Journalism is twerking? How web analytics is changing the process of gatekeeping

Abstract

New communication technologies have allowed not only new ways in which the audience interacts with the news but also new ways in which journalists can monitor online audience behavior. Through new audience information systems, such as web analytics, the influence of the audience on the news construction process is increasing. This occurs as the journalistic field tries to survive a shrinking audience for news. In this study, I argue that how journalists conceive of the audience as a form of capital influences the extent to which journalists integrate audience feedback from web analytics in their news work. I developed this theoretical framework through case studies of three online newsrooms that included a total of 150 hours of observations and 30 respondent interviews. The findings showed the extent of influence of web analytics on traditional gatekeeping processes and on a new gatekeeping practice online, which I call the process of de-selection.

Keywords

Audience case study field theory gatekeeping journalism journalists online news twerking web analytics

Introduction

Grinding and twerking in her skimpy flesh-toned latex underwear, wagging her tongue most of the time, 20-year-old singer Miley Cyrus shocked a lot of people with her performance at the 2013 MTV Video Music Awards. The next morning, news giant CNN shocked media critics. Not only did it upload a story about the stunt, but CNN.com also used the article as the main story on its homepage (Hare, 2013). In an online article explaining the decision, the website’s managing editor Meredith Artley said, “It was an attempt to get you to click on CNN.com so that we could drive up our web traffic, which in turn would allow us to increase our advertising revenue” (The Onion, 2013). Except that it was not really Artley who wrote the piece.

It was published by satirical news outlet The Onion, which was trying to make fun of CNN’s editorial decision while also trying to make a point (Phillips, 2013). Not everyone laughed at CNN, however, for within the journalistic field, a field that is seeing a continually declining newspaper readership, many journalists understand that drawing traffic to news sites is becoming more and more crucial. An editorial from Variety pointed out that the decision was not as easy as The Onion made it sound (Wallenstein, 2013). “It’s difficult to subsist on substantive journalism,” the Variety editorial said, “without some help from more crowd-pleasing content.” In short, journalism needs the audience.

News is meant to be read. “The audience is the ultimate consumer of the media product—the end of the news process” (Shoemaker and Reese, 2013: 177). Therefore, in understanding the construction of news, one cannot divorce journalism from its audience.

The onset of online news is changing not only how users interact with the news, for it has also provided journalists with a new way of learning more about their audiences (Napoli, 2011). Through web analytics, news organizations can collect and analyze the footprints that news users leave behind, offering immediate access to an unprecedented wealth of information about audience behavior. This information, in the form of audience metrics, guides subsequent editorial decisions (Lee et al., 2012; MacGregor, 2007).

Online news has swung right into the center of the journalistic field (Domingo, 2011), and the question now is not whether journalism has changed, but how it is changing. This study is interested in understanding how journalism is changing, particularly with the increasing influence from a knowable and quantifiable audience. Specifically, this study focuses on how web analytics, a new mechanism for audience feedback, is changing journalistic norms and routines.

Literature review

The relationship between journalists and their audiences was initially assumed to exert a relatively weaker influence on news work compared with other sources of influence (Flegel and Chaffee, 1971; Gans, 1979; Shoemaker and Reese, 1996). This, however, no longer reflects the changing news media ecosystem. In most online newsrooms, audience preferences have become omnipresent through web analytics and ignoring the online audience has become difficult.

Gatekeeping theory

Gatekeeping is “the process of selecting, writing, editing, positioning, scheduling, repeating, and otherwise massaging information to become news” (Shoemaker et al., 2008: 73). It describes the news selection process that operates under several layers of influences—from the individual to the routine, organizational, social institution, and social system levels (Shoemaker and Vos, 2009). The audience, as a social institution, can exert direct effects on news construction. For example, Herman and Chomsky (2002) referred to “flak” as a filter that shapes the news. Flak refers to negative feedback from the audience, such as criticisms or threats of boycotts—a way of “disciplining” the media (Herman and Chomsky, 2002). But the audience can also have indirect influences, as when journalists are oriented to providing what they think the audience wants, even if this prediction does not match actual audience preferences. This orientation to the audience is an example of a routine-level influence (Shoemaker and Vos, 2009).

Gatekeeping provides a holistic view of the news selection process, but it does not explicitly categorize the gates that news items pass through. Thus, other scholars have offered schemes to classify specific gates in news construction. For example, Bruns (2005) proposed looking at input, output, and audience response gates. In studying participatory journalism, Domingo et al. (2008) proposed studying audience participation in at least five stages of news construction: access/observation, selection/filtering, processing/editing, distribution, and interpretation. I am adapting this classification to study how web analytics impacts each stage of news construction.

Gatekeeping theory outlines the potential influences on news construction, but it also does not theorize a mechanism through which influence comes about. Because of this, a general impression from previous theorizing is that the media are always susceptible to influences, but journalists are also capable of resisting them. Sociologists have long stressed the importance of elaborating on mechanisms (Elster, 1989; Mahoney, 2003; Tilly, 2001). Stinchcombe (1991) argued that mechanisms increase “the suppleness, precision, complexity, elegance, or believability” of a theory (p. 367). In proposing a mechanism of influence on gatekeeping, I turn to field theory.

