What clicks actually mean: Exploring digital news user practices

Abstract

This article problematizes the relationship between clicks and audience interests. Clicking patterns are often seen as evidence that news users are mostly interested in junk news, leading to concerns about the state of journalism and the implications for society. Asking and observing how 56 users actually browse news and what clicking and not clicking mean to them, we identified 30 distinct considerations for (not) clicking and classified them into three categories: cognitive, affective and pragmatic. The results suggest, first, that interest is too crude a term to account for the variety of people’s considerations for (not) clicking. Second, even if one aims for roughly estimating people’s news interests, clicks are a flawed instrument because a lack of clicking does not measure people’s lack of interest in news. Third, taking users’ browsing patterns seriously could help bridge the gap between what people need as citizens and what they actually consume. Finally, we argue that all metrics should be critically assessed from a user perspective rather than taken at face value.

Introduction

The digitalization of journalism has enabled news organizations to minutely monitor the behaviour of online news users. Through such tools as Chartbeat and Google Analytics news professionals know exactly and often in real time how many users are spending how much time on which news item. Web metrics are not only monitored by individual journalists but also displayed on big screens in newsrooms and forwarded to staff by editors-in-chief.

Scholars and news professionals have tended to take metrics at face value by assuming a close correspondence between clicks and audience interests. Since ‘most viewed lists’ are often dominated by news about entertainment, crime and sports, it is assumed that news users are more interested in ‘junk’ than in ‘public affairs’ news (politics, economics, international relationships) (e.g. Tenenboim and Cohen, 2015; Tewksbury, 2003). Boczkowski and Mitchelstein (2013) speak of a ‘news gap’ between the preferences of journalists and news users. This article problematizes the relationship between clicks and audience interests. Rather than looking at metrics, we observed how news users in everyday circumstances browse news and asked them what moves them to click or not to click. The aim of this research is to explore what (not) clicking means to people and to what extent clicks reflect their news interests.

Professional autonomy versus pleasing the masses

Monitoring audiences is hardly new. Schlesinger (1978: 111) describes how the BBC News had a large wallchart tracking how its Nine O’Clock News was doing in the ratings. However, monitoring was done mostly to track how they were doing relative to their competition. Other than today, the audience in itself was not an important consideration for journalists (Darnton, 1975; Gans, 1979; Schlesinger, 1978). Journalists had neither the tools nor the need for knowledge about their audience, as one producer illustrates: ‘I know we have twenty million viewers, but I don’t know who they are. I don’t know what the audience wants, and I don’t care’ (Gans, 1979: 234). Indeed, journalists actively resisted audience feedback. In the early 2000s, public TV journalists interpreted any discussion of audiences as potentially compromising journalistic autonomy (Costera Meijer, 2003). Journalists also feared that taking audience preferences into consideration equalled lowering journalistic standards (Costera Meijer, 2013; Gans, 1979). This binary opposition between professional autonomy and pleasing the masses, between making quality journalism while users apparently prefer trivial news, is deeply ingrained in the journalism profession.

The impact of clicks

Despite the autonomy-popularity binary, today, journalistic considerations have become audience-centric. Research shows how by and large, news organizations are having metrics inform their editorial decisions, from news presentation (news placement, headline adjustment) to news production (expanding or following up heavily clicked stories) (Anderson, 2011; MacGregor, 2007; Vu, 2014). Cross-lagged analyses show that audience clicks affect both news placement (Lee et al., 2014) and subsequent reporting (Welbers et al., 2015). Tandoc (2014) illustrates how editors select and de-select news items based on the web traffic they generate.

News organizations aiming for popularity monitor clicks most closely, whereas those whose brand identity hinges on quality emphasize the importance of their professional judgment (Welbers et al., 2015). Yet, Tandoc (2014) found no evidence of journalists weighing editorial autonomy and accommodating audiences while observing their everyday practices. Christin (2014) illustrates an ambivalence towards clicks: even journalists critical towards click-chasing do ‘understand online success as a signal of professional value’ (n.p.). Similarly, Usher (2013) found that journalists at Al Jazeera English, who do not have to take economics into consideration, were nevertheless monitoring metrics for ‘personal validation’; they want their stories to do well (p. 346). Karlsson and Clerwall (2013) found that while public service journalists may not track metrics in real time, they do look to clicks to ‘prove’ their public relevance and thus provide legitimacy. Finally, journalists in the same study suggest clicks deserve not just a critical attitude as they can also help increase the quality of journalism: if clicks indicate ‘important’ stories did not reach the audience, journalists can take action (Karlsson and Clerwall, 2013).

