The Meaning of Like: How Social-Media Editors and Users Make Sense of Social Media Engagement

ABSTRACT

News organizations today rely on various metrics to assess users’ engagement with news. Focusing on social media engagement metrics, the present study argues that their non-proprietary nature sets them apart from other measurement systems. These metrics were introduced to invoke social activities on social media platforms among users. Therefore, they serve as a unique case in which both users and social media editors are obliged to take part in the same interpretive process. Based on 117 semi-structured interviews with Israeli users, and 18 interviews with Israeli social-media editors, the study outlines the reciprocal interactions between these two stakeholders that shape news dissemination on social media. In particular, the study discusses how users and editors assign meaning to social media engagement metrics, based on the social rules that govern news sharing on these platforms. It also finds that, editors are fully aware of users’ “social rules” of engagement, and try to harness them to advance the revenues they obtain from their social media activities.

KEYWORDS:

• Social media
• audience metrics
• news engagement
• editorial processes
• news engagement
• affordances

Introduction

Metrics assessing media audiences which are based on “big data” (Nelson and Webster 2016) represent the latest stage in the “industrial construction of audience perspectives” (Anderson 2011), and are now considered the standard measure of users’ tastes and preferences. Due to their objective, quantified nature, metrics are often perceived as an accurate reflection of users’ preferences (Meijer and Kormelink 2020). However, this perception has been revealed to be a mirage (Nelson 2018). Much like other audience measurement systems (Napoli 2014), news organizations need to make sense of these metrics in order to gain insight into their audiences.

The current study delves into this sense-making process and its implications for news production by focusing on social media engagement metrics: the quantified representation of users’ engagement with content posted on social media (i.e., interaction scores). These metrics include, for instance, the visual representation of the number of Likes, Shares, and Comments on Facebook, or the number of retweets on Twitter. It is argued here that their non-proprietary nature (Chua and Westlund 2019) makes these metrics a unique case—a focal point that captures many of the tensions influencing journalism today. Social media engagement metrics are characterized by an inherent duality (Paulussen, Harder, and Johnson 2016): Since they were not created as an audience measurement system but rather as a tool for users to encourage interactions among them (Gerlitz and Helmond 2013), they serve, simultaneously, two different stakeholders with different agendas. When users are concerned, engagement and engagement metrics are indeed social connectivity tools, in keeping with their original design. At the same time, however, editors are also aware that scoring high on these metrics is crucial, since it is a key element in social media curation processes (Lee and Tandoc 2017): Both algorithms and users reward content with high engagement scores. This entanglement of purposes, motivations, and uses makes the interpretation of such metrics a complex matter indeed.

Considering the complexity of social media engagement metrics, this study takes a closer look at the sense-making process carried out by both users and social media editors, and at the influence of this process on editorial practices. While research hitherto has mainly focused on practitioners’ interpretation of audience measurement systems, the duality of social media engagement metrics requires that attention be given to users as well. Since users are also exposed to these metrics and use them when choosing content (Meijer and Kormelink 2020), this is a unique case in which they are obliged to take part in the same interpretive processes as do editors. Therefore, the current study builds on the strength of a conjoined analysis by conducting interviews with Israeli social media editors and Israeli Facebook users. The aim is to identify how users and editors grasp the complex nature of social media matrices, and how their understanding is reflected in their everyday practices. The study addresses three main questions, which are elaborated below. To begin with, a conjoined analysis can identify both parallelism and divergence between users’ and editors’ perceptions of social media matrices (Heise et al. 2014). The question that arises from this approach is, What meanings do both these stakeholders read into these matrices, and to what degree do both sides share these meanings? Second, it was suggested that social media had become a shared space for interaction between news organizations and their audiences (Lawrence, Radcliffe, and Schmidt 2018). A question to probe this aspect would be, Do users and editors perceive social media matrices as a communication tool through which users can convey their opinions and preferences to the news organization? And finally, centering on matrices as an audience measurement system, one should consider the influence of editors’ understanding of these matrices on news creation and distribution. This could be articulated as the following question: Do editors identify the complex nature of these matrices, and if so, how does this understanding interlace with their editorial decisions? All of these issues have direct relevance to our ideas regarding journalism work. Meijer and Kormelink (2020) claim that, “[s]ince audience metrics have become pervasive forces for commercial and public service media, paying attention to the usage patterns of audiences has become crucial for the survival of journalism” (12). Understanding the delicate interactions between users and editors is one step further in this direction. Such an approach can reveal the paths through which social media affordances shape the news flow: how the framework of social motivations and norms influences the distribution of news, not only by users but also by editors.

What Sets Engagement Metrics Apart from Other Audience Measurement Systems

The term “social media engagement metrics” designates the aggregate score of social interactions with social media content (van Dijck and Poell 2013). Accordingly, all news engagement behaviors—Likes, comments, sharing, tagging, etc.—become indicators of attention that are presented in a quantifiable manner alongside the news post (part of what Nelson (2019) defines as reception-oriented audience engagement). As such, they are a specific case of metrics used in the news-room to track users.