Field theory

Field theory seeks to study the journalistic field as a whole by accounting for not only the relationships between agents within the field but also the relationship of the journalistic field as a whole with other fields, such as the political and the economic fields (Benson, 2004). The framework of field theory builds on four key concepts: field, capital, doxa, and habitus. A field is a “field of forces” which agents struggle to either transform or preserve (Bourdieu, 1985, 2005). The journalistic field is an example (Benson, 2006; Siapera and Spyridou, 2012). It is a “microcosm with its own laws, defined by its own position in the world at large and by the attractions and repulsions to which it is subject from other such microcosms” (Bourdieu, 1998: 39). A field differentiates itself from other fields through an endemic system of rules that govern its own game. This is called the doxa, which refers to a “‘universe of tacit presuppositions’ that organize action within the field” (Benson and Neveu, 2005: 3). It refers to the rules of the game. News values are examples of journalistic doxa—an enduring set of criteria of what makes something newsworthy (Willig, 2013). Norms dominate in part because of the individuals who perpetuate them. So while the field is differentiated by its own set of rules, the rules coalesce through the agents that enforce the rules. This link between the field and its agents, between the micro- and the macro-levels, is provided by the concept of habitus. This refers to dispositions, accumulated through an agent’s experience over time, that generate “practices and perceptions” (Johnson, 1993: 5). In the journalistic field, the habitus “implies understanding the journalistic game” (Willig, 2013: 8). It refers to one’s historical trajectory, a collection of personal and professional experiences accumulated from social positions that produces knowledge and understanding of the game. Finally, capital refers to “the specific forms of agency and prestige within a given field” (Sterne, 2003: 375). It refers to various forms of resources, some more dominant than others. Forms of capital enable agents to participate in the struggle within the field (Handley and Rutigliano, 2012). Thus, capital is also something that agents seek. Field theory refers to two main forms of capital: economic and cultural (Benson and Neveu, 2005), but other scholars also refer to social and symbolic capital (Sallaz and Zavisca, 2007; Willig, 2013).

Considered the dominant capital in most fields, economic capital refers to money or assets transformable into money (Benson, 2006). In journalism, it has been operationalized in terms of advertising revenues, circulation rates, and audience size (Benson, 2004; Benson and Neveu, 2005; Siapera and Spyridou, 2012). Cultural capital refers to possession of competence in a socially valued area (Sallaz and Zavisca, 2007), often operationalized as possession of journalistic excellence or quality as conferred by professional or academic groups, such as the Pulitzer Prize (Benson and Neveu, 2005). However, other scholars defined it in terms of accumulated knowledge that can come in embodied, objectified, or institutionalized forms (Bourdieu, 1986; Siapera and Spyridou, 2012). Social capital refers to “the sum of all those people one knows or groups to which one belongs” (Siapera and Spyridou, 2012: 81). A journalist’s social media network is a form of social capital. Finally, symbolic capital refers to prestige or status which legitimates dominance (Johnson, 1993; Siapera and Spyridou, 2012) that comes in the form of credibility or good reputation (Bourdieu, 1986). In this study, I will argue that Bourdieu’s discussion of forms of capital is an example of mechanistic explanation that could be adapted in understanding influences on gatekeepers.

A mechanism of influence

In proposing a mechanism of influence, I will argue that journalists, as rational agents (Fengler and Russ-Mohl, 2008), realize the required capital they need to participate in the eternal struggle within the field. They become susceptible to influences from various levels when they experience or perceive some form of instability in their capital accumulation. Instability can refer to either a perceived increase or decrease in capital. Thus, when journalists perceive some instability in their capital—economic, cultural, social, or symbolic—they become susceptible to influences as they seek for capital stability. How a journalist responds to perceived instability depends on one’s habitus and the field’s doxa. For example, a survey of online editors found that perception of uncertainty predicted monitoring of audience metrics (Lowrey and Woo, 2010; Figure 1).

Figure 1.

The mechanism of audience influence.

In media studies, audiences have been considered as both consumers of news content and commodities sold to advertisers (Butsch, 2011; Loosen and Schmidt, 2012; Webster and Phalen, 1994). This is consistent with the conceptualization of economic capital. But audiences are also essential in understanding the normative roles of journalism (Siebert et al., 1963). For example, the press is considered the fourth estate, consistent with the ideal of serving the public (Schultz, 1998). In journalistic role conception literature, roles are oriented toward audiences (Donsbach, 2008; Tandoc et al., 2012; Weaver et al., 2007; Weaver and Wilhoit, 1986, 1996). An informed audience speaks of the kind of journalism a news organization produces, a form of prestige that confers legitimacy. The audience, therefore, can also be conceived of as a form of cultural, social, and symbolic capital. How journalists conceive of the audience as a form of capital will also influence how they respond to any perceived capital instability.

Web analytics in gatekeeping

Journalists used to ignore, if not reject, feedback from the audience (Beam, 1995; Gans, 1979; Schlesinger, 1978). They substituted their own preferences for those of their actual audience. Feedback used to come in the form of letters to the editors or phone calls to the newsroom considered unrepresentative of the actual audience (Schlesinger, 1978). As years passed, and even with newer audience measurement systems such as readership surveys and television ratings, journalists still relied on their supervisors, peers, relatives, and personal preferences for their news judgment (Beam, 1995; Sumpter, 2000).

The literature offers four plausible reasons for this institutional rejection of audience feedback. First, journalists mostly dislike numbers and statistics (Curtin and Maier, 2001; Gans, 1979), the form audience research most often takes. Second, journalists and audiences have different preferences (Boczkowski, 2010; Boczkowski et al., 2011; Mitchelstein and Boczkowski, 2010; Shoemaker and Cohen, 2006). Third, a vague notion of who the audience was, as well as the divergence in the preferences of journalists and their perceived audiences, could have led many journalists to have a pejorative view of the audience (Atkin et al., 1983; Schlesinger, 1978). Finally, a fundamental explanation to rejection of audience feedback among journalists lies in the journalistic norm of autonomy. Beam (1995) found that journalists feared that incorporating results of readership surveys in editorial operations might erode journalistic quality. “The conflict between researchers and journalists is over the priority of commercial versus professional considerations in story selection” (Gans, 1979: 233).

