder of the CMI also describes content marketing as attracting and retaining customers through
storytelling. Therefore brands need to behave like media agencies (Pulizzi, 2012). According to Pulizzi (2014), the formal definition of content marketing is:
“Content marketing is the marketing and business process for creating and distributing valuable and compelling content to attract, acquire, and engage a clearly defined and understood target audience – with the objective of driving profitable customer action.” (p. 5).
This definition has common ground with the version by Holliman and Rowley (2014). However, the goals of content marketing reach beyond engaging customers only.
Liebrecht (2015) compares the academic definition from Kerkhof and the practitioners definition held by Pulizzi. The two differ mainly regarding the goal of content marketing. Kerkhof (2010) states that content marketing is not only focused on promoting products and therefore sales, while Pulizzi’s (2014) objective is profit. Liebrecht (2015) explains this disagreement through the different fields in which content marketing is executed. In public relations the focus is on relationships and mutual understanding, which fits the first definition, while in marketing the goal is persuasion through promotional messages, which better suits the ideas of Pulizzi. The holistic view held in this publication resulted in three main
characteristics of content marketing. The practice is focused on receivers, is goal centred, and utilises owned media (Liebrecht, 2015). Others also use a combination of the academic definitions and refer to work by Pulizzi (e.g. Antheunis, Van Kaam, Liebrecht, & Van Noort, 2016; Järvinen & Taiminen, 2016). For the purpose of this thesis, content marketing is defined by Pulizzi’s formal definition stated above (2014), to which is added that the content needs te be distributed via owned media.
As suggested by Holliman and Rowley (2014), content marketing offers a solution to the declining effectiveness of traditional marketing techniques. Content marketing can complement or offer an alternative for traditional marketing, because it is oriented on the receivers and focussed on pulling them towards the content. Traditional marketing on the contrary, is more about pushing content and this can result in resistance from the audience (Peelen, 2017).
Resistance towards persuasion and content marketing
When people experience a reaction against change, it is called resistance. Resistance in that sense is being able to withstand a persuasive attempt by an advertiser. The cognitive, affective and behavioural elements of resistance can result from pressures of change, as well as being a motivational state to counter those pressures. The concept of reactance is an external force of resistance. This uncomfortable feeling arises when people have the feeling that their freedom of choice is being threatened. People resist advertising because this feeling of reactance motivates them to restore their freedom of choice (when the freedom is important enough and the threat is direct). Besides reactance, other faces of resistance are general distrust towards persuasive messages, general scrutiny activated by awareness of influence attempts and inertia – focus on keeping existing attitudes intact instead of resisting the change (Knowles & Linn, 2004). Fransen, Verlegh, Kirmani and Smit (2015) follow this idea and developed a typology for resistance strategies that people employ. Avoiding advertising is one way of resisting a persuasive message. People can avoid in cognitive, behavioural or mechanical ways, which are reactions to cluttered advertising. They avoid advertisers’ push strategies and prefer a pull strategy, where they can select the information to consume themselves (Kelly, Kerr, & Drennan, 2010). Results of focus groups and
interviews show that teenagers are more likely to avoid advertising on social media when they have prior negative experiences, the ads are not relevant to them personally and when they are skeptical towards the medium and the message (Kelly et al., 2010). The second type consists of contesting strategies, where people challenge and refute the information in the ad by countering the content and the source of the message or the persuasive tactics that are applied (Fransen et al., 2015). The third strategy type is empowerment, which is
related to the receiver instead of the content. By bolstering their pre-existing attitudes, seeking social validation by others or by self asserting reminders of their confidence in attitudes, people help themselves in reassuring their existing attitudes (Fransen et al., 2015). That is why it is expected that the pressure to change, inherent to a persuasive message, motivates people to resist that message and employ resistance strategies.
The role of persuasion knowledge
Fransen et al. (2015) explain contesting strategies based on the Persuasion Knowledge Model (PKM) by Friestad and Wright. Persuasion knowledge is constantly developing and acquired through experiences and observations. This knowledge is immediately available and, when activated, affects responses people have towards
advertising shaped by three interacting structures: persuasion knowledge, agent knowledge – beliefs people hold about specific traits, goals and competences of the perceived sender – and topic knowledge – beliefs people hold about the message’s topic. When the first is activated, the other two will be activated subsequently at some point (Friestad & Wright, 1994). The awareness that the advertiser tries to make a persuasive attempt on them, makes people want to refute the message with counter arguments (Fransen et al., 2015).
No research on the influence of content marketing on the PKM is found. However, the succes of native advertising, a form of sponsored content, might be assigned to consumers not knowing that they are consuming commercial content. According to the PKM, the recognition of advertising affects the attitudinal evaluations and intended behaviours (Wojdynski & Evans, 2016). Experimental results show that not recognising an ad has positive effects on the perceived news credibility.
Based on the PKM it is thus expected that when people do not recognise advertising and their persuasion knowledge is not activated, they will resist the content less.
H1: Content marketing, compared to traditional marketing, is less likely to be recognised as advertising.
H2: The recognition of advertising, compared to when it is not recognised, is more likely to result in activation of persuasion knowledge.
H3: The activation of persuasion knowledge, compared to when it is not activated, makes people more likely to resist.
Effects of content marketing
Attitudes and intentions. Based on the PKM, Van Reijmersdal et al. (2010) predict that the perceived credibility and persuasion attempts of customer media are affected by the commerciality of those magazines. This in turn might have an effect on consumers’ attitudes towards the content. Results of their experiment show that more commercial magazines are perceived as more persuasive and less credible, which leads to more negative evaluations. Another study compared editorials, advertorials and advertisements in print. The results in this experiment also show that the perceived credibility of editorials is higher than for the more obvious commercial advertorials and ads (Attaran, Notarantonio, & Quigley, 2015). Besides that, people have a higher purchase intention for products shown in editorial context compared to advertorials and ads. Liebrecht (2015) specifically looked at the credibility of different types of content marketing. The results of her experiment show that informative or instructive content is perceived as more credible compared to entertaining content.