Notably, however, there is a key distinction between social media engagement metrics and other metrics that are sometimes referred to as web-analytics (Tandoc and Vos 2015). Web-analytics track users’ footprint within the news website itself (i.e., page views, time spent on website, number of clicks on a headline). Therefore, they are only visible to the news organization,¹ and play but a small part in users’ perceptions of news content. Social media engagement metrics, on the other hand, are the result of users’ behavior on platforms non-proprietary to the news media (Chua and Westlund 2019), and consequently, follow a different logic.

Social plug-ins, the features that allow users to engage with news, were introduced as a social activity, “an easy way of sharing web content with one’s contacts in order to invoke further social activities on the platform such as resharing, commenting and later Liking” (Gerlitz and Helmond 2013, 152). Thus, they are first and foremost a social tool for users, and their functionality for publishers is an epiphenomenon of their original design. In fact, the literature on social media engagement metrics suggests that they are visible to four different stakeholders (Meijer and Kormelink 2020; Nelson 2018; Oeldorf-Hirsch and Sundar 2015; Picone 2015; van Dijck and Poell 2013): (1) users can see the engagement metrics of each news post in their feed; (2) other members in a user’s network can see if a specific user engaged with a news post; (3) news organizations, and mostly social media editors, who are in charge of the organization’s accounts; (4) the platform itself collects data from these metrics and uses it in their algorithmic curation processes.

The present article makes a case that this duality—being simultaneously a social tool and a measurement system—makes social media engagement metrics an interface between news organizations and users that warrants investigation. I specify three inter-related areas in which the dual nature of social media matrices could be of significance. The first is the sense-making process that these metrices necessarily involve: What meaning do users and editors assign to them? Do these actors recognize the duality of social media matrices as both social tools and measurement systems, and if so, how do they understand it? The latter, general, question can be narrowed down to one germane specifically to communication: To the extent that social media matrices are both social tools and media matrices, are they perceived as a tool whereby users communicate with news organizations? Finally, insofar as other measuring systems are known to shape news production, how does editors’ interpretation of these matrices translate into editorial practices?

Each of the following sections is devoted to one of these areas, elaborating on current literature and presenting a rationale for the respective research question.

Interpreting Social Media Engagement Metrics

Metrics are the newest audience measurement system, and much like their predecessors, they are conventions, or facts by human agreement (Napoli 2014). As Napoli puts it, since these systems have all aimed to “serve as currency in the audience marketplace[,] such currencies involve the unanimous acceptance and utilization of a particular set of social constructions as facts” (227).

What makes deciphering social media engagement metrics particularly complex is their visibility. Because users’ engagement with news is visible to other members in their network, it is fraught with “concerns about personal reputation and privacy: what and to whom do I communicate about myself by engaging in these practices?” (Meijer and Kormelink 2020). Thus, engaging with news is used for impression management purposes, to maintain ideal self-images (Duffy and Ling 2020; Picone 2015; Picone, De Wolf, and Robijt 2016) and to strengthen social capital (Gil de Zúñiga, Jung, and Valenzuela 2012). A slew of engagement practices—posting, commenting, Liking, and tagging—allow users to share or demonstrate their involvement with news either more or less publicly (broadcast level: Oeldorf-Hirsch and Sundar 2015), and consequently, involving different levels of effort (Picone et al. 2019). Thus, according to Meijer and Kormelink (2020), social media engagement practices “exist on a spectrum from smaller to minuscule acts of communication and from content-related acts (engaging with news itself) to phatic acts of communication” (35). These authors suggest that tagging marks the “phatic” end of the spectrum, whereas commenting marks the “content” end. As such, users treat these features as social gestures that help sustain different engagement practices. Hence, these metrics are by no means an objective measure of users’ interest.

While in most cases matrices are only visible to news professionals, which means they are the only ones interpertaing them (Napoli 2014; Nelson and Webster 2016), social media engagement metrics are public, and guide users’ behavior on social media. Thus, studies have documented the impact of engagement metrics on users’ consumption of news and their engagement with it (Meijer and Kormelink 2020). The effect of matrices depends on both their quantity (e.g., the number of Likes, comments, etc.) (Messing and Westwood 2014), and their quality (the type of engagement measured by the matrices) (for review see: Dutceac Segesten et al. 2022). This finding implies that users also need to make sense of these metrics. If engagement involves phatic communication (Duffy and Ling 2020), and if users differentiate between types of engagement, then they should carry out an interpretive process that helps them construct the meaning of this phatic communication. This makes social media engagement matrices a unique case in which both editors an users must deploy sense-making processes. Indeed, since other users are the target of phatic communication, their interpretation of these metrics is as crucial as that of an editor. Due to the social origin of these metrics (Gerlitz and Helmond 2013), it is the establishment of a social norm regarding the meaning of a Like or a Share that infuses them with value. In this sense, metrics are a mirror that reflects users’ behavior to editors and users alike.

