Introduction

In an increasingly fragmented digital media landscape, news and broadcasting organizations are pressured to maintain visibility across multiple online platforms (Meyer, 2025). Audiences now consume content not just via newspapers, TV or news websites, but across a range of social media channels (Pew Research Center, 2024). Each social media has its own technical requirements, varying formats, and different algorithmic structure to provide personalization (Presuel & Sierra, 2019). This diversification has increased the workload for journalists and marketing teams, who must continually adapt and distribute content to ensure that their audiences grow (Kumar, 2023).

Media Gridz is a software tool designed to address this challenge. It automates the process of transforming news articles into platform-specific social media posts and publishes them automatically, allowing newsrooms to distribute content more efficiently, consistently, and effectively.

While Media Gridz is designed to increase web traffic by transforming articles into visually engaging social media posts, its effectiveness compared to traditional link-based posts remains an assumption that requires empirical validation. To date, there is a lack of academic research using statistically significant data to evaluate if link-based posts indeed perform worse compared to image-based posts (Lauron, 2025).

This issue is particularly relevant for media organizations operating in today’s digital landscape, where media outlets depend heavily on web traffic for revenue through advertising and subscriptions (Mitchell & Holcomb, 2014). In such an environment, optimizing post format to increase traffic is a strategic necessity. Therefore, this paper aims to provide knowledge in Social Media Optimization for newsrooms on the type of format of post they should be utilizing to achieve higher organic web traffic from social media, and respectfully higher advertising or subscription revenue (Anderson, 2024).

According to data from Pew Research Center, Facebook remains the leading social media platform for news consumption in the U.S., with approximately one-third of U.S. adults reporting that they regularly get news on the site (Pew Research Center, 2024). Moreover, medias still consider Facebook as primary platform to promote their content (Hille and Bakker, 2013). This positions Facebook as a critical channel for news distribution, making it highly relevant for evaluating content format aimed at increasing website traffic. Lastly, Facebook's support for various post formats, including traditional link-based posts and image-based posts with links in the comments, aligns well with the functionalities offered by Media Gridz (Meta Business Help Centre, 2025).

From an algorithmic perspective, driving website traffic from Facebook requires a post to reach an audience, catch their attention and encourage exploratory behavior, specifically, clicks to read more (Pócs et al., 2021).

Extensive research has been conducted on the factors that drive attention and exploratory behavior on social media. Most academic studies to date have focused on content-related variables, such as message framing, audience alignment, shareability, virality, and headline construction, including the use of clickbait (Paulussen et al., 2016; Qiu & Golman, 2024; Scacco & Muddiman, 2020). However, significantly less attention has been given to the impact of post format and its impact on visibility and click behavior amongst users. This represents a critical gap in the literature. As Media Gridz is not concerned with altering content itself, but rather with how content performs in terms traffic generation, this underexplored aspect of format becomes central to its value proposition.

Previous non-academic research shows variations in content performance. In 2023, Duree Company found that video content achieved 30% more reach and 135% more engagement than images (Jenna, 2023). In contrast, a 2024 Hootsuite study comparing engagement found that multiple-image posts outperformed videos, links, and single-image posts (Lauron, 2025). However, both researchers were not executed following specific methodology and standardization, so the reliability and validity of their experiments is very low (Carroll, 2023). 

Given these mixed findings and the absence of academic research on organic format performance on Facebook, this study aims to fill the gap by examining whether link-based posts or image-based posts with links perform better in terms of organic web traffic generation from social media regardless of the content or headline. The findings will help Media Gridz statistically validate its potential for generating higher organic social traffic while also contributing to the academic literature on organic social media content performance and Facebook algorithm structure.

Literature Review and Theoretical Framework

This literature review examines the determinants of social organic traffic from Facebook to news articles. The first section synthesizes existing research on the algorithmic mechanisms that govern content distribution on Facebook, identifying the key factors that influence a post’s visibility in users’ feeds. The second section reviews studies of user behavior on social media, with particular emphasis on the motivations driving news consumers on Facebook to attend to, engage with, and demonstrate exploratory behaviors toward journalistic content.