Race to the bottom?

Scholars typically evaluate clicks from a critical perspective because clicking patterns are seen as evidence that users prefer junk news over news about public affairs (Boczkowski and Mitchelstein, 2013; Tenenboim and Cohen, 2015; Tewksbury 2003). This leads to worries about the future state of journalism and the implications for society. Although Nguyen (2013) notes that metrics ‘provide a considerable amount of accurate and reliable information for journalists and news executives […] to serve people in a more considered, more scientific manner’, he warns that using them uncritically can lead to ‘the dumbing down of news’ and ‘a disaster for public life in the long term’ (p. 157). Tandoc and Thomas (2015) argue that the use of metrics ‘has the potential to lock journalism into a race towards the lowest common denominator, ghettoizing citizens into bundles based on narrow preferences and predilections rather than drawing them into a community’ (p. 247). Such observations echo journalists’ assumptions about the narrow scope of the interests of the general public. As one journalist in Usher’s (2013) research noted, ‘On a certain level you just can’t give the masses what they want. You are selling your soul’ (p. 343).

The relationship between clicks and interest

Scholars – and journalists – typically measure clicks in terms of the most read or viewed news stories, and in turn use these as a proxy for people’s ‘preference’ for or ‘interest’ in news (e.g. Boczkowski and Mitchelstein, 2013; Schaudt and Carpenter, 2009; Tenenboim and Cohen, 2015). For instance, Schaudt and Carpenter (2009) conclude from most-viewed stories lists that readers ‘most preferred’ the news values ‘proximity’ and ‘conflict’ and ‘least preferred’ ‘timeliness’ and ‘prominence’. Similarly, Tenenboim and Cohen (2015) argue that ‘sensational topics and curiosity-arousing elements’ being most heavily clicked indicates ‘that news consumers are mostly interested in non-public affairs news’ (p. 212). However, other research suggests that clicking patterns may not accurately or fully capture the interests or preferences of news users. We previously found that people engage in online user practices that do not necessitate clicks but do express interest in news, such as ‘checking’, ‘monitoring’, ‘scanning’ and ‘snacking’ (Costera Meijer and Groot Kormelink, 2015). Von Krogh and Andersson (2015) found that measured in clicks (page views), ‘public sphere’ accounted for 9 per cent of online news consumption, whereas measured in spent time it made up 20 per cent. Therefore, our aim is to explore what it actually means when users click on news and also what it means when they do not click, and how this relates to their (lack of) interest in or preference for news.

Methodology

To explore what clicks actually mean, we researched people’s considerations for clicking and not clicking by looking at their everyday online news browsing. Our approach was mixing interviews with sensory ethnography and the think-aloud protocol (cf. O’Brien et al. (2014), who combined the think-aloud protocol with a simulated work task scenario). First, participants were asked to describe how they use news throughout the day, focusing on the ‘multisensoriality’ (Pink 2009: 1) of their experiences (e.g. what they feel, taste, smell, hear or see when using news). This approach allowed us to get a layered picture of their news use (e.g. checking news with an espresso or while riding the bus) and enabled our participants to call to mind their news user practices. Subsequently, using the concurrent think-aloud protocol (Van Den Haak et al., 2003), participants were instructed to browse news as they normally would – using their own devices and preferred websites and apps – and to say out loud all their steps and considerations. Participants were encouraged to comment on actions they failed to mention spontaneously. We argue that the subsequent loss of natural flow was warranted given our aim of uncovering considerations for (not) clicking; indeed, subtle or subconscious actions like scanning or scrolling past a headline were as important as consciously clicking on news.

It should be noted that although most participants had little problem verbalizing their motivations, news users may not know precisely what they want and why they want it. Yet, we argue that having participants provide their own account of why they did (not) click might give a more accurate reflection of what clicking means to them than having them choose from pre-selected categories, as is often the case in uses and gratifications research (see, for an overview, Ruggiero, 2000). Although socially desirable answers should never be ruled out, the ease with which participants ‘admitted’ to reading entertainment or being tired of news about Syria suggests we obtained a fair picture of the news they would normally (not) click. We also sought to limit social desirability by having the interviewers demonstrate the think-aloud protocol to participants using such ‘interviewer self-disclosures’ (Lindlof, 1995: 182) as ‘I usually go to the entertainment section’. Finally, participants were selected from the social network of the interviewers as ‘the development of a personal relationship’ is crucial for interviews that go ‘deeply into the person’s experiences’ (Lindlof, 1995: 171).