Seeing as both users and editors participate in similar sense-making processes, the study compares these two stakeholders’ understanding of social media engagement metrics, in an endeavor to elucidate the interactions between them. Indeed, due to the strong element of reciprocity in audience-journalist relationships on social media (Lewis, Holton, and Coddington 2014), it is important to establish whether users and editors ascribe the same meaning to social media engagement metrics.

RQ1: How aligned are users’ and social media editors’ interpretations of social media engagement metrics?

Social Media Engagement Metrics as a Shared Space

The first research question of this study pertains to the visibility of all of one’s news engagement practices to one’s social network. The second question pertains to the visibility of social media engagement metrics to both users and news organizations, each side being potentially aware that the other is tracking these metrics.

Scholars largely concur that social media is the main platform on which news organizations and their audiences interact (Lawrence, Radcliffe, and Schmidt 2018), and that they do so in direct, reciprocal, and visible ways (Lewis, Holton, and Coddington 2014). The directness, reciprocity, and visibility of such interactions all manifest themselves in social media engagement metrics, rendering them a “shared space” (Heise et al. 2014). By means of these metrics, users can now assess the performance of news posts and discuss it with other users and with the news organization. This is one of many examples for the changing balance of power between news organizations and their audiences as a result of the visibility enabled by the digital age (Wilhelm, Stehle, and Detel 2021). The questions that are salient in this regard would be, Do users and editors see social media engagement metrics as channels of communication? and, Do users take advantage of these channels to convey their preferences to the news organization—and, for that matter, do editors believe users are doing so? In sum:

RQ2. Do users and social media editors see social media engagement metrics as a communication tool?

Social Media Engagement Metrics and News Production

Research heretofore has demonstrated how audience metrics guide news production processes (Lee and Tandoc 2017). For instance, metrics were found to influence issue coverage and placement: Topics that received greater attention in the past tended to be covered more often and placed in a more prominent place on the website (Lee and Tandoc 2017). Social media engagement metrics are potentially influential in this regard as well. First, like any other type of metrics, they reveal users’ tastes; at the same time, as opposed to other metrics, they also determine the level of exposure a news post receives on social media. In a process known as algorithmic curation (van Dijck and Poell 2013), social media algorithms are programmed to push content that scores high on engagement metrics, ensuring it reaches more users (Tandoc and Maitra 2017). Editors are therefore geared towards improving their “metrics score” to foster exposure.

Against this background, this study explores how interpreting these metrics translates into editorial practices. Put differently, the study moves away from explanations for the influence of user engagement based on quantity alone. Rather, it endeavors to establish whether editors have developed a more nuanced, qualitative understanding of engagement that translates into different practices.

RQ3. How does social media editors’ interpretation of metrics influence news production?

Method

Semi-structured interviews were conducted with 18 social media editors and 117 Facebook users. The study focused on Facebook since it is the most used platform in Israel, by both news organizations and users, with news organizations devoting most of their resources to this platform (Dvir-Gvirsman and Tsuriel, 2022).² Israeli media market is characterized by a hybrid media system (Hallin and Mancini 2011; Tenenboim-Weinblatt 2014); originally it demonstrated party-press parallelism, and still maintains parts of this logic, but neo-liberal processes drive it to adopt the characteristics of a Liberal system. It is also characterized by high levels of news consumption and of social media use (Wike et al. 2022).

Social Media Editors

This study focuses on social media editors, since they are in charge of news organizations’ social media activities, and their day-to-day job is to handle engagement metrics (Ferrer-Conill and Tandoc 2018). Social media editors were contacted directly by the author and asked to participate in the study. These editors work for 18 news organizations, representing the majority of Israeli media: 5 print newspapers which also operate online via their websites; 6 digital-only news websites (though 1 of them is part of a large media group that owns a print paper and might use online items in print); 4 TV stations which also use websites (3 broadcast, of which one is public and 2 are private; 1 cable); 1 magazine; and 1 radio station (the 2 latter organizations operate online).³

Due to Covid-19, the interviews were conducted via phone, Skype or Zoom. Each interview lasted around half an hour; the interviewees were not remunerated for their time. The interview guide included 33 questions, targeting their use of social media and media metrics, their perceptions of users’ engagement (Likes, comments etc.), and their beliefs about users’ views on engagement.