Functionality and relations of variable within Facebook algorithm

Several interrelated factors determine if organic web traffic will be generated via a post on Facebook. To measure traffic, which is key for media’s revenue, Google Analytics provides an option to do that via the variable “sessions”. Sessions are marked in Google Analytics “when a user either opens your app in the foreground or views a page or screen and no session is currently active, for example, their previous session has timed out.” (Analytics Help, n.d.)

Organic web traffic generation via Facebook is defined in Vhatkar’s research as “website links shared on business page. Users see these links & when they click on links they are redirected towards website.” (Vhatkar, 2016). A business page means a profile on Facebook which represents a company (Meta Business Help Centre, 2025). Therefore, in the context of this research, links published through business page are essential components of a post to generate traffic.

Links can be placed in variety of posts including “Photos; Videos; Live broadcasts; Carousels; Stories” as well as purely link-based posts. Moreover, links can be used within the description of the post, as well as added as a comment via the business page (Meta Business Help Centre, 2025). For a post to be seen, it must first reach the user (Pócs et al., 2021). In the context of Facebook, reach refers to the number of unique users who have seen a post (Tan & Lim, 2020). There are two types of reach: organic and paid. This study focuses solely on organic reach defined as “the number of people who saw an unpaid post from a business page on their screen.”, as the goal is to help medias understand how to improve organic web traffic from social media page (Meta Business Help Centre, 2025). In practice, Facebook doesn't track whether a user actively engaged with the content but simply counts that the post was displayed on a user’s screen (Zote, 2024). Based on resources from Meta (2025), total page likes and follows of a business page play a significant role in determining the reach of posts. Total page likes refers to the number of users who have clicked the "Like" button on a Facebook Page to show their support and receive updates, and Facebook, notes that when someone likes your Page, they automatically follow it, ensuring that they may see your Page’s updates and posts in their feed. Similarly, total page follows, as defined by Facebook, indicates that users who follow a Page may also see its updates and posts in their feed (Meta Business Help Centre, 2025). Additionally, engagement plays a key role in determining how the algorithm boosts reach (Lipsman et al., 2012). Engagement includes user interactions such as reactions, shares, comments and clicks directly on the post (Mishnick & Wise, 2024). All these elements are interwind and contribute to the success or failure of a post to generate organic web traffic. Therefore, the next section, examines how the algorithm perceives each of these elements in determining the post’s reach.

In a study from 2018 using a data mining approach it is identified that the most critical factor influencing organic reach is audience interaction, including total page likes of business profile, followed by the type of content (video, text, links, images, album images), as well as the timing of publication, including the month, day, and hour (Huang et al., 2018).

According to the research of Pócs et al. “likes” on post do not always translate to higher organic reach. There is a statistically significant negative correlation between total organic reach and the “like”, which challenges the assumption that all interactions increase visibility. On the other hand, “comments,” the “love” and the “haha” engagements increase total organic reach. In contrast, a high frequency of “wow” or “sad” reactions tends to correlate with negative Facebook reach. The impact of “shares” on organic reach is complex, while they positively influence non-fan reach, they have a negative correlation with fan reach (Pócs et al., 2021).

Data from the research of Qiu and Golman from 2024, showed that, though, engagement interactions sustain long term reader’s engagement, but it’s impact on short term curiosity (clicking) is negatively correlated, meaning less engagement might lead to more clicks (Qiu & Golman, 2024).

Another important factor are links. Nonacademic studies from 2024 and 2023 found that posts without a links have almost 3 times more engagement than link-based posts. This suggests that Facebook may deprioritize posts with links in their algorithms, potentially reducing their visibility (Cashyap, 2024; Froome, 2022). Nonetheless, this hypothesis is not statistically confirmed. A study from 2021 suggested that placing the link elsewhere in the post, such as in the comments, could exploit a loophole in Facebook’s algorithm and help maintain reach (Chawla & Chodak, 2021).

Moreover, user engagement through comments is another crucial determinant of organic reach. The number of comments on a post creates a feedback loop that further increases reach (Shahbaznezhad et al., 2021).

Last, but not least, it’s important to mention that Facebook also uses algorithmic personalization to prioritize content in users' feeds (Kaluža, 2022). Academics argue that this leads to filter bubbles, whereas the theory highlights how algorithmic curation personalizes content based on previous interactions, potentially limiting exposure to diverse viewpoints (Holone, 2016). Algorithmic personalization of Facebook also leads to echo chamber effect, which results in users who are exposed to content that aligns with their existing interests and beliefs, reinforcing their perspectives over time and increasing confirmation bias (Cinelli et al., 2021). The echo chamber effect leads to limited access to news that might propose alternative opinions (Holone, 2016).

