From permissive to resistive tactics: How audience members engage with and make sense of datafied journalism

Abstract

While audience data are pivotal to producing journalism, audiences’ perspectives on the issue have received relatively little attention. Addressing this gap, the paper examines audience members’ tactics for making sense of and engaging with the datafied journalism into which they contribute with their data. Empirically grounded in group interviews and instant-messaging group chats with 21 readers of prominent Finnish tabloid Iltalehti, the author identified four tactics, along a continuum from permissive to resistive: an audience member may 1) happily benefit from datafied journalism; 2) be resigned to it yet reflect critically on it; 3) act to prevent effects on personal news-consumption patterns, by curating the content; or 4) entirely restrain themselves from engaging with it. Awareness of these tactics, which help individuals cope with and navigate the datafied-journalism landscape, facilitates grasping the factors in audiences’ relations to datafied journalism and, thereby, understanding their consumption of news and their relationship with journalism.

Introduction

Scholars have reached general consensus that datafication is inseparable from day-to-day life. The term ‘datafication’ denotes a trend whereby human action in the digital environment is transformed into numeric form, with the resulting data output getting collected, stored, and analysed within almost all organisations, across society (e.g. Cukier and Mayer-Schoenberger, 2013; Hintz et al., 2019; Mejias and Couldry, 2019). Datafication has visible impact in the fields of health (Mahnke et al., 2023; Ruckenstein and Schüll, 2017), and insurance (Tanninen et al., 2022). Furthermore, it is evident in media and journalism too (Loosen, 2018), which forms the focus of this study. With societal developments from the onward march of datafication, the early stance of enthusiasm for what is often denoted as ‘big data’ (Cukier and Mayer-Schoenberger, 2013) has given way somewhat to more critical angles, especially among media-studies scholars. Researchers have pointed to concerns and emerging concrete problems related to such matters as data capitalism (Couldry and Mejias, 2020); surveillance capitalism (Zuboff, 2019); and issues of privacy, algorithms’ control, discrimination, and connections with inequalities new and old (Kennedy, 2018).

Turning a critical eye to the new structures entails a need to study the practices whereby ‘ordinary people’ interact with datafied systems to which they have no direct power in terms of the core functions of the algorithms or the platform structures – only in terms of their own interpretations and actions with such systems (e.g. Kennedy, 2018; Ruckenstein, 2023). Research in this vein has focused on how algorithms’ functions have been understood and made sense (Lomborg and Kapsch, 2020; Swart, 2021; Ytre-Arne and Moe, 2021a) or felt (Ruckenstein, 2023) but also on user agency. For instance, researchers have delved into citizens’ activism in encounters with datafied systems (Kennedy, 2018; Velkova and Kaun, 2021), ways of consuming social-media-conveyed news (Swart, 2021), and how they try to affect what personalisation algorithms ‘spit out’ (Swart, 2021; Thurman and Schifferes, 2012).

Researchers probing datafication of the everyday have observed that users are rarely fully aware of how datafied systems operate. Their understanding and control over the personal data involved, as well as the systems’ utilisation of said data, are quite limited, and that inability to influence the operations can lead to uneasiness (Ruckenstein, 2023) manifested in ‘data anxieties’ (Pink et al., 2018). Draper and Turow (2019) refer to the notion of digital resignation, where people desire better control over datafied systems but often feel unable to achieve it. From these foundations, user studies have examined how people manage living alongside datafied systems, such as by coping (Møller Hartley and Schwartz, 2020), disengaging (Gangneux, 2021), or disconnecting (Moe and Madsen, 2021). All of these mechanisms often tie in with reflexive practices whereby users seek fuller understanding of the datafied systems and negotiate their engagement with them (Gangneux, 2021; Lomborg and Kapsch, 2020).

Journalism studies discipline overwhelmingly acknowledges datafication’s influence on journalism, especially its influence on journalists’ work. As they have access to wide range of sources to open data and capacities to process it into information or public knowledge, and newsrooms can automatically publish articles with the help of algorithms and collect data of their own news sites’ users – the resulting journalism, can be called datafied journalism (Loosen, 2018). The use of audience data has also changed journalists’ understandings of their audiences. While journalists have always envisioned imagined audiences guiding their work (Nelson, 2021), in the datafied context, the audience is perceived as measurable and algorithmic (Anderson, 2011; Zamith, 2018). Journalists continuously monitor audiences’ online news consumption, influencing story volume, topic placement, selection, and news presentation formats and styles (Fürst, 2020). In these ways, news consumption – what users click and read – influence journalistic production.