Facebook Users

A total of 117 Facebook users were interviewed, as described below. First, we performed a secondary analysis of existing interviews (Heaton 1998) conducted with 83 students with active Facebook accounts who reported that they obtained news from their feed at least occasionally (Set 1). This research population comprised 40% men and 60% women, ranging in age between 18 and 35 (average age: 24.8). In addition, a second set of interviews (Set 2) was conducted, tailored to the current study. This set included 34 participants who were chosen to complement the first set in terms of age and socioeconomic background: 20 of the 34 were 40–60 years old, 10 of these had an academic education; eight were aged 24–39 and of lower socioeconomic status; six were young adults with low socioeconomic backgrounds and secondary school diplomas. Genders were divided evenly. All participants were recruited through a Facebook post (Whitaker, Stevelink, and Fear 2017). For the first set of interviews (Set 1), participants (students) were recruited through Facebook groups for university students. For the second set (Set 2), the objective was to reach interviewees from different socio-demographic backgrounds, and therefore the recruitment posts were shared in local neighborhood groups, which differed in economic levels. All interviews in Set 1 and three interviews in Set 2 were conducted in person. Most interviews in Set 2 were conducted via Zoom due to Covid-19. The study employed convenience sampling, and therefore all analyses are based on samples that are not representative.

Following Kümpel, Karnowski, and Keyling (2015), at the beginning of the interviews participants were instructed to browse their Facebook accounts for as long as they wished (average browsing time was approximately 10 min) and were informed that the session would be recorded. This recording was used at the end of the interview: the interviewer and the participant watched it together and discussed the latter’s actions during the browsing session and their rationale (i.e., reading/not reading a news story, commenting, etc.). The average length of each interview was one hour, and participants received approximately $35 as compensation for their time.

Interviewers were guided by a list of 15 questions, but conducted fluid discussions following the participant’s train of thought. We asked the interviewees to describe their general news consumption habits, overall use of social media, and engagement with news on social media (Facebook and otherwise): how they choose between the different affordances and why. We specifically asked them to elaborate on the way they interpreted engagement metrics; this line of questioning included such questions as, What does it mean when someone Likes or comments on a news story? What does it mean if a news story has only few / a lot of Likes, comments etc.? Lastly, we asked them about their reasons for engaging with news on social media and about the importance of engagement scores in this respect.

The interviews from both groups were subjected to a thematic analysis through deconstructing texts into units and identifying common meanings and patterns. In the interviews with users, no major differences emerged between the two sets in respect of the themes identified; hence, in what follows, the entire sample is treated as a single unit.

Findings

In what follows, the findings are organized in accordance with the three research questions. First, the meaning users and editors assigned to the matrices is described, with focus on the effect of the dual nature of these matrices on interpretation processes. Second, the idea of matrices as a communication tool is addressed. Lastly, the impact of editors’ interpretation on editorial practices is discussed.

The Meaning and Interoperation of Social Media Engagement Metrics (RQ1)

Understanding Users’ Motivations

Before addressing editors’ and users’ interpretations of social media metrics, it is important to situate the interpretive process within the context of news sharing practices, as any reading of these metrics hinges on the original behavior which they measure.

Research hitherto has highlighted that news sharing is utilized to maintain one’s self-presentation online (Duffy and Ling 2020; Picone 2015; Picone, De Wolf, and Robijt 2016). In the current study, too, users described news as material employed to establish one’s desired online social persona. User1, one of the few participants who shared news, provided the most powerful testimony: “It was important to me for people to see it and say: “NAME [interviewee’s name] understands economics.” When I share such content [news about the Minister of Finance], I say to myself: “I don’t really know that much about economics, but this will designate me as someone who does.” I’m presuming to be something I’m not by participating in this semi-intellectual sharing.” However, for the most part, concerns over their social appearance prevented users from sharing (Meijer and Kormelink 2020), since they did not want “to “look like idiots,” “embarrass themselves” (User2) or be perceived as a “normie” (User3) by associating themselves with issues that are neither “cool” nor socially desired.

Interviews with editors revealed that they are fully aware that users share news as a social commodity. In their view, news is used for phatic reasons as a “conversation starter” (Meijer and Kormelink 2020). For instance, one editor reflected on the time when they posted an item about a shortage of mayonnaise in McDonald’s, and the post went viral. He explained that “there were a lot of people who were bothered by it, but not enough for it to surface. Once we posted it, we gave them the platform to convey it among themselves.” Second, editors believed that users shared news to better situate themselves within their network. Thus, in respect of Facebook users, Editor1 noted: “It’s a way to show the world that you are politically involved, that you care; to feel active without [actually] being active. But I’m not saying anything new, it’s almost a cliché.” It’s not just the appearance of involvement, it’s the social value gained: “By sharing, you get your friend to Like you: it gets you positive value” (Editor2). Editor3 demonstrated this point by drawing a contrast between Facebook and Twitter: sharing on Facebook is more visible and this visibility invests news sharing with greater symbolic value, or as he put it, with higher “visual dividends.” Therefore, according to Editor3, supporting a news item by sharing it on Twitter is less about appearances, and hence is more authentic. The above quote is noteworthy, as it demonstrates that editors recognize the power of visibility, which is at the heart of engagement metrics’ uniqueness.