In the context of Media Gridz, these theories suggest that once a user engages with certain content, the platform is more likely to show them similar content in the future. It is important to mention that the algorithmic personalization does not focus on personalizing the format of content, but its essence, which means the content inside of it. For news medias, that means that when you are exposed to a certain perspective, you interact with it, you will be provided with similar perspectives in the future (Kaluža, 2022). However, this is problematic for the organic traffic of medias that strive to present multiple points of views (Chueca Del Cerro, 2024; Holone, 2016).

In conclusion, organic reach is driven by multiple interrelated factors, including channel size, audience interactions, content type, timing, and platform algorithms.

News consumption in social media. Factors that drive attention, engagement and exploratory behavior

Although the algorithm is driven by specific factors, as discussed above the main force behind it are its users and their interactions with the content. Therefore, this chapter presents literature that explains how news consumers consume news content on Facebook, what triggers their attention, and what motivates them to produce exploratory behavior in the form of click.

Attention capture on social media

Past research of news consumers identifies new trends in how news content on social media captures attention of users. In the context of this research attention capture happens when the brain focuses more on one stimulus over others, giving it priority and directing more cognitive energy towards it (Öhman et al., 2001). Firstly, given the vast amount of content competing for attention on social media, users naturally filter out what seems unimportant (Pentina et al., 2015). Broadbent’s Filter Model explains that individuals prioritize sensory input based on physical properties like color, motion, and contrast, meaning that visually striking content is more likely to stand out (McLeod, 2023). Furthermore, in 2016 research of font sizes in relation to eye fixation, it is identified that the larger the font size, the higher average time of fixation on content. This means that more prominent fonts on images of news are expected to capture more attention of a user (Rello et al., 2016). Moreover, in eye-tracking research of visuals on social media Vraga et al. identified that attention to social news and political publications is often a result of visual design (Vraga et al., 2019). Sunarso et al. also discovered that visual content plays a fundamental role in enhancing engagement and strengthening brand recognition, supporting storytelling, and improving web traffic (Sunarso et al., 2023).

Despite previous research that indicates visual content as essential for capturing attention, some research suggests this is not always the case. For example, an experiment on Twitter found that tweets without links or images sometimes performed better in terms of engagement, challenging the assumption that every post must include a captivating visual to be effective (Sehl, 2021). Furthermore, in 2024 study in how attention in social media newsfeed works, it was discovered that attention capture of visual content decreases when users are in public setting as well as when they scroll on mobile, and instead they focus more on textual elements (Mayer et al., 2024).

Although, most research shows that visuals are expected to produce more attention capture on feeds, there are some cases where visuals, might not always capture the attention. Meanwhile, Media Gridz promise is that a visual including a headline with bigger font, will lead to higher attention capture. Therefore, it is crucial to confirm this hypothesis.

Attention retention on social media

Berger et al. discovered that capturing attention is not enough to produce engagement. A piece of content also needs to retain that attention and only after that a user can decide to like, comment, share or click to learn more. Therefore, this section will present what factors drive attention retention or as defined by Berger et al. “hold the attention” of the user on a certain piece of content (Berger et al., 2023). 

Sülflow et al., 2019 found that factors like credible source, as well as more comments on the content will hold the attention for longer (Sülflow et al., 2019). Meanwhile, Kanuri, Chen and Sridhar (2018) discovered that content which requires higher cognitive processing might lead to higher attention retention (Kanuri et al., 2018). However, this conflicts with the research of Alter and Oppenheimer from 2019, which discovered that the easier the processing, the likelihood of holding the attention increases (Alter & Oppenheimer, 2009). Moreover, emotions can heighten attention, but not all emotions work the same way. Language linked to uncertainty (e.g., anxiety or hope) and high arousal (e.g., excitement or threat) is more likely to hold attention because it activates cognitive and physiological vigilance. In contrast, low-arousal emotions like sadness are less effective in sustaining attention (Öhman et al., 2001; Teeny et al., 2020). Furthermore, uncertainty and certainty also have impact. On one hand, certain language will lead to higher engagement in the form of likes (Pezzuti et al., 2021). On the other hand, uncertainty will sustain the attention for longer, as users will be looking for a resolution, so they will spend more time attaining to the information (Tiedens & Linton, 2001).