Some of these datafied operations happen on third-party social media platforms, where news organisations must adapt to platform power. However, the majority of these operations occur on the news outlets’ own web sites and news apps, where news organisations collect data, and their staff uses it. Consequently, discussions about the impact of user data on journalism are closely linked to debates about journalism’s responsibility and role in relation to the public. Because journalism is a ‘cultivator of publics’, the datafication of journalism transcends experiences of individuals – it reaches between them, exerting effects on how the publics are constructed (Møller Hartley et al., 2023) and how journalists fulfil their tasks in a well-functioning democracy, such as informing people and acting as watchdogs over those in power.

The public role of journalism raises questions about the type of data operators users perceive the news organisations to be. For instance, Ehrlén et al. (2023) found in interviews with young adults that news media are considered more trustworthy and ethical data collectors than social media companies, despite audiences possibly being unaware of news organisations’ data practices (as noted by Heikkilä, 2022; Tandoc, 2019). This observation prompted further questions about whether users respond differently to datafied journalism than to other aspects of datafied everyday life. Therefore, this paper contributes to studies on the datafication of everyday life and journalism. Specifically, it examines users’ news-consumption perspectives – an area that has received less attention.

The study presented here is rooted in actively produced qualitative material obtained through group interviews and ‘group chats’ conducted via instant messaging with members of the audience of mainstream Finnish tabloid Iltalehti. The paper applies a framework anchored in the distinction between strategies and tactics articulated by Michel De Certeau, 1984. The key argument is that audience members hold only tactical power over their actions when operating within the structures of datafied journalism (news outlets). These structures are created by news entities that possess strategic power to define them. The central research question of the article is: what kind of tactics have the audience members created in response to datafied journalism? This inquiry allows to explore how news users make sense of and what they do with the space allotted to them, how they either conform to or challenge the structures of datafied journalism. Additionally, since some aspects of these tactics are reflected in newsrooms through audience data, and then transmitted back to news users via the content they consume, studying these tactics also sheds light on how users’ engagement with journalism may shape journalism and public discussion.

Tactical and strategic power in datafied audience–journalism relations

In The Practice of Everyday Life, Michel De Certeau, 1984 applies distinction between tactics and strategies to describe how people conduct their everyday activities. Taking a city as an example, he illustrated how some actors hold strategic power to dictate the framing structure: where the streets run and the locations for hospitals and schools. Thus, he articulated that those with strategic power can exercise it over the state of affairs and the future. Still, the people walking that city as ‘ordinary’ citizens may appropriate it or strive to resist the planning frames by taking shortcuts instead of following the main roads. They can interpret and contemplate the city to find the ways of living in it that suit them best. Accordingly, citizens can exert power over their own actions: they can operate within the structure, using it in the manner they find beneficial. Through their tactics, citizens shape their practices within the space allotted to them, and these practices may influence future planning. However, the authority to define structures rests with those who hold strategic power.

Henry Jenkins (2013) applied this distinction when examining fandom. He observed that although fans are tactical operators, they exert significant power when reinterpreting the meanings of original media products through collective reading, interpretation, and fan fiction creation. In a more recent context, studies have applied the strategic versus tactical power lens when addressing issues of users’ relations to platforms, algorithms, and data (e.g. Møller Hartley and Schwartz, 2020; Velkova and Kaun, 2021). Møller Hartley and Schwartz (2020: 12) posited that users employ ‘coping tactics’ that represent ‘practices of resistance and appropriation by the seemingly less powerful in the response to asymmetric structures of power’. Additionally, Velkova and Kaun (2021) asserted that users can create ‘tactics of repair’. These tactics may involve users ‘complying with algorithmic logics but resisting their output, tricking algorithms to work toward unintended ends’ (Velkova and Kaun, 2021: 525). These tactics underscore not only users’ subordinate position in power relations with platforms and algorithms but also their methods of influencing on those.

Similar power dynamics are present in journalism. Journalists are indeed subordinate to their managers and news organisations’ shareholders, who wield strategic power – for instance, in resource allocation decisions. However, with an increasing amount of news dissemination and consumption occurring on social media and as news organisations rely on third-party metrics providers, news organisations find themselves in a subordinate position relative to tech-giants. In this context, newsrooms negotiate their infrastructural position relative to big-data giants, encompassing the extent to which they act such that they embrace these players’ platforms or resist the algorithms (Chua, 2023), but also manage their dependence on providers of other services, such as audience-analytics programs (Kristensen and Sørensen, 2023). These various technological platforms frame journalists’ abilities to function in the digital age and thus offer them a role of actor with tactical power, yet much strategic power between audiences and journalism resides in newsrooms. Journalists have command over both the content of news production and its form, and news organisations decide how and to what degree they allow analytics and algorithms to inform their operations.

