1. Introduction
The concentration of so-called hybrid forms of media content in the contemporary online environment is becoming increasingly intense (Balasubramanian, 1994; Macnamara, 2014; Taiminen et al., 2015). The lack of a clear theoretical and practical distinction between sponsored content (Tutaj & Reijmersdal, 2012), content marketing (Pulizzi, 2014; Pulizzi & Piper, 2023), brand journalism (Cole and Greer, 2013), and native advertising (Verčič & Tkalac Verčič, 2016) often results in confusion among readers, who are finding it increasingly difficult to recognize the commercial nature of the content being presented to them. For most researchers, native advertising serves as an umbrella term intended to simplify the classification of hybrid forms, as native advertising is, by definition, a paid advertisement designed to resemble editorial content. Its purpose is to attract a target audience with useful content while simultaneously promoting a brand, its values, and its products (Cole & Greer, 2013; Howe & Teufel, 2014; Wojdynski & Evans, 2016). On the other hand, research by Taiminen et al. (2015) revealed that most public relations and marketing professionals prefer to use the term content marketing when referring to commercial hybrid forms of online content.
In addition to the obvious terminological confusion, three other problematic areas of native advertising need to be addressed:
1. Application area: Is native advertising a public relations technique, a marketing tool, or a tactic indistinguishable from traditional advertising?
2. Authorship of native ads: Who is the creative author and who bears responsibility for the content?
3. Legal and ethical framework: How is native advertising legally defined, and what is the practice of media outlets in labelling such advertisements? (Trbušić et al., 2024)
According to the Interactive Advertising Bureau (IAB), there are three fundamental forms of native advertising: In-Feed/In-Content native advertising, Content Recommendation Ads, and Branded/native content. The latter refers to “paid content by a brand that is published in the same format as full-fledged editorial content on the publisher’s website, usually in collaboration with the publisher’s own teams” (IAB, 2019). This form of native advertising (branded/native content) will be the focus of the research component of this paper because it is the only type among the mentioned forms of native advertising that allows for a structured analytical approach.
The fact that native ads closely resemble editorial content in both appearance and format on online news portals implies that such ads incorporate essential elements of journalistic texts, such as headlines, visual elements (graphics and infographics, photo, video, branded visual content), and sources of information, which may include quoted statements or comments from common people or influential figures and representatives of organizations. Therefore, the main objective of this paper is to analyse which types of headlines, visual elements, and sources of information are most used in native ads. Furthermore, the study aims to determine, based on the Elaboration Likelihood Model (ELM), whether these elements belong to the central or peripheral route of persuasion. Finally, the paper seeks to identify differences in the effectiveness of ads based on these elements, enhancing understanding of the key factors that influence the success of native advertising.
In line with the research objectives, the following hypotheses are proposed:
H1: Native ads employing clickbait headlines will be more efficient than ads with neutral headlines.
H1a: There is a significant association between industry type and the type of headline used; specifically, industries with a pronounced consumer focus (e.g., healthcare, financial) employ emotional or uncertainty-based clickbait headlines more frequently than industries with less consumer focus.
H2: Native ads with rich visual elements will be most efficient compared to those with minimal visual components.
H3: Native ads featuring influencers as sources of information will achieve highest reach, thanks to the influencers’ reputation and appeal.
H3a: There is a significant association between industry type and the information source used in advertisements. Specifically, ads in industries that depend on personal preferences and trends (e.g., lifestyle, travel) will more frequently engage influencers as sources of information, whereas ads in professional and technical sectors (e.g., financial, ICT, healthcare) will more often rely on expert authorities or official institutions as sources.
In the following chapters, the paper elaborates on the theoretical framework, further develops the proposed hypotheses, describes the methodology and analytical design, and concludes with an interpretation of the results, conclusions, and limitations of the research.
2. Literature review and hypothesis development
This section is structured and divided into two parts. The first part provides an overview of the theoretical framework of the study, while the second part explains and contextualizes the variables to be analysed in the research segment in accordance with the theoretical foundation.