Editors likewise acknowledged that, in consideration of impression management, users may choose not to engage with news publicly (Swart, Peters, and Broersma 2018). Editor4, for example, said: “Some people say, why do I need it? I don’t want the entire country to know what I think, that I’m bored, have nothing better to do than comment on Facebook.” Editor5 expressed a similar sentiment: “There are types of content people won’t Like, but they will click on the post. Sometimes people will comment, “Hey, why do you post items that nobody Likes?” but you can tell from the data that it is interesting; people just don’t want others to see. There is an awareness of what is OK to be seen and what is not.” Editor5’s words emphasize that editors are aware of the distinction drawn here between web-analytics like click ratio metrics and social media engagement metrics. Editors understand that the latter metrics are influenced by social motivations.

Interpreting Social Media Engagement Metrics through the Lens of Social Motivations

Participants in this study distinguished between different social practices based on social media features, in keeping with Meijer and Kormelink (2020). However, whereas Meijer and Kormelink (2020) described a spectrum, participants described a hierarchy where some features are used carelessly while more public and informative practices, such as comments and posts, are the most valuable and rare. User1 illustrated this vividly when he explained why he never comments on controversial issues, but might Like them regardless: “You’ll see hundreds or even thousands of Likes, so I’m right there on the list, but if I express a specific opinion [by commenting], I’m opening myself up to attack.” Thus, sharing practices are valued according to the strength of the affinity they create between user and content. This affinity is based on broadcast level (Oeldorf-Hirsch and Sundar 2015), i.e., its reach and the extent to which the user can control its target audience and the amount of information it provides to its originators (Meijer and Kormelink 2020). That the desire to manage one’s impression manifests itself through social media sharing practices is critical to the argument made here regarding the nature of social media engagement metrics. This finding establishes the link between users’ social motivations and engagement metrics, a link which is unique to social media setting due to the inherent duality of these metrics.

As sharing and selecting news are reciprocal behaviors in social media, the social value of sharing practices applies to the selection processes as well. When explaining how he selects news posts, User7 noted: “Likes, no. Posting is more interesting, because it’s more meaningful than Liking … [posting] is more of a challenge … Likes are less interesting unless the number is absurd.” User8 echoed the same sentiment: “When I Like something, I have no expectation of influencing anything [other users].” User8 remarked that comments influence him but Likes do not because “in a comment, someone wrote his own opinion; he made the effort.” One participant explained that having a friend share a news post was a strong enough indicator for her to Like the post even without reading the news story (together with her trust of the news website from which the post originated).

Editors also acknowledged the users’ hierarchy and that it is determined by the level of commitment and “how much I want to echo it in my community … active sharing [sharing while writing additional text] which means [the user] wants to echo it to the fullest” (Editor7). Editor4 said: “More shares means that what we posted is really good. If people want to publish it on their platforms, it’s not merely pressing Like and goodbye.” The choice of the phrase “their platforms” in this quotation is meaningful: it suggests that, from an editor’s point of view, by sharing content, users come to own it. By contrast, as Editor8 puts it, “[Likes] tell nothing and have no currency.” Editor9, for his part, goes so far as to quantify the relative value of Likes and shares, pricing the former at roughly half of the latter. Editors’ disparagement of Likes is partly attributable to their belief that “only news freaks read today, most people look for TL;DR (too long, don’t read) because they don’t have the time. It’s 2020!” (Editor10); in effect, users Like content they don’t actually read.

The findings obtained in the analysis indicate that users’ and editors’ interpretations of metrics align, as per RQ1. More importantly, they establish the argument made here about the dual nature of these metrics, which sets them apart from other audience measurement systems.

Social Media Engagement Metrics as a Communication Tool (RQ2)

Editor5’s quote above, to the effect that users comment, “Hey, why do you post items that nobody Likes?”—hints at the idea that, since social media engagement metrics are visible to users and editors alike, they constitute a space for communication on journalist practices (Heise et al. 2014). Users see these metrics and could, potentially, judge a news organization’s performance, comment on it, etc. Yet both sets of interviews revealed that this kind of direct acknowledgment of news organizations is rare. Addressing a news organization directly was hardly mentioned at all. Only very few (young) participants said they would Like a post as a way of sending a message to a news organization: “so you will know I saw it and liked it and you will continue posting such material!” [User4]. While users do not communicate with news organizations, they do communicate with journalists by following them on social media and/or engaging with content posted on their personal/professional accounts (Hedman 2020). These journalists are viewed as autonomous and not representatives of media organizations. Therefore, any communication with them takes a personal form, as the following two examples demonstrate. User5 noted that she follows a columnist devotedly, but she never responds because the columnist “has enough Likes.” User6 relayed how she had once commented on a story published by an older journalist, a well-known figure with a long history in Israeli public broadcasting. She wrote to him, saying she liked his recent story, and to her surprise, he later called to thank her. Thus, perhaps not surprisingly, interviewees applied different logic when considering journalists and news organizations. It was abundantly clear to them that journalists act like any other social media user: tracking the amount of social support and attention they receive from others in the network. Still, they could not imagine that, somewhere in the newsroom, there was a person whose job it was to sit and count Likes.