But new information technologies are offering new ways of interaction between audiences and messages. Newsrooms in the United States have embraced web analytics, and editors can now monitor how users behave online based on audience metrics (Lowrey and Woo, 2010). Web analytics programs provide a range of audience metrics, such as number of page views, amount of viewing time, and site referrals, among others (Kaushik, 2010; Napoli, 2011). Having access to metrics allows website developers to decide beyond their intuition (Duncan, 2010). In journalism, metrics now guide editors in their decisions of where to place stories on the website (Anderson, 2011a, 2011b; Lee et al., 2012). A web-based survey of 318 newspaper editors in the United States found that 84% monitored traffic regularly and 31% used audience metrics “to plan content production” (Vu, 2013: 10). Interviews with 19 online journalists found that “Some admit now that they double-check their instinctive guesses with tracking data” (MacGregor, 2007: 294). A content analysis of three online newspapers in the United States and their lists of most viewed stories found that audience clicks, measured by most viewed stories, affected subsequent placement of stories in the homepage (Lee et al., 2012). Anderson (2011a) found that journalists “now expect reader feedback, even if they do not like it, agree with it, or see it as enhancing their ultimate journalistic product” (p. 558).

But journalists also remain loyal to norms that have dominated the journalistic field. MacGregor (2007) found that while journalists started using web analytics, many of them still did not want metrics to get in the way of traditional news values in guiding story selection. Organizational structures, work practices, and representation of audiences, among other factors, can influence how technological adoption will impact news organizations (Boczkowski, 2004, 2005). In this study, I am focusing on how audience feedback impacts news work. This depends on how journalists conceptualize the audience as a form of capital. Thus,

Research question 1 (RQ1). How do journalists use web analytics in their news work?

RQ2. What factors influence how journalists use web analytics in their news work?

Methodology

Case study is a methodology with an epistemology that emphasizes the need to understand cases holistically instead of dissecting a case into decontextualized segments (Creswell, 2007; Yin, 2009). It relies on a range of methods to achieve the purpose of understanding a case or cases in depth (Eisenhardt, 1989). Building on studies of journalistic uses of web analytics, mostly based on surveys, this study hopes to contribute to this research area using the methodology of case study.

This study involves three cases—instances of the phenomenon being studied—following a replication logic (Eisenhardt, 1989; Yin, 2009). Three online newsrooms that used web analytics were selected. These three cases are considered theoretical units of the same phenomenon. They all use a third-party web analytics program for real-time monitoring, a typical set-up for news sites in the United States. Studying multiple cases yields a “more robust, generalizable and testable theory” compared with single case studies (Eisenhardt and Graebner, 2007: 27).

The websites of the three online newsrooms included in this study are among the 50 biggest news sites in the United States based on traffic (the first newsroom’s website gets about 3 million unique monthly visitors, the second newsroom about 3.9 million, and the third newsroom about 3.8 million). These newsrooms all have newspaper products, also among the 50 biggest newspapers in the United States based on circulation. Thus, the online desks are located in newsrooms converged with the editorial desks of their respective newspapers.

Research methods

I used participant observation and interview methods. The use of ethnographic methods is appropriate to the goals and assumptions of multiple case study, an argument for referring to this approach as an ethnographic case study (see, for example, Boczkowski, 2005). Paterson (2008) argued that only ethnographic methodologies “can come close to providing an adequate description of the culture and practice of media production, and the mindset of media producers” (p. 2).

Participant observation

I assumed the observer-as-participant role (Lindlof and Taylor, 2010), taking notes during my observations and establishing good rapport with many of the editors (Tracy, 2013). I attended daily budget meetings and observed online editors do their work—checking emails, writing and editing stories and photo galleries, tweeting, using Facebook, taking phone calls, and talking to other journalists. I focused only on the online news process and sat in the online news desks, jotting down what editors did and said that were related to the online news operations. My observations across the three newsrooms took 150 hours (110 hours in the first, 10 hours in the second, and 30 hours in the third), spread out in the months of March and May 2013.

Respondent interviews

I used ethnographic interviews, or informal conversations during the course of observations, and respondent interviews, or sitting down with an editor at a specified time usually after work hours (Tracy, 2013). I also implemented semi-structured interviews: I was guided by a set of prepared questions that explored how journalists encountered and responded to audience feedback at each stage of their daily news work, but I also asked follow-up questions, particularly on relevant aspects of work my interview guide had not anticipated (e.g. a recent capability of web analytics software to conduct headline testing). The interviews were digitally recorded. I interviewed a total of 30 journalists (15 from the first newsroom, 8 from the second, and 8 from the third) from different levels of the hierarchy, ranging from the publisher and editor-in-chief to the web editors, for a total of 668 minutes.

Data analysis

I used constant comparative approach, a common analytical strategy in grounded theory (e.g. Corbin and Strauss, 1990; Glaser, 1965) that has been used extensively in other studies that did not use this methodology (Creswell, 2007; Tracy, 2013). I also used NVivo, a computer-assisted qualitative data analysis software (CAQDAS), to organize and code the data in an iterative fashion, using concepts from my proposed theoretical framework (e.g. audience as economic capital) as well as using in vivo codes emerging from the data (e.g. click-bait) (see Tracy, 2013). This process yielded 135 codes for primary cycle coding. These codes were later lumped together under conceptual bins during secondary cycle coding (see Saldaña, 2009; Figure 2)

Figure 2.

Theoretical framework from case studies.