2.1. Elaboration likelihood model (ELM)
The Elaboration Likelihood Model (ELM) provides a comprehensive framework for organizing, categorizing, and interpreting the fundamental processes that underlie the effectiveness of persuasive communication (Petty & Cacioppo, 1986). As such, it is a model applicable to various areas of theory and practice, including psychotherapy and counselling (Petty et al., 1984a; Cacioppo et al., 1985) as well as mass media advertising, selling and marketing communications (Petty et al., 1984b; Cacioppo & Petty, 1985; Kemeter & Bušljeta Banks, 2013; El Hedhli & Zourrig, 2022).
At its core, the Elaboration Likelihood Model (ELM) explains how persuasive communication can influence attitude change. It posits two distinct routes through which individuals process persuasive messages: the central and peripheral routes to persuasion (Petty & Cacioppo, 1986). The central route suggests that individuals are sometimes motivated to pay attention to the content of a message and accept it when the arguments presented are sufficiently persuasive. In such situations, individuals engage in a thorough evaluation of the message and process its content in depth. Conversely, under different circumstances, individuals may lack the motivation to carefully assess information and instead focus on superficial and external cues. This approach is referred to as the peripheral route to persuasion, as it relies on elements that are secondary to the message itself (Petty et al., 2003; Aronson et al., 2013).
Given that the Elaboration Likelihood Model (ELM) explains how individuals process persuasive messages through the central and peripheral routes of persuasion, examining the language and communicative features of marketing communication becomes essential for understanding the mechanisms underlying its effectiveness. One of the key factors in marketing success is the application of an appropriate linguistic style, which serves as a powerful persuasive tool (Blankenship & Craig, 2011). Moreover, the way a message is formulated often plays a more significant role than its actual content, as it can influence the perception and interpretation of information (Brennan & Williams, 1995). In this regard, linguistic style not only facilitates the comprehension and processing of a message but also shapes the perception of the message sender (Areni & Sparks, 2005).
In the context of native advertising, where persuasive messages are integrated into journalistic-style content, the analysis of linguistic style becomes even more crucial. If we assume that linguistic style encompasses not just verbal expressions but also that structural and visual elements influence how the audience perceives the message, we can argue that headlines, visual design, and source attribution play a pivotal role in shaping the persuasive effect of an advertisement. These elements can function as peripheral cues, influencing how recipients process and interpret advertising messages, particularly in situations where deep cognitive analysis of content is absent. Therefore, considering the applicability of ELM (Elaboration Likelihood Model) to advertising and marketing communication, it is justified to investigate how various aspects of linguistic style, including headlines, visual design, and source attribution, influence the effectiveness of native advertisements.
Previous research on persuasive communication within the Elaboration Likelihood Model (ELM) framework highlights key factors influencing attitude change. Petty et al. (1981) found that personal relevance determines whether individuals process messages via the central route (focusing on argument strength) or the peripheral route (relying on cues like speaker prestige). Cacioppo et al. (1996) further demonstrated that individual differences in need for cognition (NFC) affect message processing, with high-NFC individuals engaging in deeper analysis. Petty and Brock (1981) showed that distractions and cognitive load lead individuals to rely more on peripheral cues. Chaiken (1980) and Petty et al. (1995) found that attitudes formed via the central route are more stable and resistant to change, whereas those shaped by the peripheral route are more susceptible to persuasion over time. When it comes to native advertising, research is related to the analysis of readers’ cognitive perception of native ads and their persuasive elements (Wojdynski & Evans, 2016), analysis of communication strategies in native ads (Wang & Li, 2017), perception of the organization or the sponsor of native ads and its socially responsible activities (Jung & Heo, 2018; Krouwer et al., 2020; Beckert et al., 2020; Wu & Overton, 2021), perception of the credibility of online portals that publish native ads (Howe & Teufel, 2014), and analysis of the basic features of native ads based on the most-read selected Croatian online portals (Trbušić et al., 2024).
2.2. Review of key variables
This section reviews three critical variables: clickbait headlines, visual elements, and information sources, that shape the effectiveness of native advertising, and introduces corresponding hypotheses suggesting that attention-grabbing headlines, strong visual design, and reputable or influential sources function as powerful peripheral cues to drive higher audience engagement.