We seek to map the whole spectrum of considerations for (not) clicking rather than look for the distribution, frequency or representativeness of clicking patterns. Yet, common user patterns found across a relevant variety of news users might point to firmly anchored user patterns in general. A total of 56 people were interviewed in an everyday setting, typically their home. Participants were selected using ‘maximum variation sampling’ which seeks to generate a wide range of data by including a broad spectrum of users (List, 2004). To enable capturing a variety of consideration for (not) clicking, our selection included 28 younger (19–35) and 28 older (50–65) users with various news habits (e.g. light or heavy digital use). The participants were from various (e.g. rural, urban) parts of the Netherlands, a country characterized by high rates of Internet penetration (96%) and online news use (81%) (Newman et al., 2016). We might, therefore, assume that their routines or preferences rather than obstructive technology (e.g. bad Internet connection) were the main factors in the participants’ browsing behaviour. The interviews were conducted in February and March 2014 by seven Journalism MA students from Vrije Universiteit Amsterdam and typically lasted 20–40 minutes. This included the browsing of websites and apps, which ranged from quick ‘checking cycles’ (Costera Meijer and Groot Kormelink, 2015) to lengthier reading sessions, depending on how the participant would normally use news. Devices used included computers, laptops, tablets and smartphones. Visited websites and apps varied but often concerned major Dutch titles including Nu.nl, NOS.nl and Telegraaf.nl. The interviewers received extensive interview training and exhaustive feedback after each interview round. All interviews were recorded and transcribed. The process of analysis drew from the Grounded Theory Method, using constant comparison between data and analysis and allowing categories to emerge from the data themselves (Corbin and Strauss, 1990). Initial open coding was done by the first author, whereas the labelling of the concepts and subsequent integration of concepts into categories was done in collaboration with the second author.

Because we are interested in participants’ own considerations for (not) clicking, the categories are illustrated through interview quotes. Even if some labels seem self-evident, exploring the meaning of clicks for users demands taking seriously the perspective of the participants. Also, participants often had multiple reasons for (not) clicking on one particular headline, but since we want to map the variety and range of user patterns, the quotes illustrate the categories in their ‘purest’ form.

Considerations for clicking and not clicking

Following the procedures of the Grounded Theory Method (Corbin and Strauss, 1990), we found 30 distinct considerations for clicking and not clicking. After an extensive process of axial coding, the first major distinguishing factor between the considerations appeared to be whether or not they were content-related. The content-related considerations proved to be further divisible into cognitive and affective considerations. Here, ‘cognitive’ refers to considerations where the decision whether or not to click was made predominantly on a mental level (‘thinking’), whereas ‘affective’ refers to considerations where the decision was made predominantly on an emotional level (‘feeling’). We use the term ‘predominantly’ because the distinction between cognition and affect was more gradual than absolute. In the third category, not participants’ thoughts or feelings about content but their pragmatic considerations were their dominant reference point for clicking or not clicking. For each consideration, we will note whether it concerns a reason to click or not to click (or both), and where applicable, we will discuss how it relates to selection criteria of news professionals (cf. O’Neill and Harcup, 2009).

Cognitive considerations

Recency and importance might be expected to be dominant considerations from a production perspective (cf. Golding and Elliott, 1979), yet were not mentioned much by our participants. Recency refers to whether the participant sees the news as timely or current. The limited mentions of this consideration seem to contradict research that indicates how users expect being presented with the latest news online (Bergström, 2008; Groot Kormelink and Costera Meijer, 2014). However, we argue that from a user perspective, recency may constitute a general prerequisite for online news but not an important consideration when deciding which particular news item to click on.

Importance refers to whether the participant views the news as significant in the conventional sense. However, if from a professional perspective importance is about ‘need to know’ (Golding and Elliott, 1979: 118), from a user perspective ‘ought to know’ is a more accurate description. Sandra (25) illustrates how the placement of news on a website influences how important she perceives it to be:

‘Cabinet: no clear picture of money laundering’, I couldn’t care less, so wouldn’t click on that. […] If it was REALLY important it would have been big at the top [of the homepage]. Then maybe I would’ve clicked on it.

Online news presented as important through prominent placement on the website or news app is experienced as more worthy of knowing; if the same news is placed less prominently, it apparently is not significant enough to deserve a click. Like recency, importance is not a dominant consideration when deciding which individual headline to click on. They are not so much selection criteria for (not) clicking on news as prerequisites for selecting a news site or app in the first place. Indeed, reflecting the original function of the front page of newspapers, users do expect (professional) news websites or apps to show them what is recent and important (Groot Kormelink and Costera Meijer, 2014).