When asked if, in their opinion, users were aware that news organizations monitor their behavior on social media, editors tended to reject the idea. Their responses can roughly be grouped into three categories. The majority think that users are naïve, and one even expressed disbelief that they understand anything at all: “Gee whiz … You [the interviewer] have just flattered them all.” One editor explained that she uses her mother as an exemplar of the average reader and that her mother “has no idea, has no understanding whatsoever of what she is doing.” The second group feel that users might be aware of tracking systems, but are not bothered enough by this to give it a second thought: “99.99% of the public don’t care a fig about privacy.” The few exceptions made up the third group, most of them editors who cater to a more literate audience. For example, Editor6 said: “When someone is critical of you, they won’t share the link to your article. People understand that [sharing] is worth money … They are telling me, Why won’t you try to please me? Don’t you want traffic?”

To conclude, both users and editors perceive users’ engagement with news on social media as being driven by users’ social motivations. Engagement has a lot to do with communication among individuals, and less so with communication between news organizations and their audiences.

Influence of Social Media Metrics on Editorial Practices (RQ3)

In the past, it was suggested that engagement metrics enable editors to identify users’ tastes, and to gain a wider reach (Lee and Tandoc 2017; Tandoc and Maitra 2017; Vu 2014). Interviewees in the present study showed less concern for the former, while emphasizing the latter. In fact, one editor mentioned that they were considering using a survey to gain a better understanding of topics that interest their readership, without giving a second thought to the behavioral data available to him.

The emphasis on engagement as a vehicle for traffic is most clearly demonstrated in the way the editors thought of the sharing practices’ hierarchy. They may recognize the hierarchy ascribed to these practices by users, but this does not mean that their ranking judgments are the same. Instead, editors’ ranking of sharing practices is shaped by an imagined algorithm—how they believe the algorithm favors the various features. For instance, although both users and editors agreed that sharing marks the highest level of support for content, Editor11 preferred comments since, he said, comments have recently been prioritized by the algorithm: “A comment tells Facebook that the person read the post and has something to say: it’s meaningful. Share used to be a very strong tool, but now it doesn’t amount to much.” Editor12 echoes this sentiment, saying: “What Facebook told me is that what they now value the most is comments, because of the whole issue with Trump … it’s the most valuable measure, more so than shares. So, you know, you can encourage it: there are survey elements, or ask a question on the post—“what do you think?”—or ask them to tag friends.” When asked directly if his perceptions regarding importance were filtered through what he believes is important to Facebook, he replied: “It also serves me, it’s a win-win situation.” Thus, comments are highly valuable because of algorithmic curation, while in terms of journalistic norms they are perceived as “part of the corruptive culture of social media: people comment and most of the time it has nothing to do with the article” [Editor13].

Comparing users’ and editors’ perspectives is fruitful not only because it demonstrates the discrepancies between editors’ understanding of users and the latter’s actual practices, but also since it highlights the similarities between the two groups. In this regard it is worth noting that users we interviewed also frequently mentioned Facebook algorithms and, much like editors, imagined these algorithms’ “preferences.” They showed some understanding of algorithmic curation, and as such, added and removed Likes to influence the filtering process. Thus, one participant said: “I’m very meticulous. If I see something that doesn’t interest me I will unlike it so it will stop appearing,” and: “I try to make it [the algorithm] understand, to create a more rounded [Facebook] personality so I get more diverse content” (User20). While engagement metrics were considered a space shared by users and news organizations, it seems that both sides are as much geared towards the platform itself as they are to the other side. This stems directly from the non-proprietary nature of social media engagement metrics.

In Editor12’s quote cited above, he described the editorial tactics used to encourage users to comment. Indeed, the findings demonstrate that the editors’ ranking of practices is not a matter of preference alone: it also translates into actions. For example, Editor12 suggested that editors can more or less tell what type of content will provoke a specific response. Editor13 said: “There are certain posts that we call ‘tagging posts’ [which raise an issue or describe an event that might be relevant to some people]; this is the only reason they will succeed.” Editor14 expressed a similar view: “The type of content and the way you phrase it influences the type of response you will get.” Editor19 explained how they “sacrifice some posts” by supplementing them with a request for readers to add their opinion, in an effort to make them comment. “When we do it, it automatically cuts our traffic to the item,” but as a general strategy it is worthwhile because the page gets more visibility. This common practice nudges users not to read the articles—contrary to the normative role of journalism and consistent with social media logic (van Dijck and Poell 2013). Hence, in a somewhat ironic twist, the tool that was supposed to help gain insight into users’ preferences and to foster more reciprocal communication is used by editors to manipulate users so as to conquer the algorithm. While previously the platforms were described as an intermediator between news organizations and their audience (van Dijck and Poell 2013), here a more complex pattern emerges: the platforms stand between news organization and users, but users also “stand” between the news organization and the platform.