Therefore, this research needs to consider the post source credibility (Facebook business profile credibility), complexity of message, emotional language and level of uncertainty presented in content, while measuring engagements in the form of comments to avoid these variables to influence the results.

Engagement on social media

As discussed before, engagement is another factor that will directly influence the organic reach of a content.

Prior research identifies a variety of content-related factors that influence engagement on Facebook, including post topic, emotional triggers, and visual elements. For example, political content often prompts more comments and shares due to its identity-driven and emotionally charged nature, while health-related posts tend to be shared for their perceived utility, even if they don't spark active discussion (Tenenboim, 2022).

The format of a post also plays a critical role in shaping engagement. Shahbaznezhad et al. (2021) found that photo content consistently generates higher engagement in the form of likes, comments, and shares, compared to other formats like video. While both photos and videos can encourage passive engagement (e.g., likes), photos are more likely to prompt active interactions, such as comments and shares. Furthermore, when the content is rational or information-based, typical for news posts, using a photo format leads to significantly more likes than comments, suggesting that even subtle format differences can shift the type and depth of user engagement (Shahbaznezhad et al., 2021).

Emotional valence also influences behavior. High-arousal emotions such as anger have been shown to increase all forms of engagement, while sadness, a low-arousal emotion, tends to result in passive engagement like liking or sharing but not commenting (Karnowski et al., 2021; Tenenboim, 2022; Valenzuela et al., 2017)). Posts that include surprising elements or prominent figures also tend to perform better across engagement metrics (Tenenboim, 2022; Trilling et al., 2017).

Exploratory behavior on social media

While engagement metrics like likes and shares reflect surface-level interaction, clicking on a link to access additional content represents a deeper form of engagement, referred to as exploratory behavior (Qiu & Golman, 2024; Reio, 2011). For news organizations on Facebook, this behavior is critical to generating organic web traffic (Anderson, 2024). Understanding what drives users to click involves examining post design, audience characteristics, and cognitive processing models.

One of the main drivers of exploratory behavior is information seeking. Several researchers have discovered that the headline is one of the main drivers for this behavior. The more information gap or uncertainty is present, the higher the clicks. Furthermore, when curiosity is created through the headline structure, the clicks to read more are also higher (Qiu & Golman, 2024; Scacco & Muddiman, 2020). Another factor is topic’s relevance and consumer interest in it (Tenenboim, 2022). Nonetheless, as priorly mentioned, Media Gridz as a product, does not alter headlines, so this research requires isolation of the level of curiosity, uncertainty and relevance.

Research on social media news consumption also shows that users rely on source credibility and social proof, such as visible likes and comments, to determine whether content is worth clicking on. Posts that show more engagement tend to attract higher click-through rates (Mayer et al., 2024).

Crucially, link placement plays a defining role. A large-scale analysis of Facebook posts across 172,000+ business pages found that posts with links embedded directly in the post body receive substantially higher engagement compared to those with links placed in the comments. Photo posts without links received 68% more engagement than those with links, and text-only posts without links had 125% more engagement than linked versions. Moreover, the practice of placing links in the comments is extremely rare, used by just 0.24% of brands, likely due to reduced visibility on mobile devices and the extra effort required to locate the link (Joyce, 2021). These findings suggest that placing links outside the main body of the post introduces friction, which in turn reduces the likelihood of clicks. This is particularly relevant when evaluating whether image-based posts with links in comments can perform as well as traditional link-based posts. If locating the link disrupts the user’s scrolling experience, the action may be abandoned altogether. The Dual-System Theory can be used to explain this phenomenon, as most clicks on social media occur via System 1 processes. This means users are more likely to engage with content that is easily accessible and requires minimal effort, such as clicking a clearly visible link within the post itself. In contrast, when a post format forces users to search for the link (e.g., in the comments), it disrupts this automatic behavior and shifts the task into System 2, increasing cognitive load and potentially reducing click-through rates (Kannengiesser & Gero, 2019).