Participants often clicked on news that had personal relevance, relating to their everyday life, including work. This consideration is dual, meaning that it counts as reason to click when present and as reason not to click when absent. Henry (55), who invests, clicked on a news item about the stock market, but skipped a headline concerning the shares of a specific company: ‘I [don’t invest] in companies, so the particular company mentioned here I couldn’t care less about’. Matthew (25) clicked on a headline about Samsung Galaxy S5: ‘because I want to buy a new phone’ but skipped news about rented housing because ‘after [I leave my student house] I’m not going to rent, I will buy something immediately’.

Golding and Elliott (1979) distinguish between professional selection criteria geographical proximity and cultural proximity, and from a user perspective we found a similar distinction. Both considerations are dual. First, participants tended to click if they saw the headline as concerning news taking place within their immediate surroundings, regardless of absolute distance. Bianca (54) clicked on a headline about a dead body found 20 km away from her hometown: ‘[City] is so close, I just wanna know. […] And if it’s not so close then it’s not interesting’. Yet, Tracy (53) skipped a headline about an accident that happened within a similar distance because she did not experience it as nearby: ‘I think it didn’t happen in this region but somewhere in the south. No, that doesn’t really interest me’. Golding and Elliot’s (1979) ‘cultural proximity’ depends ‘on what is familiar and within the experience of journalists and their audience’, but for our participants, more specifically, it refers to whether they recognize a kinship with the subject of the news, again regardless of absolute distance (p. 166). Leonard (24) clicked on sports news concerning compatriots: ‘I like cycling, especially if Dutch people are participating. […] I don’t have to know if some Slovak won a round in Poland’. Conversely, Dutch native Andrew (58) did not click on a headline regarding Antilleans in the Netherlands because he does not feel a connection: ‘It may be important, but […] not for me right now. […] Because I don’t do anything with Antilleans. […] I mean, I don’t know one Antillean and I don’t know if they’re good or bad’.

Whereas for journalists unexpected refers to rare, out-of-the-ordinary developments (Galtung and Ruge, 1965), from a user perspective it is about whether the news fits their idea of what is common. Lilly (26) clicked on a headline about a joint action from a trade union and an employers’ organization: ‘Seems interesting, I’m curious why [they] are on the same page here, seems a bit illogical’. It is important to stress that what is unexpected to journalists may not be experienced as unexpected by users, and vice versa. For instance, Anita (21) did not click on news about a man lighting himself on fire: ‘Yeah, it’s bad, but it’s, I don’t care […] because uhm, yeah it happens regularly’.

Related to ‘unexpected’ is the reason this is logical, where the user does not click because from their perspective the news is (too) obvious. Regarding the headline ‘Nokia unsure about brand name for the future’, Nanda (21) noted that she already knew Nokia was not doing well: ‘Then this seems like a logical continuation. Then I don’t have to read it, because I already know why that is’.

Like journalists selecting stories already in the news (Galtung and Ruge, 1965; Harcup and O’Neill, 2001), participants regularly clicked on follow-ups to stories they had read before. Lauren (26) illustrates, ‘What catches my eye immediately is the headline [….] “Exam fraud [school] costs 3 million euro.” I’ve followed [that story] before’. A dominant reason not to click was that the participant already knew about the news. Not to be confused with ‘follow-up’ where users click on a new development, here they are already familiar with this particular development, as Karen (50) indicates: ‘[This] I already just heard, so I’m not going to read that again’.

A dominant dual consideration was whether the subject of the headline rang a bell with the participant. This concerned famous people but also names or events the participants recognized but could not quite place, as Nina (54) illustrates: ‘That Benno L., you’ve heard something about that before and then [you’re] like, gosh, who was that Benno L. again?’ Conversely, Eddy (53) asks why he would click if the subject matter does not ring a bell: ‘“Fight parenting clinic and insurer resolved,” well, I wouldn’t know what a parenting clinic is, so (laughs) I’m like, why should I read that?’

More detail on particulars comes into consideration when the headline raises a question in the participant’s mind, causing them to want to know more about the situation, as Jack (56) illustrates: ‘Heavy weather in Italy, I see […] (clicks) What is going on here?’ For a similar reason, Karen (50) clicked on a headline about a fishing ban: ‘Then I’m like, what do we catch there? […] What kind of fish is swimming there?’

Another reason to click was that the news enables participants to join in conversation. Rod (24) explains why he clicked on a headline about the Winter Olympics:

Because if you start a conversation with people then often you want to talk about things that uh are recent and speak to a lot of people and uh the Winter Olympics I think are a part of that, so uhm to be able to join in the conversation, so to speak.