2.2.1. Clickbaits
A headline is an integral component of a media text, yet its distinct positioning makes it the first, and often the only, element that the audience encounters (Kanižaj et al., 2022). Headlines are crafted to attract readers (Silić, 2006; Hudeček & Mihaljević, 2009), and since capturing their attention is challenging, publishers frequently resort to clickbait techniques. Clickbait occurs when a publisher posts a link with a headline that entices people to click to see more, without providing substantial information about what they will find (El-Arini & Tang, 2014). When readers realize that they have landed on irrelevant or misleading content, they often experience disappointment (Kanižaj et al., 2022). However, clickbait can sometimes be identified in advance, as it frequently follows stereotypical sensationalist constructions and linguistic patterns (Chen et al., 2015). Thus, clickbait headlines can be recognized through the following patterns, which will serve as a framework for the typology of native advertisement headlines in this research:
1. Uncertainty
2. Multiple combinations
3. Use of numbers
4. Emphasis on emotions
5. Use of vague pronouns
6. Reverse narrative style
Accordingly, the following hypothesis has been formulated:
H1: Native ads employing clickbait headlines will be more efficient than ads with neutral headlines, as such headlines more strongly arouse curiosity and act as a peripheral cue.
H1a: There is a significant association between industry type and the type of headline used; specifically, industries with a pronounced consumer focus (e.g., healthcare, financial) employ emotional or uncertainty-based clickbait headlines more frequently than industries with less consumer focus.
2.2.2. Visual elements
It is well-documented that modern digital environments increasingly prioritize visual content (Paska, 2019), a development that has consequently shaped the online media sphere. A substantial body of research has demonstrated the superiority of pictorial elements over verbal ones (Childers & Houston, 1984; Mitchell, 1986; Obermiller & Sawyer, 2011; Wedel & Pieters, 2000). The notion of visual equipment also encompasses video formats, particularly shorter videos (from just a few seconds to a couple of minutes), which rank among the most popular forms of video content in the information era (Wang, 2020; Xiao et al., 2023). Recently, an increasing number of businesses have allocated resources to promoting their products via short-form video advertising (Meng et al., 2024).
Therefore, we have formulated a second hypothesis that will be subject to empirical testing:
H2: Native ads with rich visual elements will be most efficient compared to those with minimal visual components, because appealing design and visual content provide powerful peripheral cues that quickly capture attention.
2.2.3. Information sources
According to their most fundamental definition, sources encompass any individuals or platforms from which, or on which, information about relevant facts can be obtained (Bobić, 1987). Such sources may be diverse, including other individuals, government agencies, other media organizations, and non-governmental organizations (Malović, 2005). In an advertising context, they can also include employees within an organization, product or service users, and others. In recent years, influencer marketing has evolved into a key component of digital marketing strategies, serving as a critical touchpoint for reaching target audiences (Kim & Kim, 2021). At its core, influencer marketing is a type of endorsement marketing that employs influencers’ product recommendations to boost sales; however, its objectives extend beyond mere revenue generation, aiming to capture the attention of potential customers, stimulate word-of-mouth effects, and foster customer engagement with a brand or product (Brown & Hayes, 2008).
Hence, the following hypotheses have been formulated:
H3: Native ads featuring influencers as sources of information will achieve the highest reach, thanks to the influencers’ reputation and appeal, which serve as a strong peripheral cue for the audience.
H3a: There is a significant association between industry type and the information source used in advertisements. Specifically, ads in industries that depend on personal preferences and trends (e.g., lifestyle, travel) will more frequently engage influencers as sources of information, whereas ads in professional and technical sectors (e.g., financial, ICT, healthcare) will more often rely on expert authorities or official institutions as sources.
3. Methodology and analytical design
In this section, we detail the selection and analysis of 543 native advertisements published over six months on six major Croatian news portals, focusing on how different headline types (standard vs. clickbait), visual elements (photos, videos, graphics), and information sources (e.g., influencers) were identified and measured. We then discuss how these variables were correlated with ad reach to assess the impact of peripheral cues across different industries.