Results

The door swung open after long, silent minutes of people going in and out of the photo room speaking in hushed tones. A tall man, probably in his late 50s, emerged out of the room, his face flushed red, his eyes wet, and his voice cracking. “Thank you all,” the veteran journalist said, addressing a newsroom of about 30 people. Normal chatter ceased and gave way to a respectful applause. “Thank you,” he said after almost 2 minutes of applause. “This newsroom has been my family for 23 years.” Then, the newsroom veteran, a videographer per his most current job description, made an informal salute and left. A web editor—she had been here for about 6 months—was speechless, her jaw wide open. Another web editor, who also came in just a few months ago, covered his mouth with both of his hands. Layoffs have been ordinary in this newsroom in the last year or so. But nobody gets used to seeing people go.

Using web analytics

It is painful to see these layoffs happen, but older editors in the upper echelons of the newsroom are no longer surprised. Many of them started their careers at a time when newspapers flourished, and they rose to the ranks of the newsroom hierarchy as the industry continually declined. The newsroom’s deputy managing editor for digital finished college in 1996, a period when, she said, “everybody can get a job.” Then she and her contemporaries witnessed how everything changed: The company filed for bankruptcy in 2009. The new editor-in-chief, who has spent most of his life as a journalist, now finds himself having to make decisions on which among his staff members will go jobless next. “The landscape changed very quickly,” he said, reflecting on what he had witnessed across the years. “Lots of journalists lost their jobs. It meant smaller newsrooms had to do more with less … And it was kind of, you know, either change or die.”

This perception of instability is a driving force for the institutionalization of using web analytics in these three newsrooms. Since the new owners brought him in just a year ago, the first newsroom’s editor-in-chief has been trying to emphasize how the newsroom should think of itself as a digital company. This entails pursuing new ways of interacting with the audience, delivering content, and accessing data. In general, training on web analytics is not formalized in most newsrooms—either a young editor is expected to know it prior to working in the newsroom or an old editor being moved to digital is expected to learn the technology by using it. The editor-in-chief at the first newsroom only referred to his experience working in an online-only newsroom prior to him coming to his present newsroom. His managing editor said he learned the tricks in the newsroom, being personally interested in new technologies. A web editor said most of what she knew about web analytics was self-taught.

The managing editor of the first newsroom compared using web analytics with getting hooked on drugs. “It’s like crack,” he said, grinning. “You can sit here and watch it, popping all night.” Online editors would have their analytics program open all day, and the numbers on the analytics dashboards would change all the time. The metaphor is funny, but it also has deeper implications. The reason illegal drugs are outlawed is because a drugged person might pose danger to herself and to others as she loses control and becomes unable to function normally.

The first two newsrooms used both Omniture and Chartbeat, while the third newsroom used Omniture and Visual Revenue. These are all paid web analytics programs. Omniture is used to analyze long-term data, such as comparing this month’s total traffic with the same period last year. Chartbeat and Visual Revenue are used for day-to-day operations and real-time data. In response to RQ1, the following sections describe how journalists use web analytics at each stage of their news work (Domingo et al., 2008).

Access/observation

Online traffic is a sure way to get the attention of editors. Stories and topics that trend on social media, for example, always get noticed. In the budget meetings at the second and third newsrooms, the managing editors would always ask the social media manager or the assistant managing editor what topics were being discussed on social media. Stories that had gotten good traffic based on web analytics would get updates and follow-ups. Topics that have done well in the past also tend to get assigned more. One time, during the budget meeting at the second newsroom, an online editor proposed putting together a photo gallery of the then newly elected Pope Francis. “There’s a lot of interest over the weekend about the Pope being a ‘regular’ guy,” the editor said, adding that stories on the new Pope did well on Chartbeat. “It was really resonating over the weekend.”

Selection/filtering

The homepage is the prime space for a news site, and it is where homepage editors exercise a lot of judgment in terms of story selection and placement. Only a limited number of stories and story elements are placed on the homepage to make sure it does not get cluttered and readers can easily navigate the site. The website also needs to remain “fresh” with a regular supply of new stories. Readers are not expected to come back if they will just find the same set of stories on the homepage over and over again. When numbers start to dip, web editors take that as being the right time to update the homepage by moving around stories or adding new ones to replace those that have stagnated based on traffic. These judgments—selection and what can be described as de-selection—are often based on web metrics. One time, a web editor at the second newsroom took out a business story from the homepage to give way to a new sports story he was sure would get a lot of traffic. However, when he checked Chartbeat again, the business story he had taken out was actually the eighth most viewed story so far. He decided to put it back. Instead, he took out a story on a girl who was shot dead, saying, “If it doesn’t get any traction in 20–30 minutes, we usually pull it out.”

Sometimes, the recommendations for selections and de-selections are made by the algorithms in the software. For example, Visual Revenue provides placement recommendations with an estimated number of unique visitors that following each suggestion would bring to the site. A web editor talked about a story that quoted a football player saying “bad barbecue makes me want to fight.” The article was a blog post buried in the sports page and had gone unnoticed by the homepage editors. But then, out of the blue, Visual Revenue recommended putting it on the homepage. The web editor followed the recommendation, took an underperforming story out and replaced it with the “bad barbecue” story which ended up being the day’s most viewed story.

Processing/editing

An interesting use of web analytics comes at the stage of writing headlines. The web editors at the third newsroom had put words like “bra” and “vagina” on their headlines. One story was about the New York City police denying actress Amanda Bynes’ allegation that her arresting officers had slapped her privates. The web editors exchanged ideas on how to write the headline, until one of them decided to use the word vagina on the headline. The story attracted a lot of traffic as expected and was turning up in search engines. That was the first thing they checked after uploading the story—traffic. Sometimes, when a web editor would be dilly-dallying on how to write the headline, he would just use Visual Revenue’s headline testing service. One time, a web editor wrote two headlines for the same story. Visual Revenue randomly exposed readers to one of the two headlines. His first version got a 9% rating while his second version got 42%. The ratings were based on how many readers clicked on the story after reading the headline. He decided to go with the second headline based on Visual Revenue’s recommendation.