Moreover, texts are perceived as more credible and the brand as more trustworthy, compared to video formats.
Therefore, according to the PKM it is to be expected that when an ad is not recognised and persuasion knowledge is not activated, people are likely to experience more positive evaluations and higher behavioural intentions, because content marketing is perceived as less commercial and thus more credible. Especially in the context of informative and instructive texts.
H4a: Content marketing, compared to traditional marketing, is likely to result in a more positive attitude towards the advertised brand.
H4b: Content marketing, compared to traditional marketing, is likely to result in a higher intention to purchase the advertised brand.
Advertising recognition and persuasion knowledge. In an experiment that
compared sponsored content with a more traditional way of advertising (banner ads), Tutaj and Van Reijmersdal (2012) found that sponsored content is perceived as more informative, more amusing and less irritating. Banner ads on the contrary evoke a higher persuasion knowledge and are more recognised as advertising. In turn, the recognition of online native advertising results in more negative evaluations (Wojdynski & Evans, 2016).
Wei et al. (2008) expect that the perceived appropriateness of covert marketing and product placements plays a role. The results of their multiple experiments with covert advertising on the radio point at a negative impact on brand evaluations when participants are aware of the marketing tactic – when they have persuasion knowledge. They note however that activation of persuasion knowledge not always has a negative effect. If the perceived appropriateness of the persuasive tactic and the familiarity with the brand are high, the negative effects can decrease or even turn positive.
The studies mentioned researched different media and different types of marketing and advertising, but they all conform the idea that recognition of the advertising and activation of persuasion knowledge results in more negative outcomes. It is expected that the type of marketing affects the advertising recognition and consequently the activation of persuasion knowledge and resistance. Additionally, attitudes can help predict behaviour (Spears & Singh, 2004). Attitude and behavioural intentions are related when the target and action elements correspond (Ajzen & Fishbein, 1977). That is why the following mediation-hypothesis is proposed:
H5. Content marketing, compared to traditional marketing, results in a more positive attitude towards the brand and a higher purchase intention mediated by advertising recognition, activation of persuasion knowledge and resistance towards persuasion.
Presenting a conceptual model
Based on the PKM a conceptual model is hypothesised (Figure 2.) in which exposure to content marketing would result in less recognition of the advertising, after which
persuasion knowledge is not activated, resulting in no resistance towards the persuasion and consequently more positive attitude towards the brand and a higher purchase intention. On the contrary, traditional marketing is recognised as advertising, activating persuasion
Figure 2. Conceptual model.
Note: The hypothesised mediated pathway.
knowledge and therefore evokes resistance, which results in a more negative brandattitude and a lower purchase intention.
Preliminary study: Expert interviews
To explore how content marketing is currently being experienced and practiced in the advertising and marketing field a preliminary study was conducted. This qualitative research aimed to answer the following research question:
What are the perceptions and experiences of, and attitudes towards content marketing held by advertising and marketing practitioners and how do they use content marketing in their jobs?
Method
Semi-structured interviews. Qualitative research in the form of semi-structured expert interviews was conducted for the exploration of this field. This was an appropriate method, because semi-structured interviews are flexible and allow to explore the interviewees’ point of views and perspectives. Furthermore, semi-structured interviews enable the addressing of specific issues – such as content marketing –, while the general formulation of the questions makes rich and detailed answers possible (Bryman, 2012).
Interviewees. The participants (N = 4) were recruited from within the professional network of the researcher. The interviewees all work in the Netherlands and are each in some way engaged with content marketing in their job. Interviewee #1 is an online marketer for an online marketing agency, interviewee #2 is the director of an art fair, interviewee #3 works as a content marketer for an energy comparator for businesses, and interviewee #4 is a brand specialist for an e-commerce company. The interviewees work in different
companies, which resulted in insights from a variety of job positions and B2B- and B2C-oriented markets. Other differences were found in the total years of work experience of the interviewees (M = 7.75, SD = 8.18), their years of experience with content marketing specifically (M = 3.00, SD = 1.41). The sample also represented variety in age (M = 33.00, SD = 9.42), and an equal distribution of gender (2 males, 2 females). They have a
Procedure. People interested in partaking the study were contacted and informed that the interview would take up about 30 to 45 minutes, in which they could talk about their experiences with content marketing. After agreement to participate, an appointment was made and a link to a short online survey was sent to the interviewees. Via the survey they provided demographic data and gave their informed consent. The accompanied text showed them the definition of content marketing used in this study beforehand.
The final interviews that provided the data for this preliminary study were conducted between November 28 and December 6 2018. The appointments were all in meeting rooms at the offices of the interviewees or the researcher, in Amsterdam, Hoofddorp and Utrecht. During the interviews the researcher took notes and the interviews were also recorded on a mobile phone, using Apple’s Voice Memos application, to enhance the reflexivity of the researcher and the validity and reliability of the study. The total length of the interviews varied between 30 and 38.5 minutes. The interviews were guided by a predefined topic list
(Appendix A.). The topics were introduced by the researcher by means of general questions that invited the interviewee to freely talk about the topic. Where needed, the researcher asked questions to get more detailed answers and to make sure all the elements of the topics were discussed. Subsequently, the recordings were transcribed using the AI-powered software Trint, after which the transcripts were open and axial coded in Atlas.ti by the
researcher (see codelist, Appendix B.). By implementing a CAQDAS strategy, the trustworthiness of this study was enhanced.