Rod’s reason for clicking is the social utility function the topic provides: fodder for conversation. Teacher Joe (26) similarly clicked on a headline about the ‘largest lunar impact ever recorded’ ‘because I also talk about that with my students’.

Participants also clicked if they had their own opinion about a headline and wanted to see how it was discussed in the article. Jenna (27) clicked on the headline ‘World Bank freezes aid to Uganda over gay law’ because ‘I personally have an opinion about it, so I’m curious on what grounds the World Bank does something like that’. However, this consideration was uncommon; like in Donsbach’s (1991) study that relativized the influence of cognitive dissonance on readers’ selections, our participants rarely expressed strong opinions about headlines. If they did, disagreement was not a reason not to click.

Participants regularly did not click on news they thought was repeating itself. We labelled this supersaturation. Bruce (55) noted about the ongoing crisis in Syria: ‘Because every day it’s the same, same, same, at some point it becomes less interesting. Even though it’s not less terrible’. This is less about ‘compassion fatigue’ (Moeller, 1999) than about how hearing about it again does not provide new insights. The headline does not invite a click anymore, as Jeff (58) illustrates: ‘You actually drown in that kind of news. At some point you’re like, it’s not going to stop anyway. It’s not that it’s not important, but it doesn’t stop’. As we will elaborate later, not wanting to click on a headline does not mean the user does not want to see it. But for now the headline itself provides a sufficient update about the situation; it is not until ‘something completely new’ happens that Jeff (58) will click again.

Some participants clicked on headlines that offered a new perspective. This is not about the news event being unexpected but about the headline offering ‘the other side’ of a topic. Such news inspires because it adds to your knowledge or broadens your horizon and as such enables an aha-erlebnis (cf. Costera Meijer, 2013). Corbin (24) illustrates,

Here’s an article called ‘According to these three imams the Koran has nothing against gays’. That’s interesting to me [because] you have this image that in the Koran it says that homosexuality is wrong and here it says something completely different, and I’m curious how that is substantiated by those imams.

Rather than the topic of homosexuality and the Islam it is the original angle of the headline that makes Corbin click.

Sometimes, participants clicked on a headline because they wanted to see for themselves or ‘experience’ what happened. We labelled this participatory perspective. An example is Nick (24), who clicked on the headline ‘Man makes illegal base jump from moving ski lift’ because he ‘can’t really picture how anyone would do that’ and hoped to see it in a video.

A reason not to click was that the headline was just an opinion. Regarding a developing story about the possible resignation of a minister, Tara (20) noted, ‘If a decision really has been taken, I’ll find it interesting, but […] nine out of ten times it’s blether. […] If [prime minister] says “[He] is staying,” then that’s not a truth but just an opinion’. What keeps Tara from clicking is the lack of validity or decisiveness.

A similar reason for not clicking was disjointed news fact, where the participant does not want to read a story until it is finalized. Tara (20) is not interested in clicking on isolated updates about developments she is already aware of: ‘I don’t need to have that information in between, […] I want the answer, you know, the conclusion’. From a user perspective, even the conclusion of a story can be a disjointed news fact. Mark (52) did not click on a headline concerning a resolved conflict, because he was not aware of the problem in the first place: ‘You have to know what the problem is [and] then you can also know: what is the solution? […] But yeah, just an isolated little fact, I would never read that’. About a headline regarding the Ukraine, he similarly argued that it concerned a detail too small to warrant a click. If he were to click, he would also want to know the context: ‘What is the cause? How did it happen? What happened? Why do they do it? What do they want to achieve?’ This suggests that Mark would appreciate a headline like ‘5 things you should know about the crisis in the Ukraine’ that allowed him to get a full picture of the situation within one article.

An important finding was that sometimes the participant did show (signs of) interest in particular news items and yet did not click. The narrative construction of the headline appeared to be a relevant factor. A frequent occurrence was that the participants showed interest in the news itself but the headline was informationally complete and consequently, they did not expect to be better informed by clicking. Lauren (26) noted, ‘“More than 4 million viewers for Olympic finals 1500 meters,” that’s a fun fact to know, but I know that this is usually all the information you’re gonna get, so I don’t really have to click it anymore’. This is the opposite of clickbait: Lauren is interested in the topic, but there is no need to click because the headline tells the whole story. Nick (24) similarly illustrates, ‘I see it says “Final will be great,” so I already know they’re in the final so I don’t necessarily have to click it’.