This is further confirmed in studies from 2009 and 2014, which discovered that the presence of link within the image post, led to higher click rates, particularly for mobile audiences, where fast, intuitive browsing dominates. These effects are consistent with the idea that users engaged in low-effort news encounters are more responsive to easily processed behaviors (Grabe & Bucy, 2009; Townsend & Kahn, 2014).

Audience familiarity also influences exploratory behavior. Pócs et al. (2021) found a strong positive correlation between clicks and nonfan reach. In other words, non-followers, who are less familiar with the page’s content, are more likely to click to learn more. This suggests that exploratory clicks are driven by curiosity and novelty, particularly when users encounter unfamiliar content through the platform’s algorithmic distribution.

Exploratory click behavior on Facebook is shaped by a combination of design clarity, link visibility, audience familiarity, and cognitive effort. Posts with embedded links and visual elements are more likely to trigger intuitive, low-effort engagement, particularly among mobile users. In contrast, formats that increase friction or disrupt habitual browsing behavior may reduce the likelihood of clicks, even when the content itself remains the same (Kannengiesser & Gero, 2019). These insights provide a clear rationale for examining how post format alone, image-based posts with links in comments versus standard link-based posts, affects organic traffic.

Results

What is the difference between link-based posts and image-based posts in terms of organic reach on Facebook?

The analysis of organic reach for both link-based posts and image-based posts reveals that image posts have a higher reach compared to link posts. As seen from Table 1 the higher mean of 1197 for image posts suggests that image-based posts tend to reach a larger audience compared to link-based posts. Furthermore, the standard deviation for image posts is notably higher than for link posts, indicating that the reach for image posts is more variable.

Table 1
Descriptive Statistics comparing Means of reach, sessions from Facebook to article, shares, comments and reactions between link and image-based post.

Equal variances are assumed between link posts and image posts via Levene’s Test for Equality of Variances (p = 0.063). An independent samples T-test performed indicates that, the t-value is found to be -2.762, with 6 degrees of freedom, and the two-tailed p-value was 0.016, so the difference in reach between image-based posts and link-based posts is found to be statistically significant, indicating that image posts generally have a higher reach compared to link posts. Table 2 shows the relationship between reach and comments, reactions, and shares. A strong positive correlation is present between reach and comments, with a Pearson correlation coefficient of 0.870 and a p-value of 0.005, indicating that as the number of comments increases, the reach of a post also tends to increase. A moderate positive correlation is observed between reach and reactions, with a Pearson correlation coefficient of 0.760 and a p-value of 0.029, suggesting that posts with more reactions also tend to have higher reach. Although, a moderate positive correlation between reach and shares is found, the p-value of 0.104, indicates that while a relationship exists between reach and shares, it is not statistically significant enough to conclude that more shares directly lead to higher reach.

Table 2
Correlation analysis

What is the difference between link-based posts and image-based posts in terms of engagement in the form of likes, shares, or comments on Facebook?

Table 3, 4, 5 give descriptive analysis of how shares, reactions and comments performed. The analysis of engagement metrics for link-based and image-based posts shows that, on average, image posts outperform link posts in terms of shares, comments, and reactions. Based on the mean of shares it is suggested that image posts generally receive more shares than link posts, though the higher standard deviation for image posts indicates that the number of shares is more variable (see Table 1).

Table 3
Descriptive Statistics frequencies of shares of each of the 8 posts

Table 4
Descriptive Statistics frequencies of comments of each of the 8 posts

Table 5
Descriptive Statistics frequencies of reactions of each of the 8 posts

The independent sample t-test conducted indicates that the difference in shares between image-based and link-based posts is not statistically significant, as the Levene's Test for Equality of Variances shows no significant difference (p = 0.140), so equal variances are assumed and the independent samples t-test reveals a t-value of -0.758 with 6 degrees of freedom and a two-tailed p-value of 0.239.

For comments, link posts had an average of 0.75 comments, with a standard deviation of 1.500, while image posts had a higher average of 2.75 comments, with a standard deviation of 2.363. Image posts generate more comments on average compared to link posts, but, similar to shares, the higher standard deviation for image posts suggests greater variability in the number of comments across posts. Levene's Test for Equality of Variances is not significant (p = 0.418), indicating that the assumption of equal variances holds. The t-test shows a t-value of -1.429 with 6 degrees of freedom and a two-tailed p-value of 0.101, which is greater than 0.05. This result suggests that the difference in comments between image-based and link-based posts is not statistically significant.