Results
The interviews are discussed, based on the topics of the interview guide. Definition of content marketing. The sensitising concept of the first topic was ‘definition of content marketing’. The interviewees were encouraged to talk about their own perceptions of content marketing, how they applied this practice in their jobs, and how content marketing is used within their organisation. The definition of content marketing used in this study, was read aloud by the researcher. After which the interviewees were asked to comment on the given definition.
The interviewees were mostly in agreement with the definition of content marketing and some talked about the aspect of indirect sales. One interviewee compared content marketing with traditional marketing and pointed out that, among others, the branding aspect could indirectly lead to sales, while traditional marketing is more focussed on direct sales. Other interviewees pointed out that the relevance of the content is important and that the content should respond to the customers’ needs and wants. By creating a loyal following and satisfying those needs, you can eventually sell more.
Interviewee #1: “Traditional marketing is very much about direct sales. You create something that should lead directly to sales. Content marketing can also just be informative, […] but then create a certain experience or weave a kind of branding into it, which then eventually can indirectly lead to sales.”
Interviewee #2: “That is exactly what I did. It [content marketing] is indeed
reassessing the relevance. Making sure that you create a certain warm feeling in your audience, so they know you and keep following you.”
Interviewee #4: “Like we say: 'Content should arise from the customer's needs’ […] and eventually, the goal of course is that you want people to take action.”
Interviewee #3 mentioned that content marketing is all about a sustainable relationship with your customers, besides the endgoal of a conversion. She also agreed with the
definition, but wanted to add something:
Interviewee #3: “So I think SEO also really is an important part of content marketing. […] With valuable content certainly.”
Search engine optimisation (SEO) is a practice that the interviewees apply in their content marketing activities. Interviewee #4 implements SEO, because using the right keywords the e-commerce company can serve customers with the right content at the right time in the customer journey. SEO is also an important motivation for the interviewees to write blogs and content for their corporate websites.
Interviewee #4: “Content marketing, yes I always immediately associate that with SEO, because that is something I am very busy with.”
Interviewee #1: “I have written a lot of blogs […] just to have content on the website.”
Another motivation for some interviewees to execute content marketing is becoming an expert.
Interviewee #3: “If you have qualitative content […] you really are an authority in your field, specialised in something, then you really generate customer trust I think.”
On the contrary, interviewee #2 mentioned that content marketing is not about claiming expertise, but to inform and help customers with content.
Interviewee #2: “We do not say that we know everything, but we do say where to find it [the information] and we can help you find it.”
Forms and types of content. The second topic covered the different types of content the interviewee has used in their work, and how commercial this content was or how
prominent the brand was in this content.
As mentioned in the first topic, the interviewees generate mostly blog and website content in their jobs. Depending on the type of company they work for, this content is created for their company’s or their clients’ corporate website. Other forms of content mentioned by the interviewees were infographics, videos, images, newsletters and social media.
Newsletters were used to inform an already engaged audience and to push them to the website. For one interviewee, branding plays an important part in the newsletters, specifically showcasing the brands of the exhibitors at a fair.
Interviewee #2: “They get double: A, they are publicised; and B, they see that what [the content] they deliver is used for the public.”
Social media was mostly used as a platform for distributing the created content. Interviewee #3 mentioned that she posts blogs on social media. At the moment it is mainly one-way communication, but she would like to create more engagement and associations with the brand in the future. However, the company interviewee #4 works for utilised Facebook mainly for paid advertising. The focus on social media was on the overall brand and not the product categories, by responding to current events in a humorous way to create reach.
Video content was used in an infotainment way or to initiate a conversation with possible partners. Videos can add value by supporting the text or vice versa. The e-commerce company uses this type of content on specific product pages for instructive reasons. One interviewee pointed out that she would like to create professional videos but
that this process is costly. Interviewee #2 confirms this threshold, by saying it was a real investment for her to develop video content.
Regarding the commerciality and the prominence of the brand, the interviewees made some interesting remarks. At the company interviewee #1 worked for, brand awareness and recognition are crucial. In every type of content they created the brand name and logo were present in a prominent way. For interviewee #2 the brand was also leading, only in a more subtle way: by displaying the core values of the brand in the content. Depending on the content, adding a logo could even be counterproductive. She explained:
Interviewee #2: “If it [the content] was without branding, because it comes from our channels, […] then it was shared and you got a lot more comments”. The mission of the brand – transparency in the market – was guiding for interviewee #3 as well, but she did add links and buttons to her brand’s product in the content. In the content created by interviewee #4, apart from the logo on the website itself, the focus was more on the brands the company sells. On commerciality he commented:
Interviewee #4: “Of course it is commercial, in the sense that you want people to buy the product, but […] it should just be honest content as well.”
Interviewee #1 also mentioned that to stand out on the internet it is important to have uniform communications, and companies should adopt a branding strategy. This is related to using the content to display the brand’s values. Overall, a healthy balance between valuable content and the presence of a brand appears to be important:
Interviewee #3: “Content marketing is also that […] people go to your website, because they want to know something about a certain subject or something. And then you should not immediately push with commercial messages, for example.” Goals of content marketing. For the third topic, the interviewees were asked to discuss the different goals for which they have used content marketing, and whether specific types of content work better for a certain goal. The interviewees spoke briefly of goals in the previous topics. Branding, brand awareness, conversion, informing, better search results and sales as endgoal were already brought up.