Last but not least, it should be noted that the influence of editors’ perceptions of users does not stop at the granular level of specific engagement metrics. Since editors are quite realistic in appraising the social arena in which users operate, on realizing that news engagement sometimes has little to do with news itself, they adapt by developing new practices which ensure revenues. They do this by deploying editorial tactics and technological features that will nudge users to engage with news via visible social media plug-ins, which will increase traffic and performance. For example, Editor16 deplores users’ tendency to share screenshots instead of links: “I understand them. First of all, in this way the person becomes the news organization. He sends it, it comes from him no matter what the title says. Second, it spares people everything they hate: ads, waiting for the content to upload—torture. They want it now. From the perspective of the news organization, it’s clearly unfair. Millions of people work on it, this is the economic model [traffic] and you [the user] don’t want to do it so you take a screenshot and take it out? Who will pay the reporters’ salary? But if I were a user, I would do the same probably. If I want to send something to my mom, she doesn’t know how to open a link but she knows how to open a picture.” To deal with this practice, editors yet again try to accommodate through editorial choices: “When you use a mind-blowing title, users don’t need to go to the article, it’s enough to take a screenshot and send it, because the headline tells you everything and it is easier than going to the article and sharing the link. This way you have to create added value [in the article itself] so people will have to share and not rely on screenshots” (Editor4). This explanation as to why an article needs to include substantive materials has nothing to do with journalistic practices and norms: rather, it revolves around getting users to act as prosumers, distributing content in the most effective way. Alternatively, editors accommodate through technological solutions. One website has a very successful feature where journalists’ tweets are presented in a news ticker format. The editors discovered that users take screenshots of this section, which for them was counterproductive. They therefore decided to develop a button specifically to share the content.

Another example for the use of technological solutions comes from Editor8, a chief editor: “Since sharing is about image and not about reading the story, I moved the sharing buttons (e.g., social plug-ins) outside—from the story webpage to the homepage or even to the image that goes with the story. [The rationale is] to save people the aggravation [authors’ emphasis] of having to click on a story to be able to share it—a technicality that doubled the traffic.” Together, these examples represent the way in which news organizations can use social media logic to promote their own interests, thus sacrificing their normative mission of informing their readers.

To conclude, just as users have evolved news engagement practices to help them manage social media logic, so have social media editors. Editors’ practices are based on a merging of two imagined entities: audiences and algorithms. Editors are fully aware of users’ behavior and motivations, and their understanding thereof goes beyond a simple reliance on “big data” comprising audience metrics (Nelson and Webster 2016). Their knowledge, however, is not put to the service of journalistic norms, nor does it serve purely to pander to popular tastes and give the audience what it wants. Rather, it is mostly harnessed to satisfy the algorithm (Tandoc and Maitra 2017).

Conclusions

Very early on, as it developed online, engagement was marked as the currency of the attention economy (Gerlitz and Helmond 2013). Like other economies, it gave rise to a more complex exchange rate. Web 1.0 users are said to have lived in a link economy and early Web 2.0 users—in a Like economy (Gerlitz and Helmond 2013). Today’s social media attention economy, with its new services and features, heralds a new stage, with visibility and self-presentation at the forefront (Picone 2015; Wilhelm, Stehle, and Detel 2021). This change is epitomized in User3’s reflection: “In the early days of Facebook, I was stingy with my Likes. I saw them as something that defines you. What you Like is who you are. But gradually, as Likes became more and more popular and widespread, and with the introduction of emoticons, this has changed.”

While Likes and links were currencies that could be translated into monetary value since they control traffic (Gerlitz and Helmond 2013; Nelson and Webster 2016), the value of social plug-ins measured by social media engagement metrics is more elusive. Granted, they too influence traffic, but how, exactly, is still unknown. This ambiguity underscores the importance of the sense-making process described here, as it triggers an array of editorial practices that shape news distribution.

The findings of this study are revealing in several ways. First, while past research focused on divergences between news professionals and users (Heise et al. 2014; Schmidt and Loosen 2015), the present study found the perceptions of these two stakeholders mostly to align. Both editors and users described the same motivations and the same practices when it comes to news sharing. This convergence could perhaps be attributed to the ubiquity of social media—a space where all social media editors are users as well, and probably participate in the same impression-management practices (Picone 2015). “I’m also a user,” said an editor when describing how she falls into traps editors set for users so as to generate engagement. Recall also the editor quoted above who spoke about his mother and how he shares news with her. Editors also demonstrate a good grasp of how social media engagement metrics differ from other types of web-analytics, and the important role visibility plays therein (Wilhelm, Stehle, and Detel 2021). Taken together, the portrayal of editors that emerges from this study’s findings is as of professionals who fully understand both their users and the world in which they operate. The discrepancy between the findings here and those in Schmidt and Loosen (2015) or in Heise et al. (2014) could have three reasons. First, the current study centered on social media, while previous research investigated an array of tools, including emails, forums, and more. Second, as professionals, this study interviewed only social media editors. It could be that focusing on the type of professional who is in constant interaction with users (Ferrer-Conill and Tandoc 2018) underscored the part of the newsroom that sets store by an alignment between users and editors. Lastly, the current investigation was carried out considerably more recently than the previous ones, and in the meantime social media editors specialized and became more professional (Ferrer-Conill and Tandoc 2018). Either way, the difference between the current and previous findings highlights the importance of the comparative approach adopted here, as well as the dearth of such research, especially when users’ perceptions are concerned.