Audience members can wield their tactical power to engage with the news in unconventional ways, such as starting to read the print newspaper from the last page – a practice that newsrooms can be unaware of (Heikkilä et al., 2023). Although such news-consumption-related tactics have existed in the pre-digital era, current audience analytics now provide data about the tactics to the newsrooms at a much higher pace and volume. Journalists actively interpret and act upon such data, observing how citizens consume news online. However, they often lack insight into broader tactics behind the numerical data they receive. This underscores the significance of studying news users’ tactics within the context of datafied journalism to understand how news users navigate the space allotted to them and make sense of their own news consumption and the data practices of journalism institutions. Therefore, the present study elaborates on this conceptualisation by examining tactics as a combination of 1) discursive reflection, through which individuals make sense of datafied journalism and its functions and 2) their actions (news consumption etc.) influenced by that reflection.

However, the audience members are often left to guess the degree to which their own data influence journalistic outputs they consume – that is, to what extent the journalistic outputs present a reflection of user’s past behaviour however distorted. Thus, consumers can adjust how they engage with the black box of datafied journalism. Therefore, we can regard tactics as a performance of data reflectivity, through which people interpret and act upon the structures created by datafied journalism.

The research setting

The project was set in Finland, which Hallin and Mancini’s well-known taxonomy of media systems situates alongside other Nordic countries under ‘Democratic Corporatist Model’, a category characterised by features such as early democratisation, a strong welfare state, strong professionalisation, extensive newspaper circulation, and strong public-service broadcasting (2004: 68). These countries are noteworthy for advanced Internet penetration, use of online news and other digital services, and trust in legacy media among their citizens (Harrie, 2017: 13). Finland, with its 5.6 million inhabitants, stands out especially for high trust in news: according to the digital news report from Reuters, 69% of Finns surveyed stated that they trust most news most of the time (Reunanen, 2023). Such a context of strong democratic traditions and public respect for the journalistic profession presents a specific but fruitful environment for examining audience’s experiences with datafication in journalism.

The terrain of Finland’s news media landscape comprises one widely read national daily and two popular evening tabloid news magazines, a strong public-service broadcaster, and various local and regional newspapers (Reunanen, 2023). The dawn of the 2000s saw tabloid magazines begin a transformation from crime-news outlets to popular daily papers covering political and economic news (Kivioja, 2018). This study focused on readers of one of the above-mentioned high-circulation tabloids, Iltalehti. Thanks to substantial investments in online journalism in recent years and its active exploitation of audience analytics, Iltalehti boasts a readership of 2.7 million people weekly for its free online content, and the print version still reaches approximately 175,000 readers daily (Media Audit Finland, 2023). While the majority of its income comes from advertisement revenues, ongoing efforts saw it introduce a paid online service centred on human-interest and wellness stories. In the 2 years since the early-2021 launch of that service, Iltalehti Plus, it has amassed a base of nearly 50,000 subscribers. ¹ Such developments render Iltalehti an ideal case for research on datafied journalism.

For the study, conducted in spring 2023, participants who read Iltalehti were sought via an announcement placed on Iltalehti’s online site. Nearly 400 people who filled in the associated Web form soliciting demographic background expressed willingness to take part in the research. From among these, five groups, totalling 21 individuals, were formed. The participants were grouped by age, since similarities in life situations/patterns among people of similar ages may influence approaches to participating and communication style, especially in digital environments. Notwithstanding efforts to recruit as rich a pool of informants as possible, women and people with a higher education were disproportionately prominent among the 400 volunteers relative to the readers overall, whereas young adults and pensioners were considerably more difficult to recruit. The five groups cover an age range of 24 to 71 years, with the average age being 47, and only seven individuals were male. Also, the nature of self-selection entails greater eagerness for involvement among people interested in the topic at hand, who are bound to possess greater initial knowledge of it. While the sample was far from representative, many participants were by no means experts in this arena (cf. Pink et al., 2018). Therefore, these groups are able to offer a valuable window to people’s encounters with datafied journalism.

Planning of the research setting proceeded on the assumption that datafied journalism might well be an opaque phenomenon to participants. Hence, several aspects of the prompts (e.g. the examples, questions, and background on them) were designed to tease out experiences with datafied journalism. A three-phase design for obtaining the data supported this aim further.

In Phase 1, all participants were asked to take a quiz to identify true versus false statements pertaining to datafication and datafied journalism. The correct answers, along with a brief explanation for each, were supplied once they had submitted their set of responses. While offering background data, the quiz phase served mainly to introduce the topic and inform participants about it in an engaging manner.