All interviewees indicated that employing content marketing serves multiple goals. Interviewee #1 thought that blog content suited goals regarding website traffic, link value and
authority the best. In the future he would focus more on branding and creating uniformity in all types of content in a strategic way. Interviewee #2 mentioned that selling tickets is always the main goal, which all content efforts try to serve. Besides that, there are secondary goals that interact. Interestingly, she distinguishes between business-to-business (B2B) and business-to-consumer (B2C) target groups. More popular content has a certain relevance and shows exactly who the brand is and what it stands for. This also attracts new exhibitors in a subtle way, without having to argue about the arrangements. Interviewee #3 also created content for both B2B and B2C during her career. She mentioned that for B2B more
qualitative content is important because they do not have much time to read, while consumers read content from a personal interest in the topic. Besides that, service is an important goal, especially via blogs and e-mail content:
Interviewee #3: “[Content] actually facilitates the client's customer journey. If they wonder about something, […] you can explain that.” “It is a bit of service towards your customers.”
In the end, goals of content marketing are really depending on the specific situation, the target group and the current stage in the customer journey:
Interviewee #4: “It depends, at certain moments it is more brand awareness that you want to create, when there is a new product launch or a new brand for example. You have to first make sure that people know it anyway, otherwise you will not be able to sell it.”
Effects of content marketing. In the last topic the interviewees could indicate whether they noticed any effects triggered by their content marketing efforts.
In terms of SEO, content marketing did have an explicit effect on data regarding
domain values and backlinks for interviewee #1. Interviewee #2 also contributed the annually increasing number of paying visitors to her content marketing efforts. For interviewee #3 the topics of the content mainly trigger the effects:
Interviewee #3: “Current topics and money. That is what triggers people often to read something. Those blogs also get many reactions.”
Content of a high quality always has an effect, according to interviewee #4. It contributes to the company’s revenue, but also to more website traffic because of better
positions in search results. When you have a lot of competitors, content can also be a crucial deciding factor:
Interviewee #4: “If you miss just that image that shows what the product exactly does or certain information that the customer would not have otherwise: therefore it could be that they do not proceed to purchasing.”
Conclusion and discussion
This explorative study gained valuable insights in the use of content marketing in the field. The interviewees all agreed that distinctive content should be valuable, relevant, satisfying the needs of the consumer and something that the audience wants to consume themselves. Serving the content at the right moment in the customer journey is also crucial, which is in line with Pulizzi’s ideas (2014). Furthermore, the interviewees indicated that SEO is also an important part of, and even a motivation to apply, content marketing. This is not at all surprising, because both marketing tactics are applied to attract consumers and turn them into loyal consumers to retain them (Dumitrescu, Fuciu, & Gorski, 2018)
In their jobs, the interviewees mainly worked with blogs and on-page website content, which were mostly SEO proof as well. The texts were often supported by visuals, such as images, videos or infographics.
Sales is an important indirect end goal of content marketing, so the content is always branded or commercial at some level. This fits the definition by Pulizzi (2014). However, despite this commercial driving force, a healthy balance between the prominence of the brand and the content itself is desirable. Most interviewees mentioned that because the content is published on channels owned by the brand, the sender is already clear. The brand’s logo is always visually present on the corporate website for example. Some interviewees brought up that the brand can be also present in a more subtle way. The
content could express the values or the mission of the brand, which are both integral parts of a brand’s identity. These ideas are in line with the literature, which suggests that
commerciality content affects the extend to which content seems more persuasive and is therefore more negatively evaluated (Van Reijmersdal et al., 2010).
Having said that, one interviewee used content marketing purely for creating brand awareness. Thus, the brand was eminently prominent in the content he created. This
difference could be explained by the type of company. The interviewee worked for an online marketing agency and the content in question was created to promote their own company, not the brands they work for.
The interviewees were unanimous about the effects of content marketing. Qualitative content did have positive effects for them in terms of increasing conversions and better positions in the search results, followed by more website traffic. This is in line with previous research that an editorial context results in a higher purchase intention, compared to
advertorials and ads (Attaran et al., 2015). One interviewee underlined the importance of the right topic of the content. Certain types of topics did work better for her company. Also it is important to adjust content to different target groups, because they have different needs and wants.
Developing experimental materials. Interestingly, one interviewee discussed the differences between traditional and content marketing, which is also the comparison that is made in the following experimental study. According to him, with traditional marketing you create something that hopefully results in direct sales. Content marketing, on the contrary, is more about informing, branding, telling a story or creating an experience, which could be rewarded in the end with a purchase.
For the stimuli in the upcoming experiment it is thus important to distinguish between selling and adding value by telling a story, for instance. The context could be the website of a brand, because the brand is already present in a natural way. The content marketing
materials could contain text supported by some visuals, which seems common in practice according to the explorative interviews.
Main study: Online content experiment
To test the hypotheses an online experiment was conducted. This study compared the persuasive effects of a content marketing website to a website with a more traditional marketing approach focused on selling.
An experiment is a method to test causal relationships between variables, by manipulating the expected cause-variable (e.g. Field & Hole, 2003; Bryman, 2012). In this
field experiment the manipulated independent variable type of marketing was isolated. Therefore, the influence of other (external) variables can be controlled and ruled out. By randomly assigning the participants to the levels of the independent variable and keeping the conditions as similar as possible, many external factors were hold constant. This control benefited the internal validity of the study, because other explanations for the effect were eliminated (e.g. Field & Hole, 2003; Bryman, 2012). The online setting of the field experiment enhanced the ecological validity of the study (Bryman, 2012), as well as the generalisability of the results to the real world (external validity).
Method
Design. The research design tested the effects of different marketing types on brand attitude and purchase intention mediated by advertising recognition, activation of
persuasion knowledge and consequently resistance towards persuasion. The experiment had a one-factor (Type of marketing: content marketing versus traditional marketing) between-subjects design.