3.1. Sample selection and data collection
The sample for the study consists of native advertisements published over six months (December 2021 – May 2022) on the six most-read Croatian internet news portals (index.hr, jutarnji.hr, vecernji.hr, 24sata.hr, telegram.hr, slobodnadalmacija.hr), which were identified as the most influential online news media in the Croatian digital space based on visitor criteria (Peruško, 2023). Following a machine-based pre-analysis that included all posts on Croatian internet portals during the observed period, totalling over 1.5 million posts on 5.104 web pages, a representative sample was selected based on visit frequency, specifically from the largest Croatian internet portals. A machine search was conducted using keywords and phrases in the content or headlines. The keywords used to identify native content included phrases such as paid content, sponsored content, native, content provided by, content created by, powered by and tnative, as native content should be clearly identified as such in the media. After selecting all articles published during the six-month period that included these keywords, a total of 811 articles were identified, from which non-native articles, primarily sponsored or PR content, and other irrelevant material were manually excluded. The final sample for the study consisted of 543 native advertisements across the six internet portals, noted for their influence and readership in Croatia. Before determining the research sample, it was confirmed that all analysed internet portals offer native advertising services through a review of their publicly available marketing, advertising, and public relations pricing lists.
3.2. Analytical framework and variables
Of the total sample of 543 native advertisements, 177 posts, or 32.6%, were from the portal jutarnji.hr, 103 posts or 18.97% from vecernji.hr, and 99 posts or 18.23% from telegram.hr. The portal 24sata.hr accounted for 88 analysed articles or 16.21%, index.hr for 62 or 11.42%, and slobodnadalmacija.hr for 14 native articles or 2.58%. The research on a sample of 543 native advertisements published between December 1, 2021, and May 30, 2022, on the six selected Croatian internet portals, was conducted using content analysis methodology. An analytical matrix was applied, containing five general categories (general identification elements, content emphasis, headline characteristics, content characteristics, audience engagement), within which 19 variables were observed for each of the 543 ads. The researchers independently read and analysed all identified ads and then compared results, achieving full agreement across all variables.
For the purposes of this study, three main categories were examined: headline characteristics, visual content characteristics, and types of information sources. In analysing headlines, the focus was on classifying different headline types, distinguishing between standard and clickbait headlines. When examining visual content, particular attention was given to identifying specific elements, such as photographs, video content, graphics, or infographics, as well as any combinations thereof. Additionally, different information sources were investigated, including representatives from the organization/company, brand ambassadors, influencers/celebrities, direct users of the product or service, and multiple sources. All these variables were then correlated with the reach of each ad to illustrate how various peripheral cues affect the effectiveness of native advertisements. Finally, an industry-specific analysis was conducted to determine whether certain headline types or types of information sources were more prevalent or characteristic of industries, and how these variables related to the ads’ reach.
4. Results and discussion
In this chapter, we analyse how various peripheral cues, such as clickbait headlines, visual elements, and information sources, shape the reach and effectiveness of native advertisements. We also examine industry-specific patterns to reveal how different sectors strategically employ these cues to engage audiences and improve ad performance.
4.1. The effectiveness of clickbait headlines as peripheral cues
The data indicate that as many as 79.01% of the observed native advertisements employed clickbait headlines, with the most prevalent techniques being those based on uncertainty (44.38%) and combinations of multiple clickbait elements (13.81%). Meanwhile, neutral headlines accounted for 20.99%, and other identified styles (e.g., emotional appeals, reverse narrative) were comparatively infrequent and less effective. In terms of reach, the data suggest that headlines containing clickbait features, particularly those emphasizing uncertainty or combining multiple styles, attract greater audience attention and encourage interaction.