Finally, sometimes there was an associative gap: despite the participants’ apparent interest in a topic, the headline did not tell them enough to want to click. Ella (51) read, ‘Pieterburen [location of a famous seal crèche] will possibly move to [island]’ and said, ‘The headline doesn’t tell me much, that’s why I don’t click it’. However, later in the interview she did click on a headline that explicitly mentioned ‘seals’ and said she was fascinated by them. Clearly, she had not made the connection between Pieterburen and seals. Based on clicks, it would be tempting to conclude that Ellen was not interested in this article, but based on her comments about how much seals ‘intrigue’ her, it seems safe to assume that she is. Similarly, Matthew (25) was clear about his interest in clicking the headline ‘Warning Kerry about Cold War Ukraine’, claiming he was following all news about the country because he planned to visit its city Chernobyl, ‘and of course I’m not gonna go there if there is almost a civil war’. Yet, he did not click on a headline about former Ukrainian president Janoekovitsj because ‘I don’t know exactly who that is, so I think I would skip that’. While this consideration is similar to ‘ring a bell’, the focus here is not the topic; instead, it is about not being able to make a connection between the headline and the user’s (pre-existing) interest in the topic.

Cognitive considerations

• Recency. Whether the user sees the news as timely or current.

• Importance. Whether the user sees the news as something they ought to know.

• Personal relevance. Whether the topic has a relation to the user’s everyday life.

• Geographical proximity. Whether the user sees the news as concerning their immediate surroundings.

• Cultural proximity. Whether the user recognizes a kinship with the news.

• Unexpected. Whether the user sees the news as surprising.

• This is logical. The user thinks the news is obvious.

• Follow-up. The user wants to know the sequel to a story she has been following.

• Already know. The user has already heard the news elsewhere.

• Ring a bell. Whether the protagonist or subject matter of the news rings a bell with the user.

• More detail on particulars. The user wants to know what exactly is going on.

• Join in conversation. The user expects to be able to bring the news up in conversation.

• Own opinion. The user wants to see how a topic they have an opinion about is discussed in the news.

• Supersaturation. The user thinks the news repeats itself too often.

• New perspective. The headline offers a different perspective that sheds new light on the topic.

• Participatory perspective. The user wants to witness the news event.

• Just an opinion. The user wants facts rather than opinions.

• Disjointed news fact. The user wants the whole story, not an isolated update.

• Informational completeness. The user has no need to click because the headline says it all.

• Associative gap. The user is unable to connect the headline to the topic.

Affective considerations

Participants clicked on disheartening headlines, similar to the news value ‘bad news’ (Harcup and O’Neill, 2001). Sarah (21) illustrates, ‘This one I would read: “Biker killed by car.” That’s just sad’. However, if participants found the headline too disheartening, they skipped it: ‘It’s such a heavy text, “Dragging patients is risky.” I prefer starting with happy news’ (Jeff, 58). Indeed, on the other side of the emotional spectrum, participants clicked on light-hearted, fun headlines that made them feel good. Isabel (30) illustrates, ‘Something about self-cleaning plastic for cars. […] Yeah, that’s a fun news item. […] It’s light, […] just nice to read’. While this corresponds to the news value ‘good news’ (Harcup and O’Neill, 2001), from a user perspective ‘feel good’ is about the impact of the news rather than its genre.

Some participants clicked on headlines because they found the accompanying image visually appealing, also a selection criterion for journalists (‘visual attractiveness’) (Golding and Elliott 1979: 155). Danny (25) is not interested in the news itself, but the picture evokes arousal: ‘On Nu.nl you often have these dumb items about, I don’t know, the New Year’s dive. Couldn’t care less, but if it happens to have a picture of a lady, I do click on it’.

Similar to the news value Harcup and O’Neill (2001) labelled ‘entertainment’, participants regularly clicked on headlines that bemused them. More specific than wanting to be amused, they feel a strong urge to click on the headline because they feel excitedly puzzled by it. Eva (19) illustrates, ‘Something provocative like “Anders Breivik: Playstation 2 instead of Playstation 3 is torture,” […] then I think what is this about? And then I click it and read it’. Such headlines usually concern remarkable or bizarre news, which might partially explain why this type of news is so heavily clicked (cf. Tenenboim and Cohen, 2015). The colloquial term for this is clickbait – headlines with a ‘what-the-hell’ factor that makes the user want to click, as Martin (24) illustrates: ‘Actually it never has any news value, but it’s usually those headlines that make you think, yeah, I’m curious what it is exactly’. A related reason not to click was that the participant felt the news was bullshit. Leonard (24) explains, ‘Now I see “German cat survives 30-meter fall.” Then you’re like, I don’t care. […] I think it’s a bit rubbish actually’. We classified this as affective instead of cognitive because it is a gut reaction dismissing the pettiness of the headline rather than a cognitive deliberation about whether or not the topic is nonsense. However, this consideration was mentioned less often than ‘bemusement’, where the silliness of the headline was exactly what does make users click.