Sample and procedure. The study was aimed at the Dutch market and therefore a population of adults who live or have lived in The Netherlands was chosen. The final sample (N = 77) consisted of volunteers, recruited within the personal and professional network of the researcher and via snowball sampling. Most of the participants drink the soft drink cola sometimes and the most frequent highest level of education is HBO and WO (Mage = 28.71, SDage = 10.97, Mocola = 2, Moedu = 6 , 67.5% female).1 They were randomly assigned to the conditions (nCM = 32, nTM = 45), exceeding the required minimum of 30 participants each and keeping the groups as similar as possible.
The participants received a link to an online questionnaire (Appendix C.) and could set the language to Dutch or English. After they were briefed about the study, they provided their informed consent. Then they filled out some demographic information. Next, they were shown one of the two versions of the manipulation, after which they responded to questions and statements measuring their attention for the stimulus materials, advertising recognition, activation of persuasion knowledge, resistance towards persuasion, brand attitude and
Multiple modes exist. The smallest value is shown.
purchase intention. The questionnaire ended with a short manipulation check and the option to leave their e-mail address behind to win a gift certificate (worth €25,-). The data was collected between December 16 and 30, 2018. Afterwards, the data was downloaded and statistically analysed using the IBM software SPSS Statistics version 24 and SPSS Amos version 25.
Materials. Participants in both conditions were exposed to a screenshot of a mock-up website of the fictitious soft drink brand Colita for at least 25 seconds, to ensure that they carefully viewed the image and read the text. Both the website and the brand were artificial,
Figure 3. Content marketing condition. Figure 4. Traditional marketing condition.
!
which contributed to the internal validity of the study. In the content marketing condition the focus of this website was on storytelling, to distinguish it from other marketing types. The traditional marketing condition consisted of a translation of the content marketing website in the form of a branded webshop with a banner ad, focussed on direct selling.
The header of the content marketing website showed the website’s title ‘The Colita Magazine’, followed by a short text about the website and the brand. After that, a button to watch a video was shown and at the bottom of the page participants could find a blog with previews of three blogposts (Figure 3.). The traditional website showed the same header image in the form of an ad for the brand. The brand name was mentioned and a price tag included. The wording in the text were changed from blog-related to selling oriented words and the button text was adjusted to ‘watch our commercial’. On the bottom of the page there was a webshop where they could order three different flavours of the soft drink (Figure 4.).
In order to control for possible cofounding variables, the materials in both conditions looked as similar as possible. The same pictures, text, colours, logo, and navigation of the website were used. Only the goal of the brand’s message, focus on attracting and retaining customers with storytelling versus focus on selling, varied.
Measures. Several demographic variables, such as gender, age, current place of residence, education level and whether the participants drink the soft drink cola, were measured to control for potential cofounding variables. Furthermore, the attention for the stimulus was checked with questions regarding the type of website and the contents of the text. A manipulation check was included to measure whether the materials had the intended effects. Participants indicated their perception of content’s approach (7-point scale, 1 = inform, 7 = persuade), the goal of the website (7-point scale, 1 = direct profit, 7 = attracting and retaining customers, reversed) and focus of the website (7-point scale, 1 = storytelling, 7 = selling). A lower score indicates that the website was perceived more as content
Advertising recognition. Whether the participants recognised the manipulation as
advertising or not was measured by responding with yes (1) or no (2) to the following statement: “The website contains advertising.” (Mo = 1).
Activation of persuasion knowledge. Five items were adapted from Boerman, Van
Reijmersdal and Neijens (2012) and Van Reijmersdal et al. (2016) and one item was added by the researcher. These items were used to measure the activation of attitudinal and cognitive persuasion knowledge. The participants indicated whether they thought the content on the website was commercial, informative (reversed), trustworthy (reversed), biased, convincing (reversed) and advertising (7-point Likert scales: 1 = strongly disagree, 7 = strongly agree).
The Principal Component Analysis (PCA) for activation of persuasion knowledge showed that the items form two components with an eigenvalue above one (eigenvalue1 = 2.07, eigenvalue2 = 1.52) and the scree plot showed a clear turning point after these components. The first factor explained 35% and the second factor 25% of the variance in the items. After a Varimax rotation the item ‘advertising’ had the strongest association (factor loading = .88) with the first factor and ‘convincing’ with the second factor (factor loading = .80). The reliability analysis for the first component showed a reasonable reliability of the scale, Cronbach’s alpha = .65. The reliability was improved to a more acceptable level, by removing the item ‘biased’ (Cronbach’s alpha = .72). This choice was made, because the word ‘bias’ has many synonyms and can therefore be open for various
interpretations. Due to probable mistranslations, the word could be interpreted differently in Dutch and in English. After removal, the scale seemed to measure a commerciality aspect of persuasion knowledge. A higher score indicates more activation of the participants’
persuasion knowledge (M = 5.81, SD = 1.24). The reliability analysis for the second component showed a reasonable reliability of the scale, Cronbach’s alpha = .63. The reliability could not be improved by deleting an item and the corrected correlations were large enough for all items to probably belong in the scale (r > .42). The scale was
score indicates more activation of the participants’ persuasion knowledge (M = 4.30, SD = 1.09).
Resistance towards persuasion. The cognitive resistance responses of the
participants were measured with four items. Sample items include, “It was easy to agree with the arguments made in the message.” (reversed) and “While reading/viewing, I
sometimes found myself thinking of ways I disagreed with the content on the website.” (e.g. Nabi, Moyer-Gusé, & Byrne, 2007; Moyer‐Gusé, & Nabi, 2010). The participants indicated their agreement on a 5-point Likert scale (1 = strongly disagree, 5 = strongly agree). The affective component of resistance towards persuasion was measured with three items aimed at negative affect. The participants indicated whether they felt angry, irritated or annoyed while reading/viewing the website (7-point scales: 1 = strongly agree, 7 = strongly disagree; e.g. Zuwerink Jacks & Cameron, 2003; Van Reijmersdal et al., 2016).