Table 1: Prevalence and reach of clickbait headline characteristics in the analysed native ads (n=543)
Source: Author’s calculations
These findings provide empirical evidence supporting H1: Native ads employing clickbait headlines will achieve greater reach than ads with neutral headlines, as such headlines more strongly arouse curiosity and act as a peripheral cue, indicating that ads featuring clickbait headlines achieve higher reach than their neutral counterparts by eliciting stronger curiosity and functioning as a peripheral cue. For illustrative purposes, we provide a select few of the analysed advertisement headlines: Do you eat fish at least twice a week? Here’s why you should (Index.hr, 2022a); Is it possible to win at an online casino, or is everything rigged? (Index.hr, 2021); How can we pull children away from their mobile phones and guide them toward more beneficial activities? A teacher offers an excellent suggestion (Marinović, 2021); What is the secret of fortunate individuals? Experts say they do these four things differently (Index.hr, 2022b).
Table 2: Wilcoxon rank sum (pairwise) tests comparing audience reach for clickbait vs. non-clickbait headlines
Source: Author’s calculations
Since the Shapiro-Wilk test for normality of audience reach data (W = 0.9763, p < 0.001) indicated a significant deviation from normal distribution, the non-parametric Kruskal-Wallis test was used to assess differences in audience reach. The significant Kruskal-Wallis result (χ² = 28.45, p < 0.001) suggested that at least one of the groups differed from the others. To investigate whether native ads using clickbait headlines achieve greater reach than non-clickbait ads, Wilcoxon rank-sum (Mann-Whitney) tests were conducted with the null hypothesis (H₀) that Not a clickbait headline does not surpass each of the six clickbait headline categories. After applying a Bonferroni correction to adjust for multiple comparisons, the results indicated that Not a clickbait headline reached significantly larger audiences than Use of vague pronouns (adjusted p = 0.0114), Uncertainty (adjusted p = 0.0110), Use of numbers (adjusted p = 0.0190), and Multiple combinations (adjusted p = 0.0041). In contrast, there was no significant difference in reach for Reverse narrative style (adjusted p = 1.0000) or Emphasis on emotions (adjusted p = 1.0000). Thus, while non-clickbait headlines appear to outperform several clickbait styles, they perform comparably to reverse narrative and emotion-focused headlines under these tests.
The Figure 1 illustrates the prevalence of various clickbait headline techniques across a range of industries. The chart visually represents how different sectors employ specific strategies to attract audience engagement, with colour intensity indicating the percentage of ads that utilize each technique within each industry. A chi-square test of independence was performed, revealing a significant association between industries and clickbait characteristics (χ² = 264.45, p < 0.001), which suggests that different industries favour specific information sources in their ads and employ distinct clickbait strategies. The respective data therefore aligns with H1a: There is a significant association between industry type and the type of headline used; specifically, industries with a pronounced consumer focus (e.g., healthcare, financial) employ emotional or uncertainty-based clickbait headlines more frequently than industries with less consumer focus, suggesting that industries with a pronounced consumer focus are indeed more inclined to employ emotionally charged or uncertainty-based clickbait techniques. A particularly telling example comes from the healthcare industry, where native advertisements prominently feature two central themes: on the one hand, narratives centred on illnesses, disorders, and diagnoses, and on the other, successful solutions and preventive measures aimed at mitigating health risks (Trbušić et al., 2024). By emphasizing fear and urgency, such ads tap into consumers’ emotional responses, fostering a heightened need for products and services that promise relief or protection. Interestingly, the real estate sector in Croatia also capitalizes on uncertainty, which is particularly evident given current market volatility and rising housing costs.
Figure 1: Distribution of different clickbait types across various industries
Source: Author’s calculations
According to recent data from the Croatian Bureau of Statistics (2025), real estate prices have experienced a steady increase over the past few years, intensifying concerns about affordability and accessibility among prospective buyers. This heightened sense of urgency provides a fertile ground for uncertainty-based advertising strategies. Below, we present only a few of the clickbait headlines from the analysed advertisements that convey a sense of uncertainty: Nearly 500,000 people in Croatia suffer from the same digestive problems. Even those under 35 should be cautious (Telegram.hr, 2022a); This vitamin helps combat winter depression – here’s how to recognize and treat it (Index, 2022c); A well-known Zagreb restaurateur reveals his biggest financial mistakes – I keep making them over and over again (Marinović, 2022); Applications for the APN program are opening soon; here’s everything you need to know if you’re buying an apartment with a subsidized housing loan (Telegram, 2022b); Will the current surge in inflation impact car insurance prices as well? An expert reveals what we can expect (Telegram, 2022c); While renovating his house, Željko from Zagreb encountered a problem many people face. I found the perfect solution. (Jutarnji list, 2022).