Another insight afforded by the findings of the current study, to the effect that users’ and editors’ perceptions largely coincide, relates to editors’ practices. Nelson (2021) pointed out, with reference to all types of analytics, that “uncertainty (around the use of metrics) stems from the fact that even sophisticated measures of audience behavior paint an incomplete portrait of who the audience is and what they want from news media.” In the case of social media engagement metrics, it seems that this uncertainty has less to do with users, and more with the platform. So long as algorithmic curation remains a mystery (van Dijck and Poell 2013), both users and editors will be trying to adapt to it and to gain control over the news feed. Taken together, the practices of users and editors reveal the same organizing principle, whereby newsworthiness, informative value, and journalistic norms are secondary to social media norms and logic. News organizations demonstrate a growing dependency on social media and its logic, which shapes their modus operandi (Tandoc and Vos 2015). The value of these platforms is grounded in the assumption that posts will translate into traffic and revenues. Yet to realize this model, organizations rely on “prosumerism,” which does not always accord with users’ practices. First, users’ reluctance to share publicly, together with their opting for private channels (Meijer and Kormelink 2020; Swart, Peters, and Broersma 2018), impedes the social metrics which accompany news posts, resulting in less traffic. Second, users’ reliance on non-clickable news information—such as comments, including summaries or screenshots of news stories—diminishes traffic to news sites. These behaviors jeopardize a major source of traffic to news organizations and somehow bring us back full circle. Thus, before Web 2.0, when news was discussed and shared via social ties, news organizations could expect little revenue from such behavior. Social media changed the situation by distributing actual articles and not only discussions about them. Nevertheless, users are now adapting to these changes in a manner that (financially) mirrors their former habits.

To fight this trend, editors adopt tactics that ensure engagement and better performance, thereby making another step in “tweaking” (Tandoc and Maitra 2017) journalistic work to fit Facebook algorithms. While past research showed how editors tweak their posts, the present study indicates that editors tweak the users. To the extent that algorithmic logic is based on a feedback loop, and since editors want reach, users are utilized to influence the algorithm, which in turn reaches more users, in a causal nexus. Consequently, once reached, the public is expected to engage with content to ensure further dissemination, and should be nudged to do so, even at the expense of actual news consumption. Thus, the study demonstrates not only how the duality of social media engagement metrics sets them apart from other types of analytics but also how this duality transforms the relations between news organizations and users.

Finally, the findings point to a sharp contrast between journalists and news organizations in how they manifest relationships and reciprocity (Heise et al. 2014; Lawrence, Radcliffe, and Schmidt 2018). The study found no evidence of direct communication based on social media engagement metrics. This was not because social plug-ins are not viewed as a communication tool, but rather because most communication was directed elsewhere: to other users, to the algorithm, and lastly to journalists’ accounts. Past studies found that some journalists promote their organizations alongside themselves (Molyneux, Lewis, and Holton 2019). The split in users’ perceptions of organizations versus journalists suggests that, even if they do so, this does not necessarily mean that users will follow their lead. Instead, users’ perceptions testify to journalists’ growing separation from their organization (Hedman 2020; Molyneux, Lewis, and Holton 2019). This split also demarcates the current boundaries of the “relational shift” in journalism studies (Heise et al. 2014). As journalists increasingly establish themselves as “brands” in their own right (Olausson 2018), scholars should explore the implications of this process not only for the relationship between journalists and their audience but also for the relationship between news organizations and their audiences. Further research is needed to identify whether the flourishing of the journalist-audience relationship on social media comes at the expense of news organizations.

Disclosure Statement

No potential conflict of interest was reported by the author.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors received funding support from European Research Council starting grant #680009.

Notes on contributors

Shira Dvir-Gvirsman

Shira Dvir-Gvirsman (Ph.D. Hebrew University, 2011) is an associate professor at the Dan Department of Communication, Tel Aviv University. Her fields of interest are: news consumption practices and their implications on political behavior.

Notes

1 With the exception of the information that the news organization decides to share with their audience – for instance, the most viewed articles.

2 Twitter is mostly used by journalist to communicate with other journalists. Only 12% of the Israeli public use it, in comparison to the 74% who use Facebook. WhatsApp, the leading messaging App in Israel, was mentioned by many of the users interviewed. However, since it does not offer engagement metrics it falls outside of the scope of the current study.

3 The Israel news market includes: 3 broadcast TV stations and 1 public TV station; 8 public radio stations and 12 local and 2 military-owned radio stations; 8 legacy media newspapers which also have websites; 2 major news websites, and 5 small niche news websites. We contacted social media editors from all national commercial news organization.