The PCA showed that the items form two factors with an eigenvalue above one (eigenvalue1 = 3.09, eigenvalue2 = 1.32) and the scree plots indicated a turning point after these components. The first factor explained 44% and the second factor 19% of the variance in the items. After a Varimax rotation, the item ‘annoyed’ had the strongest association for the first component (factor loading = .89). The reliability analysis showed a reliable scale, Cronbach’s alpha = .86. The scale was significantly improved by removing the item ‘angry’ (Cronbach’s alpha = .91). This makes logical sense, because both irritation and annoyance are feelings that reflect a slight bit of anger. Experiencing anger over a fictitious brand could be arguably too extreme of an emotion. The now excellent reliable scale seems to measure the affective resistance toward persuasion. A lower score indicates more
affective resistance (M = 4.96, SD = 1.70). ‘Doubt’ had the strongest association for the second component (factor loading = .80) after a Varimax rotation. The reliability of this scale was reasonable (Cronbach’s alpha = .67) and was improved to a more acceptable level by deleting ‘thinking certain way’ (Cronbach’s alpha = .72). This seems logical, because the content did not contain arguments that were strongly aimed at influencing the reader’s opinion. The scale therefore seems to measure the cognitive resistance towards persuasion. A higher score indicates more cognitive resistance (M = 3.17, SD = .80).
Brand attitude and purchase intention. Brand attitude was measured using 7-point
Likert scales adapted from Spears and Singh (2004). The participants indicated their feelings with the statement “The brand is unappealing – appealing/good – bad (reversed)/ unpleasant – pleasant/unfavourable – favourable”. To measure purchase intention the participants pointed out whether they thought they would buy the brand. The items were never – definitely, very little interested – very much interested, definitely intend – definitely not intend (reversed) and probably – probably not (reversed) (7-point scales; Spears & Singh, 2004).
The PCA for brand attitude showed that 4 items formed a single uni-dimensional scale: 1 component had an eigenvalue above one (eigenvalue = 2.46) and the scree plot showed a clear turning point after this component. The factor explained 61% of the variance in the items and ‘appeal’ had the strongest association (factor loading = .82). The reliability analysis for the component showed an acceptable reliability of the scale, Cronbach’s alpha = .79. For purchase intention the PCA also showed 1 component with an eigenvalue above one (eigenvalue = 3.05) explaining 76% of the variance in the items. ‘Interest’ had the strongest association (factor loading = .92). The scale was good reliable, Cronbach’s alpha = .89. It appears that the constructed scales measure attitude towards the brand (Ba) and purchase intention (Pi) respectively. Higher scores correspond with more positive attitudes and more intention to purchase the brand (MBa = 4.20, SDBa = 1.08; MPi = 3.31, SDPi = 1.47). Results
To prepare the dataset (N = 127) for the analyses, the variable names were adjusted, the correct measurement levels were selected and the data was checked for suspicious entries. There were no missing values. However, 29 cases were removed, because some participants did not complete the questionnaire. After creating a new variable type of marketing (1 = CM, 2 = TM), 21 more cases were removed because the response to the attention check-questions did not match the condition and/or the contents of the texts, resulting in the final sample (N = 77, nCM = 32, nTM = 45).
Manipulation check. To check whether the manipulation was perceived as intended, three independent samples t-test were conducted. The Levene’s tests showed significant
differences in the variance of all three dependent variables, so equal variances were not assumed (Fcontent approach = 12.23, p = .001; Fwebsite focus = 11.27, p = .001; Fwebsite goal = 10.02, p = .002). There seems to be a difference between the perception of the content approach in the content marketing condition (M = 5.13, SD = 1.91, n = 32) compared to the traditional marketing condition (M = 5.89, SD = 1.19, n = 45) that approaches significance, t(47.88) = -2.00, p = .051 (n.s.)., 95% CI [-1.53, .00]. Participants in both conditions perceived the manipulation as having a somewhat more persuasive approach, but this is almost significantly lower for the content marketing condition.
The results for website focus and website goal showed that the manipulation was perceived as somewhat focussed on selling and storytelling in the content marketing condition (M = 4.03, SD = 2.06, n = 32), compared to the perception as more focussed on selling in the traditional marketing condition (M = 5.58, SD = 1.47, n = 45), t(52.72) = -3.65, p = .001, 95% CI [-2.40, -.70]. Participants also perceived the goal to be more focussed on attracting and retaining customers (M = 2.28, SD = 1.28, n = 32), compared to a more direct profit oriented goal in the traditional marketing condition (M = 3.89, SD = 1.86, n = 45), t(74.93) = -4.50, p < .001, 95% CI [-2.32, -.90].
Normality test. The continuous outcome variables appeared to be distributed normally. All variables had a skewness between -1.00 and 1.00 (Table 1.), except for
commerciality persuasion knowledge (skewness = -1.61). The kurtosis values of all variables were between -3.00 and 3.00 (Table 1.).
Testing the hypothesised model. The main analyses were conducted with a path analysis and maximum likelihood (ML) estimations, using the Structural Equation Modelling (SEM) technique. SEM is a technique for testing hypotheses about relations between variables, by developing a model (Hu & Bentler, 1995). Two dummy variables were created for type of marketing (0 = not content marketing, 1 = content marketing) and advertising recognition (0 = not recognised, 1 = recognised). The zero values served as a baseline. The hypothesised model (Figure 5.) was then constructed in the IBM SPSS Amos software and calculated with bootstrapping (5000 samples).