Another dominant dual consideration for (not) clicking was the categorical welcome or rejection of a particular ‘beat’ or topic that participants felt, respectively, enthusiasm or aversion towards. The latter was often the case with sports news, as Ruth (24) illustrates: ‘The last [headline] is sports news, sports mean nothing to me’. Anita (21), on the contrary, categorically welcomes news about sports with which she has affinity but rejects others: ‘I don’t find soccer interesting, so I skip those headlines automatically. But ice skating and tennis, those I do follow’. While this consideration is similar to ‘personal relevance’, the emphasis here is on the feeling the headline evokes rather than the recognition of how the topic relates to one’s life.

A surprising finding was that some participants clicked on news that gleefully annoyed them. Lilly (26) clicked on the headline ‘President of Uganda will sign antigay law’ because she found it ‘particularly bothersome that again there is a country that does not understand that homosexuality is not something you should draft a law against, so yeah, I’ll read that news and be very irritated by it’. Isabel (30), similarly, clicked rather than ignored a headline that annoyed her: ‘Bart Veldkamp once again has an opinion. […] Now he thinks that the Netherlands should share their ice-skating knowledge. […] It does evoke a bit of irritation, that headline. I’m like, you became a Belgian’.

Affective considerations

• Disheartenment. The user is saddened by the news.

• Feel-good. The light-hearted news makes the user feel good.

• Visual appeal. The image evokes the urge to want to see more.

• Bemusement. The user feels excitedly puzzled by the headline.

• Bullshit. The user instantly dismisses of the pettiness of the headline.

• Categorical welcome/rejection. The user feels either enthusiasm or aversion towards the beat or the topic of the news.

• Gleeful annoyance. The user is delightfully enraged by the news.

Pragmatic considerations

Some participants did not click on news that would disrupt an otherwise smooth user experience, for instance, due to loading time or commercials when clicking videos. Bruce (55) illustrates, ‘Then you have to sit through commercials before you can watch something. Well, I won’t do that, I don’t want to’. A related reason not to click mostly associated with videos was that the item was data-heavy. Here platform-specificity also plays a role. Joe (26) does click videos about wrestling news on his computer, but not on his smartphone: ‘Videos […] I’d rather not watch on my phone because, well, data heavy’. Clicking would cost him too much.

Finally, participants did not click on news when it did not fit their routine. Josh (62) only has a few minutes to check headlines before he leaves for work, where the radio is playing the whole day. He skipped a headline about a poison gas attack in Syria, explaining, ‘That’s very important, […] but I’m sure I’ll hear it on the radio’. Similarly, Jenna (27) skipped a headline noting she would only click on it if she ‘really took the time to really dive into it’. While interested, clicking right now did not fit her schedule.

Pragmatic considerations

• Disruption. Clicking will interrupt a smooth user experience.

• Data-heaviness. Clicking will use up too much data.

• Does not fit routine. Clicking does not match with the user’s schedule.

Browsing patterns without click

Another important finding is that the participants engaged in online browsing patterns that did express interest in news, yet did not necessitate a click. Costera Meijer and Groot Kormelink (2015) have previously labelled these distinct user practices ‘checking’, ‘monitoring’, ‘snacking’ and ‘scanning’. Checking means quickly and efficiently finding out whether anything new or interesting is happening by looking at the latest headlines (Costera Meijer and Groot Kormelink, 2015). Clicks are not automatically involved, as Billy (52) illustrates: ‘For me it’s important to just quickly see things, so just a [homepage] is fine, just the [headlines]. […] I actually use it just to quickly see what the latest news is’. Just because users do not click on an item, does not mean that they do not want to see the headlines. Danny (25) explains, ‘It’s nice that you kind of know what is happening in the world. Because let’s say [the item] wasn’t there anymore. […] Then people start to talk and then you really don’t know anything about it’. For social purposes, then, he does want to check the latest important news: ‘The headlines I’d want, yes, but the articles themselves, uh, whatever’.

Monitoring is ‘actively surveying the informational environment to be able to come into action when necessary’ (Costera Meijer and Groot Kormelink, 2015: 671). Annabel (53) describes, ‘When those two criminals were on the loose last week, I constantly looked on my phone to see if they were caught. That was really scary’. Even though Annabel was continually monitoring her smartphone for updates, no clicks were registered. In similar fashion, Henry (55) uses his phone to monitor his investments: ‘I return to that at least once every two hours, because I want to see how my portfolio is developing’. In both the cases, their evident interest in news was not captured in clicks.