In contrast, more technically oriented or B2B-driven industries, such as shipbuilding, tend to favor straightforward, non-clickbait approaches, reflecting the highly specialized nature of their content and more narrowly defined consumer base.
4.2. The effectiveness of visual elements as peripheral cues
The research findings indicate that all analyzed native advertisements include some form of visual element (photograph, video content, infographic, or a combination of these). Photographs are the most frequently used, appearing in 400 ads, while the combinations of photographs and graphics, as well as photographs and video content, appear 65 and 60 times, respectively. However, further analysis did not reveal any statistically significant differences in reach across the different types of visual elements. Consequently, we are unable to either confirm or refute H2 that Native ads with rich visual elements will be most efficient compared to those with minimal visual components, because appealing design and visual content provide powerful peripheral cues that quickly capture attention.
4.3. Information sources as peripheral cues
Table 3: Representation of different types of sources in the statements and reach (n=543)
Source: Author’s calculations
A comparison of the analytical results with the proposed hypotheses H3: Native ads featuring influencers as sources of information will achieve highest reach, thanks to the influencers’ reputation and appeal, which serve as a strong peripheral cue for the audience, and H3a: Ads in industries that depend on personal preferences and trends (e.g., tourism, travel) will more often engage influencers as sources of information, whereas ads in professional and technical sectors (e.g., financial industry, ICT, healthcare industry) will prefer expert authorities or official institutions as sources, due to a stronger emphasis on credibility and technical accuracy, reveals a clear alignment. Specifically, H3 posits that native advertisements featuring influencers as sources of information will achieve the highest reach, primarily due to the influencers’ recognizability and appeal to audiences. This is corroborated by the finding of the highest mean reach (14,645.78) precisely in the category of ads featuring influencers/celebrities, a figure considerably greater than that of other source types. These results also align with previous research on the positive effects of involving influencers in marketing communication (Djafarova & Rushworth, 2017; Evans et al., 2017; Boerman, 2020). A significant difference in audience reach was also confirmed concerning the type of sources used in the ads (χ2=33.34, p<0.001). This result suggests that different types of sources, such as direct user testimonials or organizational representatives, have a measurable impact on the success of the article.
To examine whether using an influencer/celebrity category differs from other source types, Wilcoxon rank-sum (Mann-Whitney) tests were conducted with the null hypothesis (H₀) that the influencer/celebrity category does not surpass each of the six other categories. Following a Bonferroni correction for multiple test comparisons, the results revealed that influencer/celebrity sources reached significantly different outcomes compared to a direct user of the product/service (adjusted p = 0.000247) and an unrelated source (adjusted p = 0.002883). In contrast, there was no significant difference between influencer/celebrity and any of the remaining source types, namely, no source/statement, source from the organization/company, brand ambassador, or multiple sources, as their adjusted p-values all exceeded 0.05. Thus, while influencers/celebrities stand apart from direct users and unrelated sources, they perform comparably to the other categories examined in this context.
Table 4: Wilcoxon rank sum (pairwise) tests comparing audience reach for influencer/celebrity vs. other types of sources
Source: Author’s calculations
Figure 2: Distribution of different source types across various industries
Source: Author’s calculations
Additionally, a Pearson’s chi-squared test of independence was conducted to examine the association between industries and information sources (H3a), which yielded a significant result (χ² = 254.72, p = 0.000). These findings indicate that different industries favour specific information sources in their ads and employ distinct clickbait strategies. Furthermore, H3a assumes that industries oriented toward personal preferences and trends (e.g., tourism, fashion, travel) are more likely to employ influencers as sources, while professional and technical sectors (e.g., finance, ICT, healthcare) will rely more heavily on expert authorities or official institutions, owing to an emphasis on credibility and technical accuracy. The data substantiates this claim: the sports sector, as well as industries aiming to capture broad consumer attention (e.g., media, marketing agencies, etc.), prominently utilize influencers/celebrities. In contrast, advertisements in industries such as shipbuilding, real estate, and science and education more frequently feature multiple sources and/or in-house experts and official institutions.