The model did not seem to fit the data well. Nonetheless, the topic of model fit is an ongoing debate. Several researcher have established cut-off points, but it is important to note that those are only rules of thumb (Kline, 2011). A Chi-square test result indicates how much the sample and the covariances in the model differ (Hu & Bentler, 1995). The test result was significant, which is not desirable for a model that fits the data well (Bentler & Bonett, 1980), χ2(16) = 71.83, p < .001. Absolute fit indices with values > .90 are generally perceived as acceptable (e.g. Hu & Bentler, 1995; Bentler & Bonett, 1980). Others discuss a
Table 1. Skewness and kurtosis values.
Skewness Kurtosis
Credibility persuasion knowledge .21 -.53
Commerciality persuasion knowledge -1.61* 2.82
Cognitive resistance -.27 -.13
Affective resistance -.33 -1.38
Brand attitude -.31 -.33
Purchase intention .19 -.49
Note: Skewness and kurtosis values for all continuous outcome variables. * Non-normally distributed
Figure 5. Statistical representation of the hypothesised model.
Note: Hypothesised model, in which eclipses illustrate latent variables, rectangles illustrate observed variables and circles represent errors.
cut-off value of > .95 (Hu & Bentler, 1999). The results for the hypothesised model are all way below these thresholds, GFI = .79, AGFI = .53. Incremental fit indices compare the model to a baseline model (Kline, 2011). The values are judged with the same standards as the (Adjusted) Goodness of Fit (GFI, AGFI) and are insufficient, with NFI = .46, RFI = .05, IFI = .52, TLI = .06, CFI = .46. The Root Mean Square Error of Approximation (RMSEA) is scaled the other way around. The lower the value, the better the fit of the model to the data. Hu and Bentler (1999) mentioned a consensus of < .05. After researching many studies, they
concluded an alternative threshold of < .06. The RMSEA value for the hypothesised model is .21, which indicates a somewhat bad fit.
The modification indices suggested to draw pathways from credibility persuasion knowledge to the variables affective resistance, cognitive resistance and purchase intention. Paths between affective resistance and cognitive resistance and vice versa were also
suggested. Other modification indices did not make enough theoretical sense to include. The empirically developed model was modified by splitting the latent variable resistance up into two independent variabels and connecting those with each other and the outcome variables respectively. Furthermore, pathways from the credibility dimension of persuasion knowledge to all outcome variables were added.
Testing the modified model. The predictors in the modified model (Figure 6.) explained 34% (R2 = .34) of the variance in purchase intention. According to the above mentioned rules of thumb, the model was significantly improved and the modified model seems to fit the data well; χ2(8) = 9.17, p = .328 (n.s.), GFI = .97, AGFI = .87, NFI = .93, RFI = .76, IFI = .99, TLI = .96, CFI = .99, RMSEA = .04. Only the Relative Fit Index is below the suggested cut-off point.
Direct effects. Type of marketing has a moderate (b* = -.34) negative effect on
persuasion knowledge that approaches significance and becomes significant at the α < .10 level, b = -.44, p = .096 (n.s.), 95% CI[-1.00, .07]. This might indicate that the overall
persuasion knowledge score in the content marketing condition is .44-point lower and is activated less, compared to the traditional marketing condition. Type of marketing does not
have a significant effect on advertising recognition (b* = -.06, p = .638), brand attitude (b* = . 09, p = .380) and purchase intention (b* = .03, p = .726). H1, H4a and H4b are therefore rejected.
As expected, advertising recognition has a significant strong (b* = .92) positive effect on persuasion knowledge, b = 1.90, p = .012, 95% CI[.79, 3.33]. For every 1-point increase in advertising recognition, persuasion knowledge increases with 1.90. It seems likely that participants who recognised the advertising, have a more activated general persuasion knowledge compared to participants who did not recognise the advertising and therefore H2 is accepted.
The credibility dimension of persuasion knowledge has a moderate (b* = -.33) negative significant effect on affective resistance, b = -.52, p = .001, 95% CI[-.79, -.24]. There is also a moderate (b* = .39) positive significant effect on cognitive resistance, b = .29, p = .001, 95% CI[.15, .41]. This result partly supports H3. For brand attitude this variable has a medium sized (b* = -.42,) negative significant effect, b = -.42, p = .001, 95% CI[-.63, -.20] and credibility persuasion knowledge has a moderate (b* = -.29) negative significant effect on purchase intention as well, b = -.39, p = .022, 95% CI[-.70, -.06]. If credibility persuasion knowledge increases with 1-point, affective resistance decreases with .52, cognitive
resistance increases with .29, brand attitude decreases .42-point and purchase intention decreases with .39. It is plausible that the more activated the credibility dimension of persuasion knowledge, the more the participants resist, the less positive their attitudes towards the brand and the lower their intentions to purchase the brand become.
Cognitive resistance has a moderate (b* = -.24) negative significant effect on affective resistance, b = -.52, p < .001, 95% CI[-.84, -.29]. For every 1-point increase in cognitive resistance, affective resistance decreases with .52. More cognitive resistance consequently results in more affective resistance as well.
The final significant direct effect is that of brand attitude. It has a moderate (b* = .34) positive effect on purchase intention, b = .46, p = .016, 95% CI[.10, .77]. If brand attitude
increases 1 point, purchase intention increases .46-point. The more positive the attitude towards the brand, the more likely participants are to purchase the brand.