Regarding reactions, link posts had an average of 3.75 reactions, with a standard deviation of 3.304, while image posts had an average of 5.50 reactions, with a standard deviation of 5.802. Image posts also outperform link posts in terms of reactions, but as with shares and comments, the variability in reactions is higher for image posts. Levene’s Test for Equality of Variances (p = 0.221) suggests that equal variances can be assumed. The t-test results in a t-value of -0.524 with 6 degrees of freedom and a two-tailed p-value of 0.309, which is greater than 0.05. This indicates that the difference in reactions between image-based and link-based posts is not statistically significant.

The analysis also explores the relationships between engagement metrics (See Table 2). A moderate positive correlation is found between shares and comments (r = 0.782, p = 0.022), suggesting that posts with more comments tend to be shared more, and vice versa. This relationship is statistically significant (p < 0.05). Similarly, a moderate positive correlation is observed between shares and reactions (r = 0.792, p = 0.019), indicating that more reactions tend to result in more shares, and vice versa. This correlation is also statistically significant (p < 0.05). Finally, a strong positive correlation is found between comments and reactions (r = 0.908, p = 0.002), suggesting that as the number of comments increases, the number of reactions also tends to increase. This relationship is statistically significant (p < 0.01)

These findings show that while image posts generally generate more engagement than link posts, the differences in shares, comments, and reactions between the two post types are not statistically significant. However, the positive correlations between shares, comments, and reactions suggest a strong interrelationship between these engagement metrics. 

What is the difference between link-based posts and image-based posts in terms of generating exploratory behavior?

Table 6
Descriptive Statistics frequencies of sessions from the Facebook posts of each of the 8 posts

Note. SessionfromFacebooktoarticle mean Unique Sessions that are measured via Google Analytics

Table 6 provides an overview of sessions from each post. The analysis of sessions from Facebook to the article reveal that image posts slightly outperform link posts in driving click-throughs to the article (See Table 1). For link posts, the average number of sessions is 14.50, with a standard deviation of 17.137. In contrast, image posts have an average of 18.75 sessions, with a standard deviation of 13.985. While image posts have a higher mean for sessions, the standard deviation for link posts is larger, indicating more variability in the number of clicks for link posts. In contrast, image posts show more consistent click-through behavior, as evidenced by the smaller standard deviation.

Levene’s Test for Equality of Variances shows that equal variances can be assumed for the two groups (link posts and image posts) (p = 0.679). An independent samples t-test reveals that the difference in the number of sessions between image-based and link-based posts is not statistically significant, as the t-value is -0.384, with 6 degrees of freedom, and the two-tailed p-value is 0.357. Thus, despite the higher mean number of sessions for image posts, the difference in sessions between the two types of posts is not statistically significant.

Table 7 examines the relationship between reach and sessions. The Pearson correlation coefficient between reach and sessions is 0.457, indicating a moderate positive correlation. This suggests that as reach increases, the number of sessions also tends to increase. However, the correlation is moderate, meaning the relationship between the two variables is not very strong. The p-value for the correlation was 0.255, which is greater than 0.05, indicating that the correlation between reach and sessions is not statistically significant.

Table 7
Correlation between reach and sessions from Facebook to article

*. Correlation is significant at the 0.05 level (2-tailed).

Note. Sessions from Facebook to article, means clicks within the context of this research. They are captured through UTM tracking which is reflected in Google Analytics, analyzing the traffic that only comes from Facebook image-based post and link-based post. N is the number of posts done.

In summary, while image posts have a higher average number of sessions compared to link posts, the difference is not statistically significant. Additionally, the correlation between reach and sessions was moderate, but not statistically significant, suggesting that reach does not significantly influence the number of sessions to the website in this case.

Discussion

Discussion Sub-Question 1

What is the difference between link-based posts and image-based posts in terms of organic reach on Facebook?

The analysis of the difference in organic reach between image-based posts and link-based posts revealed that image posts generally have a higher reach than link posts. This difference is statistically significant, indicating that Facebook algorithm tends to prioritize image-based posts over link-based posts. The algorithmic preference for image posts suggests that this form is more likely to be prioritized and shown to a broader audience within the platform, which is also the first step required to generate any other form of engagement (Berger et al., 2023).  