To conclude, the results of this study reveal both consistencies and divergences from previous findings in the literature. For instance, Rony et al. (2017) report that clickbait posts generally reach broader audiences, yet our analysis shows that neutral headlines, in certain cases, achieved greater reach than clickbait ones. Conversely, the finding that native advertisements featuring influencers as information sources attain the highest reach aligns with expectations and prior research. A recent meta-analysis by Pan et al. (2025) further demonstrates that influencer marketing significantly affects user behavior and engagement, which is clearly reflected in our results.
Some results, such as the strong effect of influencers, were expected based on theory and earlier studies, while others, such as the limited effectiveness of clickbait elements, were unexpected and open new questions for further inquiry. According to the ‘knowledge gap’ hypothesis, one might assume that curiosity-driven, attention-grabbing headlines increase reach, but our findings suggest that this is not always the case. Such discrepancies may be attributed to factors such as audience saturation over time or decreasing trust in overly sensational headlines, consistent with research showing that clickbait can erode reader credibility and trust.
5. Conclusion
The findings of this study confirm the crucial role of peripheral cues in attracting initial audience attention in native advertising, in line with the Elaboration Likelihood Model (ELM). While reach alone does not provide direct insights into the depth of message processing, higher reach can be seen as an indication that a certain percentage of users was at least minimally motivated to click and read further. According to ELM, this suggests that native ads leveraging strong peripheral cues, such as influencers, clickbait headlines, and rich visual elements, effectively generate initial audience engagement.
This study initially found that clickbait headlines, particularly those based on uncertainty, vague pronouns, and the use of multiple clickbait techniques, can attract attention by triggering curiosity and addressing emotional or informational gaps. However, the statistical tests revealed a more complex picture. Specifically, the non-parametric analyses showed that non-clickbait headlines reached significantly larger audiences than clickbait variants using vague pronouns, uncertainty cues, numeric indicators, and multiple combinations. Notably, there was no significant difference between non-clickbait headlines and those employing reverse narrative styles or an emphasis on emotions. This suggests that although certain clickbait strategies may effectively capture immediate attention, their overall impact is not uniformly superior, and some traditional headline styles can perform just as well or even better in terms of reach. Finally, while clickbait headlines can be effective in capturing attention, their long-term impact and potential negative effects on brand perception remain open questions, as they may lead to user disappointment if the content does not meet expectations.
When examining the role of visual elements as a variable in native advertising, it emerged that every analyzed advertisement included some form of visual component. Consequently, we were unable to compare them with ads lacking any visuals, meaning we cannot definitively conclude whether visual elements, regardless of their sophistication, do or do not generate higher reach. Future studies incorporating advertisements without visual components would be necessary to isolate the precise impact of visual cues on ad performance.
Furthermore, the results confirm that native ads featuring influencers as sources of information achieve the highest reach, reinforcing the idea that an influencer’s reputation and recognizability serve as strong peripheral cues. However, reach alone does not reveal whether users engaged in deeper processing of the message. Assessing whether users read and processed the content would require additional behavioural metrics, such as time spent on the page, scroll depth, and conversion rates. This distinction is essential, as ELM differentiates initial attention capture via the peripheral route from deeper message processing via the central route. Thus, while influencer-driven ads attract attention primarily due to the influencer’s identity and reputation, further analysis would be needed to determine whether some users transition to central processing, which is more likely if the influencer is perceived as a credible expert in the respective field.
Industry-specific trends support the hypothesis that different sectors adopt distinct advertising strategies. Industries driven by personal preferences and trends (e.g., tourism, fashion, media) are more likely to use influencers to attract attention and foster audience engagement. In contrast, professional and technical sectors (e.g., finance, ICT, healthcare) tend to rely on expert authorities and official institutions, as credibility and technical precision play a crucial role in audience persuasion. This finding aligns with the Two-Step Flow of Communication Theory (Katz & Lazarsfeld, 1955), which suggests that influencers may serve as initial motivators for engagement, while a subset of users, especially those with a personal interest in the topic, may transition to deeper message processing if the influencer is perceived as an expert.