Snacking is defined as grabbing ‘bits and pieces of information in a relaxed, easy-going fashion to gain a sense of what is going on’ (Costera Meijer and Groot Kormelink, 2015: 670). Danny (25) describes how he snacks on a website about movie news without clicking: ‘I scroll a bit and look at pictures and at movies and then I click away [from the site]’.

Scanning means picking out ‘the highlights of news in order to get the gist of the story’ (Costera Meijer and Groot Kormelink, 2015: 671). Tara (20) illustrates how scanning does not necessitate any clicking. Although she does want to know about the news, she plucks words from the lead on the homepage to get the essence: ‘Like here: they “foresee no profits,” it’s about “Dutch companies,” OK, then I know enough’. As noted, even headlines can be so informationally complete that they do not necessitate a click even if the user is interested.

Conclusion and discussion

This article explored what clicks mean from a user perspective and to what extent they reflect the interests of news users. Asking and observing how people browse online news, we found 30 considerations for clicking or not clicking, classifiable into three categories: cognitive, affective and pragmatic. These differences are not normative but descriptive of the level on which the decision to click or not click is made: mental, emotional or practical. Taking an open, user-centred grounded theory approach rather than employing prefigured categories (e.g. from uses and gratifications theory) has resulted in a more complex account of people’s digital news use. For instance, cognitive considerations are not limited to information seeking (surveillance) but include the (lack of) recognition of news (ring a bell, associative gap) and the perception of how news is presented (e.g. disjointed news fact, just an opinion, new perspective). Likewise, affective considerations go beyond entertainment or positive affect and include feelings of negative (disheartenment) and mixed affect (gleeful annoyance). Our user-centred approach has also generated a vocabulary for news values and selection criteria that puts focus not on how news is sent but how it is received. The detailed labels might be relevant for journalism professionals seeking to understand what user experiences like enthusiasm and aversion are based on and provide a handle on how to effectuate or avoid such reactions.

Our results suggest that while clicking does indicate some type of interest, preference or engagement towards news, these concepts are too crude to account for the wide variety of people’s considerations for (not) clicking; our precise labels provide a more fine-grained vocabulary. More importantly, even if one seeks a rough estimate of people’s news interests, clicks are a flawed instrument. First, pragmatic considerations unrelated to interest in content interfere with users’ clicking behaviour. Second, headlines can tell users interested in particular topics too little (associative gap) to warrant a click or enough (informational completeness) not to warrant a click. Finally, digital news user practices such as checking, monitoring, snacking and scanning may not involve any clicking, but do fulfil valuable functions for users, including being brought up to speed on the latest ‘public affairs’ developments without interrupting one’s news flow. In terms of news interests, then, the news gap between news makers and news users may not be as wide or unbridgeable as Boczkowski and Mitchelstein (2013) point out. If news users appreciate browsing without having to click, future research might explore the underlying logic of these experiences as well as how non-clicking browsing patterns can be optimally facilitated and measured.

Our argument is not that clicks are meaningless, they just capture a limited range of users’ interests or preferences. Clicks may be helpful for news organizations looking to increase traffic through A/B testing of headlines or article placement (Cherubini and Nielsen, 2016). Although they have proved complex to monetize (Batsell, 2015; Cherubini and Nielsen, 2016), metrics that measure various forms of engagement (see Napoli, 2011) seem promising because they capture a broader array of digital user practices than only clicking. Such metrics could help organize websites and apps so as to accommodate users’ diverse expectations and desires at different times and in different contexts. In addition, information about users might be used not for ‘ghettoizing citizens into bundles based on narrow preferences and predilections’ (Tandoc and Thomas, 2015: 247) but for tracing and providing news that has ‘proportional relevance’ (Costera Meijer, 2003) to different communities. For instance, students are not only interested in news about students but also – as participant Matthew suggested – interested in news about the starter-home market. However, each metric should be assessed critically rather than taken at face value. Our research has shown how an open, qualitative user-centred approach can help examine what metrics do and do not measure.

If clicks only tell part of the story, our own methods are not without limitations either. The concurrent think-aloud protocol forces participants to consider and verbalize actions that in everyday life are often done automatically or subconsciously. Therefore, we encourage other researchers to further explore clicking and not clicking using different methods, such as (video) ethnography, tracking devices or screen capture tools.

Finally, and paradoxically, by giving people what they supposedly ‘want’ – as captured in clicks – news organizations could end up harming not only democracy but also themselves, as adhering to clicks might lead to the trivialization of news and thus to a decreasing interest of users.