Web metrics also affect decisions on accompanying photos, videos, and graphics. For example, at the second newsroom, a web editor was updating a news story about a homicide when he suddenly came up with an idea. Having the suspect’s photo next to a story is common, but readers would also want to see the photo of the victim. So he thought of putting the two photos side by side. “I kind of favor that,” he said. “We need her face [the victim’s] out there for it to perform well so I’m gonna talk to photo.” Then, he walked toward the photo section, came back to his workstation, and said, “So they’re producing that.”

Distribution

Lists of most popular stories, based on automatically generated audience metrics, have become staples in news sites. This is one way that web analytics affects the distribution stage. However, the first newsroom had replaced its list of most popular stories with a list of editor’s picks. There were two plausible reasons for this. First, lists of most popular stories tend to be static: an initial list functions as a guide to other readers who would view the articles listed, hence keeping them in the most popular list. A static list, however, does not encourage repeat visits. Second, relying on algorithms that populate the list based on page views takes the control out of the journalist. But another platform for content promotion and distribution is social media, as journalists take advantage of the free third-party platforms to drive traffic to their news sites. Although editors use analytics to determine popular stories, the same metrics also show them the underperforming stories. Knowing which stories are not getting a lot of traffic leaves an editor with two choices: take it out of the homepage and replace it with a click-bait, or keep it on the homepage, especially if it is something editorially determined as important and help it gain more traffic by placing it more prominently on the homepage, updating it, adding story assets to it, or by promoting it on social media. Most of the time, however, promotion on social media was done to drive traffic to the site, so that stories that are expected to be click-baits are the ones promoted most often.

Interpretation

This stage refers to user participation in making sense of the news, usually in the form of readers’ comments (Domingo et al., 2008). However, some editors in the observed newsrooms did not have a flattering view of the commenters on their sites. “I don’t know if you’ve ever bothered to read the comments sections on any of our stories on the main site,” a web editor said at the first newsroom. “It will make you sad for humanity—just the worst of the worst are usually who comment.” Instead, most web editors pay attention to web analytics for how the audience responds to content—a quantifiable measure of popularity equated with audience preference. Interpretation is reduced in terms of clicks and page views. What the audience wants is now understood through which stories attract traffic—which is easier to comprehend than going through readers’ comments. Traffic is also equated with a job well done. An editor who phoned in during a budget meeting at the first newsroom was talking about a video she had posted on the site when the managing editor told her, “Did you see my note on that? The video did very well.” He meant the video got good traffic.

Factors affecting use

In response to RQ2, this section explains the factors that affect how journalists use web analytics in their news work. Aside from the perception of instability within the journalistic field that is leading newsrooms to embrace new communication technologies, such as web analytics, another recurring factor in the interviews is the importance of the audience. The bottom line is that the online audience translates into advertising revenues that translate into paychecks. “We write stories to engage people, to inform people—the audience is critical. These are the people who pay the bill,” the metro editor at the first newsroom said. The deputy editor for digital at the third newsroom also stressed the importance of the audience:

Edson:

How important is the audience?

Editor:

It’s everything.

Edson:

Could you elaborate on that?

Editor:

If nobody is looking at the site, we’re screwed.

Edson:

Why are these numbers important?

Editor:

That’s revenue. That’s money.

It is the same business model that sustained the now seemingly unsustainable newspaper industry. What is different now is how journalists in these online newsrooms can no longer ignore the audience. It used to be that news editors did not have to think about how many people are attending to their news content, consistent with the wall of separation between the editorial department that protected its journalistic autonomy and the business department that took care of audience size and revenue. But things have changed. I asked one of the web editors at the third newsroom how he felt about what journalism critics claim to be a dichotomy between quality journalism and the goal to draw web traffic. He said, “I don’t think that we have the luxury of thinking that way because if the company’s not making money then I might get laid off. I mean, that’s just the way it is.”

The reality, however, is that while online revenue is slowly increasing, it still pales into comparison with the declining traditional media revenues. The newspaper remains to be the bigger earner in these newsrooms, and yet, in all of them there is a great push for going digital with the hope that digital revenue will, in the immediate future, be able to sustain journalism. This can explain why the editors conceived of the online audience as a form of economic capital even if online revenues remain marginal in terms of revenue generation. But this conceptualization has to be balanced with the prevailing norms in the field, or the journalistic doxa. Therefore, the integration of web analytics in the news routine is largely a constant balancing act. On one hand, journalists believe it is their role to provide the public with the information it needs, part of the journalistic doxa. On the other hand, news organizations need to attract an audience to survive. “I think our value to readers still is telling them what is news and what should be news, what are the important stories,” the editor-in-chief at the first newsroom said. “But we have to listen to the audience more than we have in the past.”

“I think now you have to sit down on the table and have an honest conversation about what the goals are,” the deputy managing editor for digital at the first newsroom said, “and knowing that having the money and being a respected journalist, organization, you know, you have to strike a balance.” However, most of the time, the balancing act does not lead to the ideal, and often, it tips toward the goal of increasing traffic by using web analytics to come up with click-bait stories. “Sometimes you have to hold your nose,” a web editor from the third newsroom said. Holding your nose is in keeping with the reality that news sites need to generate traffic. It is an unspoken rule at the first and second newsrooms; it is an explicit instruction at the third newsroom, where the publisher sets a goal of increasing traffic by 10% from last year.