Indirect effects. The indirect small (b* credibilityPK = -.09, b*commercialityPK = -.20) negative
effect of type of marketing on both dimensions of persuasion knowledge approaches significance and becomes significant at the α < .10 level, with bcredibilityPK = -.20, p = .077 (n.s.), 95% CI[-.63, .02] and bcommercialityPK = -.51, p = .077 (n.s.), 95% CI[-1.12, .05]. This might indicate that the persuasion knowledge of participants in the content marketing condition is less activated, compared to the traditional marketing condition. This effect could be mediated by advertising recognition. Contrary to the hypotheses, there is no indirect mediation effect of type of marketing on brand attitude (b* = .07, p = .195) and purchase intention (b* = .07, p = .240). Thus, H5 was not supported by the results.
Advertising recognition has a significant moderate (b* = -.19) negative effect on brand attitude, b = -.68, p = .032, 95% CI[-1.12, -.04]. The variable also has a somewhat strong (b*
= .47) positive indirect effect on commerciality persuasion knowledge, b = 1.90, p = .012, 95% CI[.79, 3.33]. When advertising recognition is 1, brand attitude decreases with .68 point and commerciality persuasion knowledge increases 1.90 point. However, because this effect goes through a variable that is not observed, it is more likely that it concerns a specification of the direct effect discovered before.
Persuasion knowledge has a significant weak (b* = -.15) negative indirect effect on purchase intention, b = -.34, p = .046, 95% CI[-1.25, -.01], but this effect is better explained by the credibility dimension of this variable. Credibility persuasion knowledge has a
significant moderate (b* = -.21) negative indirect effect on purchase intention, b = -.28, p = . 002, 95% CI[-.54, -.10]. This dimension also has a weak (b* = -.10) negative indirect effect on affective resistance, b = -.16, p < .001, 95% CI[-.28, -.08]. For every 1-point increase in credibility persuasion knowledge, purchase intention will decrease with .28 and affective resistance decreases with .16. Activated credibility persuasion knowledge is thus likely to result in more affective resistance and less intention to purchase the brand. After
mediated by brand attitude and the effect on affective resistance seems mediated by cognitive resistance.
Conclusion and discussion
This study aims to demonstrate a persuasive effect of content marketing, compared to traditional marketing, on brand attitude and purchase intention and the extent to which resistance plays a role. As expected, the results show that content marketing is less likely to be recognised as advertising compared to traditional marketing. Consequently, the
participants experience less activation of both the commerciality and credibility aspects of persuasion knowledge. When credibility persuasion knowledge is activated, the participants are indeed more likely to resist. However, to answer the research question, this effect can not be attributed to exposure to the manipulation. The same goes for the effect on the
Figure 6. Statistical representation of the modified model.
Note: Modified model with standardised coefficients. Eclipse illustrates a latent variable, rectangles illustrate observed variables and circles represent errors.
* Significant at α < .05 ** Significant at α < .10
intention to purchase the brand, mediated by the attitude towards the brand. Thus, the mediation hypothesis is not supported by the results.
Nonetheless, the experiment did result in three other significant mediated pathways (Figure 7.). Participants in the content marketing condition, compared to the baseline traditional marketing, experience less activation of their persuasion knowledge mediated by less advertising recognition. Participants that recognise advertising have a less positive attitude towards the brand, mediated by a more activated credibility dimension of their persuasion knowledge. The more activated persuasion knowledge becomes, the more affective resistance the participants hold towards the content. This effect is mediated by an increase in cognitive resistance. Furthermore, more activated persuasion knowledge results in less intention to purchase the brand, mediated by a more negative brand attitude.
These pathways are still consistent with the literature. According to Wojdynski and Evans (2016) advertising recognition has a negative effect on attitudinal evaluations and intended behaviours. Also in line with the literature is that the activation of persuasion knowledge has an effect on responses towards advertising (Friestad & Wright, 1994). Thus, resistance, brand attitude and purchase intention as well. Interestingly, the current study distinguishes two aspects of this persuasion knowledge: activation of knowledge about the
Figure 7. The mediated pathways.
Note: The significant indirect effects, with mediators. Affective resistance was scaled in the opposite direction.
commerciality of the persuasion attempt and knowledge about the credibility of the content. Credibility is a more positive framed dimension, including thoughts about the
convincingness, trustworthiness and informational attributes of the content. The negative effect activated credibility persuasion knowledge has on resistance, confirms earlier findings by Fransen et al. (2015). They found that being aware of a persuasive attempt results in employing cognitive resistance strategies to counter that attempt.
Thus, content marketing does slightly positively affect brand attitude and purchase intention, but this effect is not significant and resistance towards persuasion does not play a mediating role. Moreover, affective resistance appears to be an outcome variable of
activated credibility persuasion knowledge and cognitive resistance. General discussion
The results of this study contribute to existing literature by indicating that the PKM applies to a content marketing context. Interestingly, irritations evoked by content and thinking of counterarguments do not change the way the brand is perceived. Affective resistance and purchase intention are the outcomes of two parallel pathways, both influenced by persuasion knowledge. Wei et al. (2008) found that persuasion knowledge negatively affects attitudes for low brand familiarity or inappropriate covert marketing tactics. It is possible that the negative brand attitude, and consequently lower purchase intention, in the current study are explained by non-familiarity with the exposed brand.
Moreover, the results suggest that the audience might accept a certain amount of commerciality in content. The interviews also indicate that commerciality is not an issue, as long as it is balanced. Their knowledge about credibility aspects of the content is more important and decisive in whether they will resist the content or form negative attitudes towards the sender’s brand. When content is perceived as credible (e.g. informative, trustworthy or convincing), persuasion knowledge is less activated. This relates to the preliminary study, in which interviewees indicated that content marketing is about informing the audience and becoming an expert to create trust.
An important implication for marketers is that a healthy balance between credibility and manipulative persuasive intentions is crucial (Cotte, Coulter, & Moore, 2005). It can be