The second stage employed group interviews. Each group were invited for a respective interview conducted via the online video-call service Zoom. The interview began with orientation to the nature of the research and requesting of each person’s informed consent to take part in the study and to recordings in connection with this. ² Each group reflected jointly on various matters highlighted by the quiz, such as whether they considered themselves knowledgeable with regard to the topic. Secondly, the facilitator asked them to discuss the list of most-read articles that was displayed on the Iltalehti site’s front page. This anchoring vividly accentuates collection of information from the users and its reflection back at them. The third matter addressed was the algorithm for personalising that page. As participants were informed of its functions, they were encouraged to compare between its normal presentation for them and the view from their Web browser’s ‘incognito mode’ (which is not affected by the user’s article-reading history). They also compared with each other’s version of the front page. Via group interviews, participants exercised an opportunity to share their knowledge of the topic, thus contributing both to scholarship and to their collective evaluation of the issue of datafied journalism.

For the third phase, these groups continued their participation via instant-messaging ‘group chat’ conducted via Signal, a platform designed for privacy (this environment was chosen for assurance of data security). The group-chat interaction was framed for reporting on various encounters with datafied journalism, not only those involving Iltalehti even if the outlet provided a connecting element for the participants. With tasks assigned each day for 1 week, participants were asked to pay attention to distinctive features of datafied journalism while consuming news as usual and to report back to the group with their thoughts about their encounters and with examples shedding light on these (via screenshots and links). The tasks – related to news outlets’ topic choices, visual and textual elements, headline- and picture-testing, front-page personalisation algorithms, and paywalls – directed participants toward those elements influenced most strongly by audience data (as pinpointed via journalism studies: e.g. Rydenfelt et al., 2022; Tandoc, 2019) and to the most explicit graphic examples of the role of audience data from audience members’ perspective. The group chats facilitated interactive communication between the participants and the researcher. This allowed the researcher to provide participants with reminders about tasks, seek clarifications, and encourage them to share examples of their activities and thoughts. Participants also had the opportunity to ask the researcher questions, particularly when the tasks were unclear.

Group chats demonstrated great utility by providing many ways of sharing details of encounters (coupled with the illustrations) and thus enabled to focus on the participants’ tactics. With the help of tasks, participants actively focussed on the outputs of journalism they encountered in their everyday lives. They shared these outputs with the group and explained how and why they consumed the news as they did. As a result, participants provided explanations and sense-making of their engagement with datafied journalism and evaluated their own practices in comparison to others. While researchers seldom utilise instant-messaging mechanisms for collection of data (Kaufmann and Peil, 2020; Kümpel, 2022), these features offered convenient anchors for collective sense-making.

The resulting dataset for this paper comes from the five group interviews and these groups’ instant-messaging chat records (phase 1’s quiz responses had only a supplemental function in generating the data, so they were not analysed). In addition to textual material (80 pages of interview transcripts and 420 group-chat messages), there were 76 screenshots and 15 links, shared via instant messaging.

The paper’s content analysis was theory-informed, as the concept of tactics guided attention to the moments of reflection and action. After collecting these, the material was organised with assistance from ATLAS.ti software. Once participants’ notions from reflections and actions in relation to datafied journalism had been organised thematically, the analysis progressed to the next stage: examining what kinds of actions participants cited as performed in relation to different features of datafied journalism and the nature of their associated reflection and reasoning. This analysis identified four distinctive tactics: benefiting, resigning, curating, and restraining. The next section details these, illustrating them via extracts from the dataset where participants are identified by pseudonym and age.

Findings: The audience’s tactics

It quickly became clear that the participants applied various kinds of tactics to make sense of and engage with datafied journalism. These were informed by experiences not only with journalism proper but also with datafied systems in other domains. Similarly, as noted in earlier research on user engagement with different datafied platforms and algorithms, the participants of this study also exhibited behaviours that aligned with or resisted the formal structure of datafied journalism (e.g. Lomborg and Kapsch, 2020) and found suitable ways to engage with it (e.g. Møller Hartley and Schwartz, 2020). The attitudes toward datafication at large and in overlapping domains shaped their encounters with datafication through journalism – and hence the tactics employed in these. However, these tactics extend beyond datafication itself, encompassing not only data collection and algorithms but also journalistic content influenced by data. Consequently, this paper sheds light on how news users respond to datafied journalism, where their expectations and experiences regarding journalistic content, data and algorithmic systems intersect.