Having to attend to multiple tasks that come with the territory of being digital, web editors do not have the luxury of time to pause and consciously reflect on every editorial decision they make. The digital editor at the first newsroom is tasked to edit and upload stories to the website—but she also had to tweet breaking stories, post on the newsroom’s Facebook account, and attend to questions on social media from editors and reporters while monitoring traffic. The other web editors also experience having to wear too many hats, as digital desks continue to make sense of how the workflow should look like as news organizations embrace the digital-first mantra. In a newsroom where staff sizes continue to shrink, roles continue to evolve, and tasks continue to pile up, journalists do not have the time to engage in a personal reflection to match their belief of a need to balance editorial autonomy and audience influence with the specific day-to-day actions they undertake. Although in the interviews, web editors talked about how their news judgment reigns supreme in their news work, I never heard them during my observations refer to particular news elements when choosing which stories to upload and which stories to highlight. The only consideration routinely invoked is the metrics displayed in the web analytics software splashed on their computer monitors all day.

“You have the dessert but you also have the nutrient. It’s always like that; you have to have a balanced plate and hope it works,” a web editor at the third newsroom said. Then, he looked away and said, “I couldn’t consciously do that.”

Conclusion

The results offer empirical support to the mechanism of influence I proposed based on field theory’s concepts that can expand how we understand the gatekeeping process. Faced with the reality of declining economic capital for traditional journalism still unmatched by the slow increase in digital revenues, journalists clearly perceive capital instability within the journalistic field. This instability has opened up the gates to the influence of the audience, conceived of by journalists as largely a form of economic capital that can be maximized to preserve capital stability in the journalistic field. This mechanism explains why the editors used web analytics mainly to inform strategies to further increase web traffic. They used web analytics to monitor stories doing well, so these can be placed more prominently on the homepage and be promoted on social media to attract traffic. Backed by web metrics, editors predicted which topics, headlines, and story assets were more likely to generate traffic.

But aside from describing how journalists used web analytics and explaining potential factors that shape these usage patterns, this study also expands how we understand gatekeeping by pointing out a new process that comes after news selection. Gatekeeping explains how news gets selected (Shoemaker and Vos, 2009), but the pressure to keep the homepage updated to attract constant traffic is making the process of de-selection equally important. De-selection refers to deciding which among articles that have made it through the gates will be taken out of the homepage to be replaced by a new story. This process is largely determined by audience metrics. A dip on website traffic signals a need to replace elements on the homepage. An editor from the second newsroom even specified a particular window of time he gives a story. If it does not perform well within 20 minutes, the story will be replaced. This is motivated by the pressing goal of increasing traffic by giving the audience what it wants.

Journalists are not exactly known to be early adopters of technology, and studies showed that when they adopt new technologies, such as blogs and social media, they tend to normalize it, adapting them to fit into their existing norms and routines (Lasorsa et al., 2011; Singer, 2005). Indeed, journalists are normalizing web analytics, using audience metrics to inform their traditional gatekeeping functions. But a process of negotiation is also going on, with journalists modifying existing norms to accommodate the increasing influence of new technology. News judgment now includes acute awareness of what stories did well in the past based on traffic. Headlines are now being tested in terms of which version attracts more clicks.

This study is not without limitations. The decision to focus on three online newsrooms is based on the goal of theoretical replication and practicality of access. But future studies should also look at how other types of newsrooms use web analytics and with what results. The results cannot be generalized to all journalists in the United States, but the study hopes to provide an important glimpse into how journalism is changing because of quantifiable and immediate feedback from the audience through in-depth conversations with, and close observations of, journalists who use analytics in their day-to-day news work.

The Online Oxford Dictionary (2013) defines “twerk” as an informal verb that means to “dance to popular music in a sexually provocative manner involving thrusting hip movements and a low, squatting stance.” The journalistic field, threatened by its shrinking economic capital, is dancing to the popular music of consumer-driven logic, for it appears—at least so far—that this is the only way to survive. Still dependent on an advertising-driven model, online journalism finds itself having to chase online traffic, a routine made possible and further enabled by web analytics. In order to attract an audience no longer loyal to legacy news, journalism dances in a provocative manner—publishing stories about the wildest celebrities, uploading adorable cat videos, highlighting salacious headlines—hoping to attract attention, to increase traffic. For media critics, this is a low, almost squatting stance, for an institution that relies a lot on respect and reputation. For a few others, this is journalism trying to survive. Journalism—to some extent—is twerking.

Footnotes

Acknowledgements

I wish to thank the three anonymous reviewers for their very helpful suggestions that greatly improved this manuscript.

Funding

This work was also supported by the Center for the Digital Globe at the University of Missouri.

Author biography

Edson C Tandoc Jr is an Assistant Professor at the Wee Kim Wee School of Communication and Information at the Nanyang Technological University in Singapore. His research focuses on the sociology of message construction. He has conducted studies on the construction of news and social media messages. His studies about influences on journalists have focused on the impact of journalistic roles and audience feedback on the various stages of the news gatekeeping process. Prior to his graduate studies, he worked as a print journalist in the Philippines for 6 years.

References

Anderson

(2011a) Between creative and quantified audiences: web metrics and changing patterns of newswork in local US newsrooms. Journalism 12: 550–566.

Anderson

(2011b) Deliberative, agonistic, and algorithmic audiences: journalism’s vision of its public in an age of audience transparency. International Journal of Communication 5: 529–547.

Atkin

Burgoon

(1983) How journalists perceive the reading audience. Newspaper Research Journal 4: 51–63.