The positive correlation between engagement metrics and reach, suggests that reactions contribute to enhancing the reach of a post. In addition to reactions, the increase in number of comments on a post leads to more reach, which is also in line with previous research stating that posts that generate more discussion tend to have a larger organic reach (Shahbaznezhad et al., 2021). However, it is important to note that image posts, by their very design, tend to include a link in the comments, which may confound the results.

On the other hand, the variable of shares, which was expected to be a significant driver of reach, did not result in a statistically significant relationship (Pócs et al., 2021). This suggests that user behavior in terms of sharing content does not boost the reach of posts as much as previously anticipated.

In summary, the results suggest that image-based posts tend to have a higher reach than link-based posts due to Facebook’s algorithmic preferences. Additionally, reactions and comments play a role in enhancing reach, while shares do not appear to be as influential in driving reach as initially expected. These findings emphasize the need for media outlets to focus on creating content that generates reactions and comments, and to consider the algorithm's prioritization of image-based content to maximize their organic reach on Facebook.

Discussion Sub-Question 2

What is the difference between link-based posts and image-based posts in terms of engagement in the form of likes, shares, or comments on Facebook?

The analysis of engagement metrics such as likes, shares, and comments revealed that there were no statistically significant differences between link-based posts and image-based posts in terms of these metrics. This suggests that the format of the post itself, whether it is an image or a link, does not have a direct impact on the level of user engagement on Facebook.

Despite the lack of significant differences, the analysis uncovered an important observation: the engagement metrics are interrelated. Posts with more comments tend to be shared more, and vice versa. Similarly, reactions are found to lead to more shares, and more comments tended to generate more reactions. This interconnectedness shows that as one type of engagement increases, it can encourage further engagement, creating a feedback loop that amplifies the visibility of content.

The lack of a significant difference in engagement based on post format aligns with previous research indicating that users are more likely to engage with content that aligns with their existing interests, preferences, or current emotional state (Öhman et al., 2001; Teeny et al., 2020). This suggests that engagement is not necessarily driven by whether the post is an image or a link, but rather by the type of content and how relevant or compelling it is to the user. Users focus their attention on content that resonates with them, and they are more likely to engage with posts that align with their values, interests, or social context, regardless of the format (Öhman et al., 2001; Teeny et al., 2020; Tenenboim, 2022).The Theory of Extended Self offers another perspective on why engagement does not differ between the type of post. As individuals extend their digital selfs through their behavior and interactions on social media, users are more likely to engage with content that aligns with their aspired digital identity. Content that resonates with a user’s values, beliefs, or self-concept is more likely to be shared, commented on, and reacted to, because it validates their extended self (Belk, 2016). This reinforces the idea that engagement is not solely about the format of the post, but about how closely the content aligns with the user’s identity and values.

Furthermore, the ripple effect theory provides a useful lens for understanding how engagement metrics are intertwined. According to this theory, each user’s interaction with a post, creates a chain reaction, increasing the visibility of the content to others in their social network. The study findings demonstrates that the ripple effect is a key driver of social media interactions, where engagement leads to more engagement and more visibility (Hsu et al., 2015).

In conclusion, the findings suggest that the engagement on Facebook is not significantly influenced by the format of the post (image vs. link), but by the content itself and the social dynamics of user interactions. Content’s value and emotional valence combined with, the ripple effect, and the Theory of Extended Self all contribute to understanding how engagement metrics are interconnected and how users’ interests, emotional responses, and social contexts shape their interaction with content. The interrelationship between likes, shares, and comments highlights the complex and interconnected nature of social media engagement.

Discussion Sub-Question 3

What is the difference between link-based posts and image-based posts in terms of generating exploratory behavior?

The analysis of exploratory behavior, measured by click-throughs (sessions) from Facebook to the website, revealed that image-based posts generally showed more consistent organic web traffic rate, as indicated by the smaller standard deviation in session data. Despite the higher mean number of sessions for image posts, the difference between image and link posts in terms of organic web traffic is not statistically significant. This finding suggests that while image posts may drive slightly more web traffic, the difference is not substantial enough to be considered statistically significant. The lack of significant correlation between reach and sessions further implies that other factors, such as the content of the post, play a more substantial role in driving users' exploratory behavior.