Several limitations of this study should be noted. First, the sample consists of 543 native advertisements published over a six-month period on several leading Croatian online portals, which restricts the generalizability of the findings to other contexts or longer time frames. Second, since all analysed ads included visual elements, we lacked a control group of ads without visuals, making it impossible to isolate the effect of visual components on reach. Lastly, categories such as ‘clickbait headline’ or ‘influencer’ were defined according to our coding criteria and may overlook nuances, such as varying levels of influencer popularity.
Finally, this study provides strong empirical support for the effectiveness of peripheral cues in increasing the reach of native advertisements. However, future research should examine their impact on deeper message processing and attitude change. Further studies could incorporate behavioural analytics (e.g., A/B testing, engagement time tracking) to explore how different audience segments interact with native ads beyond initial attention capture. From a practical perspective, advertisers should recognize the effectiveness of combining engaging headlines, compelling visual elements, and influential sources of information to maximize message visibility. However, to enhance message retention and persuasion, they should also consider strategies that encourage deeper cognitive processing, particularly when targeting audiences in sectors where credibility and expertise are crucial factors in decision-making.
References
Index.hr. (2021, December 29). Je li moguće dobiti na online casinu ili je sve namješteno. https://www.index.hr/magazin/clanak/je-li-moguce-dobiti-na-online-casinu-ili-je-sve-namjesteno/2328613.aspx
Index.hr. (2022a, May 17). Jedete li ribu barem dvaput tjedno? Evo zašto biste trebali. https://www.index.hr/magazin/clanak/jedete-li-ribu-barem-dvaput-tjedno-evo-zasto-biste-trebali/2355862.aspx
Index.hr. (2022b, April 22). Koja je tajna ljudi koji imaju sreće? Stručnjaci kažu da ove 4 stvari rade drugačije.https://www.index.hr/magazin/clanak/koja-je-tajna-ljudi-koji-imaju-srece-strucnjaci-kazu-da-ove-4-stvari-rade-drugacije/2357661.aspx
Index.hr. (2022c, February 18). Ovaj vitamin pomaže kod zimske depresije – evo kako je prepoznati i liječiti.https://www.index.hr/magazin/clanak/ovaj-vitamin-pomaze-kod-zimske-depresije-evo-kako-je-prepoznati-i-lijeciti/2339004.aspx
Interacting Advertising Bureau. (2019). Native advertising playbook 2.0.https://www.iab.com/wp-content/uploads/2019/05/IAB-Native-Advertising-Playbook-2_0_Final.pdf
Jung, A., & Heo, J. (2018). Ad disclosure vs. ad recognition: How persuasion knowledge influences native advertising evaluation. Journal of Interactive Advertising, 19(1), 1–14.https://doi.org/10.1080/15252019.2018.1520661
Petty, R. E., Brock, T. C. (1981). Thought disruption and persuasion: Assessing the validity of attitude change experiments. In Petty, R. E., Ostrom, T. M., Brock, T. C. (Eds.), Cognitive responses in persuasion (pp. 55–79). Hillsdale, NJ: Erlbaum.
Petty, R. E., & Cacioppo, J. T. (1984b). Motivational factors in consumer response to advertisements. In R. Geen, W. Beatty, & R. Arkin (Eds.), Human motivation: Physiological, behavioral, and social approaches (pp. 418–454). Allyn & Bacon.https://search.catalog.loc.gov/instances/a75285f3-13b8-54c3-9648-705433e7f317?option=lccn&query=83025802
Xiao, L., Li, X., & Zhang, Y. (2023). Exploring the factors influencing consumer engagement behavior regarding short-form video advertising: A big data perspective. Journal of Retailing and Consumer Services, 70, 103170.https://doi.org/10.1016/j.jretconser.2022.103170
Clickbait naslovi i influenceri: Kako periferni poticaji utječu na doseg nativnih oglasa