Beam

(1995) How newspapers use readership research. Newspaper Research Journal 16: 28–38.

Benson

(2004) Bringing the sociology of media back in. Political Communication 21: 275–292.

Benson

(2006) News media as a “journalistic field”: what Bourdieu adds to new institutionalism, and vice versa. Political Communication 23: 187–202.

Benson

Neveu

(2005) Introduction: field theory as a work in progress. In: Benson

Neveu

(eds) Bourdieu and the Journalistic Field. Malden, MA: Polity Press, pp. 1–28.

Boczkowski

(2004) The processes of adopting multimedia and interactivity in three online newsrooms. Journal of Communication 54: 197–213.

Boczkowski

(2005) Digitizing the News: Innovation in Online Newspapers. Cambridge, MA: The MIT Press.

10.

Boczkowski

(2010) The divergent online news preferences of journalists and readers. Communications of the ACM 53: 24–26.

11.

Boczkowski

Mitchelstein

Walter

(2011) Convergence across divergence: understanding the gap in the online news choices of journalists and consumers in Western Europe and Latin America. Communication Research 38: 376–396.

12.

Bourdieu

(1985) The social space and the genesis of groups. Theory and Society 14: 723–744.

13.

Bourdieu

(1986) The forms of capital. In: Richardson

(ed.) Handbook of Theory and Research for the Sociology of Education. New York: Greenwood Publishing Group, pp. 241–258.

14.

Bourdieu

(1998) On Television. New York: The New Press.

15.

Bourdieu

(2005) The political field, the social science field and the journalistic field. In: Benson

Neveu

(eds) Bourdieu and the Journalistic Field. Malden, MA: Polity Press, pp. 29–47.

16.

Bruns

(2005) Gatewatching: Collaborative Online News Production. New York: Peter Lang Publishing, Inc.

17.

Butsch

(2011) Audiences and publics, media and public spheres. In: Nightingale

(ed.) The Handbook of Media Audiences. Malden, MA: John Wiley & Sons, pp. 149–168.

18.

Corbin

Strauss

(1990) Grounded theory research: procedures, canons, and evaluative criteria. Qualitative Sociology 13: 3–21.

19.

Creswell

(2007) Qualitative Inquiry and Research Design: Choosing among Five Approaches. Thousand Oaks, CA: SAGE.

20.

Curtin

Maier

(2001) Numbers in the newsroom: a qualitative examination of a quantitative challenge. Journalism & Mass Communication Quarterly 78: 720–738.

21.

Domingo

(2011) The centrality of online journalism today (and tomorrow). In: Domingo

Paterson

(eds) Making Online News. New York: Peter Lang Publishing, Inc., pp. xiii-xx.

22.

Domingo

Quandt

Heinonen

. (2008) Participatory journalism practices in the media and beyond. Journalism Practice 2: 326–342.

23.

Donsbach

(2008) Journalists’ role perceptions. In: Donsbach

(ed.) The International Encyclopedia of Communication. Malden, MA: Wiley-Blackwell, pp. 2605–2610.

24.

Duncan

(2010) Using Web Analytics to Measure the Impact of Earned Online Media on Business Outcomes: A Methodological Approach. San Francisco, CA: Institute for Public Relations.

25.

Eisenhardt

(1989) Building theories from case study research. Academy of Management Review 14: 532–550.

26.

Eisenhardt

Graebner

(2007) Theory building from cases: opportunities and challenges. Academy of Management Journal 50: 25–32.

27.

Elster

(1989) Nuts and Bolts for the Social Sciences. Cambridge: Cambridge University Press.

28.

Fengler

Russ-Mohl

(2008) The crumbling hidden wall: towards an economic theory of journalism. Kyklos 61: 520–542.

29.

Flegel

Chaffee

(1971) Influences of editors, readers, and personal opinions on reporters. Journalism & Mass Communication Quarterly 48: 645–651.

30.

Gans

(1979) Deciding What’s News. New York: Pantheon Books.

31.

Glaser

(1965) The constant comparative method of qualitative analysis. Social Problems 12: 436–445.

32.

Handley

Rutigliano

(2012) Journalistic field wars: defending and attacking the national narrative in a diversifying journalistic field. Media, Culture & Society 34: 744–760.

33.

Hare

(2013) Miley Cyrus twerks, stuns VMAs crowd. CNN. Available at: http://www.cnn.com/2013/08/26/showbiz/music/miley-cyrus-mtv-vmas-gaga/index.html (accessed 8 October 2013).

34.

Herman

Chomsky

(2002) Manufacturing Consent: The Political Economy of the Mass Media. New York: Knopf Doubleday Publishing Group.

35.

Johnson

(1993) Editor’s introduction: Pierre Bourdieu on art, literature and culture. In: Johnson

(ed.) The Field of Cultural Production: Essays on Art and Literature. New York: Columbia University Press, pp. 1–28.

36.

Kaushik

(2010) Web Analytics 2.0: The Art of Online Accountability and Science of Customer Centricity. Indianapolis, IN: Wiley Publishing, Inc.

37.

Lasorsa

Lewis

Holton

(2011) Normalizing Twitter: journalism practice in an emerging communication space. Journalism Studies 13: 19–36.

38.

Lee

Lewis

Powers

(2012) Audience clicks and news placement: a study of time-lagged influence in online journalism. Communication Research. Epub ahead of print 20 November. DOI: 10.1177/0093650212467031.

39.

Lindlof

Taylor

(2010) Qualitative Communication Research Methods. Thousand Oaks, CA: SAGE.

40.

Loosen

Schmidt

J-H

(2012) (Re-) discovering the audience. Information, Communication & Society 15: 867–887.