A key theory that may help explain these findings is Curiosity Theory, which posits that individuals are motivated to explore and engage with content that makes them curiosity (Scacco & Muddiman, 2020).  In this study, it seems that the content itself, specifically how the headline is structured, may be a more powerful driver of exploratory behavior than the format or reach of the post. Research has shown that when there is an information gap or uncertainty (such as a headline that hints at something intriguing), users are more likely to click through to satisfy their curiosity (Qiu & Golman, 2024; Scacco & Muddiman, 2020). This suggests that content that sparks curiosity, through its headline or topic relevance, will more effectively drive clicks, regardless of whether the post is an image or a link.

Moreover, the Dual System Theory offers valuable insights into why reach does not always correlate with more web traffic. Most social media engagement, including clicking on links, occurs via System 1 processes, meaning users are more likely to engage with content that is easily accessible and requires minimal effort, such as clicking a clearly visible link within the post. In contrast, when a post forces users to search for the link, such as when it is placed in the comments, it disrupts System 1 behavior and pushes users into System 2 thinking. This shift increases cognitive load and can reduce the likelihood of clicking, even if the format itself is compelling (Kannengiesser & Gero, 2019). Additionally, exploratory click behavior on Facebook is influenced by the clarity of post design, link visibility, audience familiarity, and the cognitive effort involved. Posts with embedded links and visually appealing elements trigger intuitive, low-effort engagement, particularly on mobile devices, where users expect seamless interactions (Grabe & Bucy, 2009; Townsend & Kahn, 2014). Formats that increase friction, such as hiding links in comments or requiring more cognitive effort, can disrupt habitual browsing behavior and reduce click-through rates.

In conclusion, while image-based posts showed slightly higher click-through behavior, the format of the post alone did not significantly influence exploratory behavior. Instead, factors such as content relevance, design clarity, and the ease of interaction might be more critical drivers of users’ decision to explore further. Curiosity theory and the System 1/System 2 cognitive framework help explain why users may choose to engage more with content that is easily accessible and aligns with their interests, while content that requires more effort or disrupts habitual behavior may reduce exploratory clicks.

Conclusion

This study explored the differences between link-based and image-based posts in terms of their impact on organic reach, engagement, and exploratory behavior on Facebook as these are the main factors that determine if social organic web traffic for news websites will be achieved.

The analysis revealed that image-based posts tend to achieve higher organic reach than link-based posts, confirming previous literature stating that Facebook's algorithm prioritizes images over links (Froome, 2022). This algorithmic preference suggests that media outlets aiming to maximize organic reach on Facebook should consider focusing more on image-based content. Reach is further increased when more presence of comments and reactions is achieved, but not necessarily more shares. It’s important to note, that the experiment deploys the A/B testing technology of Meta which shows the post mainly to a subset of Eurocom’s audience, which might affect the “share” metrics (Business Help Center, n.d.).

Despite the higher mean number of sessions inside the website for image posts, the difference in exploratory behavior is not statistically significant, indicating that while image posts show slightly higher traffic, the post format alone does not substantially affect social organic web traffic. One way that that could be explained is the Dual System Theory. The ease of engagement might be of a significant role in driving sessions, as System 1 is usually utilized by users in their social media consumption. Content that requires minimal effort, such as an easily accessible link within an image post, encourages more clicks compared to posts that require users to search for links (e.g., in comments), which increase cognitive load and discourage interaction (Kannengiesser & Gero, 2019).

However, since in terms of engagement, the format of the post is not resulting in significant differences across likes, shares, and comment, as well, another possible explanation could be factors such as content relevance, level of curiosity and uncertainty or content design (Kannengiesser & Gero, 2019; Öhman et al., 2001; Shahbaznezhad et al., 2021; Teeny et al., 2020; Tenenboim, 2022; Tiedens & Linton, 2001).

This study also discovered that engagement metrics are interrelated, where shares, interactions and comments often reinforce each other, indicating that engagement on Facebook is a dynamic, interconnected process. These findings also support the Ripple Effect theory in virtual communities like Facebook (Hsu et al., 2015).

In conclusion, considering image-based posts have higher reach, likes and comments, the recommended format for driving more organic web traffic would be the image-based format, but to essentially achieve higher web traffic, content, headline and design will still play an important role.