41.

Lowrey

Woo

(2010) The news organization in uncertain times: business or institution? Journalism & Mass Communication Quarterly 87: 41–61.

42.

MacGregor

(2007) Tracking the online audience. Journalism Studies 8: 280–298.

43.

Mahoney

(2003) Tentative answers to questions about causal mechanisms. American Political Science Association conference, Philadelphia, PA, 27 August.

44.

Mitchelstein

Boczkowski

(2010) Online news consumption research: an assessment of past work and an agenda for the future. New Media & Society 12: 1085–1102.

45.

Napoli

(2011) Audience Evolution: New Technologies and the Transformation of Media Audiences. New York: Columbia University Press.

46.

The Onion (2013) Let me explain why Miley Cyrus’ VMA performance was our top story this morning. The Onion. Available at: http://www.theonion.com/articles/let-me-explain-why-miley-cyrus-vma-performance-was,33632/ (accessed 8 October 2013).

47.

Online Oxford Dictionary (2013) Definition of twerk in English. Oxford Dictionaries. Available at: http://www.oxforddictionaries.com/definition/english/twerk

48.

Paterson

(2008) Introduction: why ethnography? In: Paterson

Domingo

(eds) Making Online News. New York: Peter Lang Publishing, Inc., pp. 1–14.

49.

Phillips

(2013) ‘The Onion’s’ CNN article draws response from network’s Meredith Artley over Miley Cyrus. Epoch Times. Available at: http://www.theepochtimes.com/n3/264928-the-onions-cnn-article-draws-response-from-networks-meredith-artley-over-miley-cyrus/ (accessed 8 October 2013).

50.

Saldaña

(2009) The Coding Manual for Qualitative Researchers. Thousand Oaks, CA: SAGE.

51.

Sallaz

Zavisca

(2007) Bourdieu in American sociology, 1980–2004. Annual Review of Sociology 33: 21–41.

52.

Schlesinger

(1978) Putting ‘Reality’ Together. Beverly Hills, CA: SAGE.

53.

Schultz

(1998) Reviving the Fourth Estate. Cambridge: Cambridge University Press.

54.

Shoemaker

Cohen

(2006) News around the World: Content, Practitioners, and the Public. New York: Routledge.

55.

Shoemaker

Reese

(1996) Mediating the Message: Theories of Influences on Mass Media Content. White Plains, NY: Longman.

56.

Shoemaker

Reese

(2013) Mediating the Message in the 21st Century: A Media Sociology Perspective. New York: Routledge.

57.

Shoemaker

Vos

(2009) Gatekeeping Theory. New York: Routledge.

58.

Shoemaker

Vos

Reese

(2008) Journalists as gatekeepers. In: Wahl-Jorgensen

Hanitzsch

(eds) Handbook of Journalism Studies. New York: Routledge, pp. 73–87.

59.

Siapera

Spyridou

L-P

(2012) The field of online journalism: a Bourdieusian analysis. In: Siapera

Veglis

(eds) The Handbook of Global Online Journalism. Malden, MA: John Wiley & Sons, Inc., pp. 77–98.

60.

Siebert

Peterson

Schramm

(1963) Four Theories of the Press; the Authoritarian, Libertarian, Social Responsibility, and Soviet Communist Concepts of What the Press Should Be and Do. Urbana, IL: University of Illinois Press.

61.

Singer

(2005) The political j-blogger. Journalism 6: 173–198.

62.

Sterne

(2003) Bourdieu, technique and technology. Cultural Studies 17: 367–389.

63.

Stinchcombe

(1991) The conditions of fruitfulness of theorizing about mechanisms in social science. Philosophy of the Social Sciences 21: 367–388.

64.

Sumpter

(2000) Daily newspaper editors’ audience construction routines: a case study. Critical Studies in Media Communication 17: 334–346.

65.

Tandoc

Hellmueller

Vos

(2012) Mind the gap: between role conception and role enactment. Journalism Practice 7: 539–554.

66.

Tilly

(2001) Mechanisms in political processes. Annual Review of Political Science 4: 21–41.

67.

Tracy

(2013) Qualitative Research Methods: Collecting Evidence, Crafting Analysis, Communicating Impact. Malden, MA: Wiley-Blackwell.

68.

(2013) The online audience as gatekeeper: the influence of reader metrics on news editorial selection. Journalism. Epub ahead of print 28 October. DOI: 10.1177/1464884913504259.

69.

Wallenstein

(2013) Why the Onion is wrong about CNN and Miley Cyrus. Variety. Available at: http://variety.com/2013/digital/news/why-the-onion-is-wrong-about-cnn-and-miley-cyrus-1200589821/ (accessed 8 October 2013).

70.

Weaver

Wilhoit

(1986) The American Journalist: A Portrait of US News People and Their Work. Bloomington, IN: Indiana University Press.

71.

Weaver

Wilhoit

(1996) The American Journalist in the 1990s: U.S. News People at the End of an Era. Mahwah, NJ: Lawrence Erlbaum Associates.

72.

Weaver

Beam

Brownlee

. (2007) The American Journalist in the 21st Century: U.S. News People at the Dawn of a New Millennium. Mahwah, NJ and London: Lawrence Erlbaum Associates.

73.

Webster

Phalen

(1994) Victim, consumer, or commodity? Audience models in communication policy. In: Ettema

Whitney

(eds) Audiencemaking: How the Media Create the Audience. New York: SAGE, pp. 19–37.

74.

Willig

(2013) Newsroom ethnography in a field perspective. Journalism 14: 372–387.

75.

Yin

(2009) Case Study Research: Design and Methods. Thousand Oaks, CA: SAGE.