Participants referred to concepts along a wide spectrum, from satisfaction to resigned powerlessness to even active resistance. All participants were active consumers of news, with none being either extremely critical or extremely permissive. Many negotiated their relation to datafied journalism simultaneously taking a critical stance but still pursuing to find the most suitable ways to act with it. The corresponding tactics may be aligned along a continuum from permissive to resistive. Same person could use different tactics depending on the target – that is, the feature or facet of datafied journalism that the individual was concerned about.

Benefiting

When participants referred to the tactic of benefiting, they focused on the positive features of the current news landscape of datafied journalism, evaluated how to use its features to get the most from it. With this tactical approach, participants acted within the structures created by the datafied journalism, neither acting against it nor reflecting on it negatively. This serves as a reminder that even considered reflections on datafied systems need not be entirely negative; some users choose a positive lens (e.g. Lomborg and Kapsch, 2020; Ruckenstein, 2023), as this tactic attests.

Tactics of benefiting targeted the current media environment’s vast array of content and platforms (Karlsen et al., 2020; Van Aelst et al., 2017), which produces a wide stream of news. Iltalehti alone publishes more than 120 articles each day, most of which are freely available. When collectively evaluating the upsides of datafied journalism, many considered to have easy access to several news sources and that they could find quality journalism, for which they would be happy to pay. Some participants expressed a desire to share examples of quality journalism with others. In this context, high consumption of quality journalism was considered one hallmark of a good citizen – a label that is rarely associated with high consumption of other types of online content.

Analytics mechanisms inform such outputs. Since journalists now monitor what kinds of topics and outputs attract readers (Fürst, 2020; Tandoc, 2019), they have been able to modify their content such that it responds better to audience wishes that get interpreted through data (Carlson, 2018). Regarding datafied journalism, the substantial volume of content did not appear to cause mere stress (e.g. Moe and Madsen, 2021). Instead, participants recognised that the extensive flow of news and easy accessibility fulfilled journalism’s role in their lives, leading to satisfaction and a sense of being well-informed.

Amid a high-bandwidth news supply, the participants found personalisation beneficial to them, in that the algorithms help them find more interesting stories and filter out the less interesting ones, a positive feature of algorithms, as noted in a prior study by Lomborg and Kapsch (2020). Recent years have seen numerous news outlets introduce algorithms for customisation (Bodó, 2021; Rydenfelt et al., 2022; Schjøtt Hansen and Møller Hartley, 2023) that many find to simplify life in media environments awash with choice (Ytre-Arne and Moe, 2021a). As indicated above, Iltalehti is among these: it uses an algorithm that curates the content for the online front page by tracking which articles the user has read historically. While 50-year-old participant Arttu had tried to train the algorithms of the news outlets he followed (for instance, tweaking the profile settings by selecting topics that interest him), such actions were rare among the participants. The majority of the participants did not seem interested in ‘training’ the algorithm (cf. Swart, 2023), and Sanna, 34, was more typical. She explained that trying to personalise the content by choosing one’s preferred topics demands too much effort, in that ‘you would have to click “like” at least three times so that you get to the profile settings’. Even when a news company offers tools whereby users can affect the news stream, these might not lead to action (as noted by Groot Kormelink and Costera Meijer, 2014): some participants were satisfied with allowing journalists and algorithms to make decisions. Tellingly, the majority of the participants did not try to influence the personalisation algorithm in any way, for instance, intentionally feeding it with their data (Lomborg and Kapsch, 2020). Its unfamiliarity – whether deliberate or not – might be one factor behind this: most did not know of its existence before taking part in the study. Likewise, as Swart (2021) pointed out, certain algorithms are easier to detect than others. This study found that personalisation algorithms on news outlets’ online front pages remained particularly opaque to participants.

In addition, the participants identified benefits in referring to the list of most-read articles that Iltalehti presents online which helps them home in on meaningful stories. Even though some reported that they rarely pay attention to the listing, let alone find it useful, several had integrated it into their daily news practices. For instance, some used it to check whether they had missed any ‘important’ stories that ‘everyone is talking about’ or things they should be aware of. Through this mechanism, news organisations provide participants with tools for engaging in a shared reality beyond personalised front pages. The list functioned as a ‘backup’ for informants, to ease the stress created by a high-choice news environment wherein one cannot read all the news.

Resigning

While participants above reflected on the upsides of datafied journalism, also critical standpoints were expressed. Although critical of datafied journalism, they still used the news, without modifying their actions in response. This meshes with Nora Draper and Joseph Turow’s (2019) definition of digital resignation, which, in the case of datafied journalism, extends beyond a sense of helplessness related to how data operators use individuals’ personal data; here, we can understand resignation as a tactic through which people who have striven to understand datafied journalism’s nature and underlying reasoning, express inability, or unwillingness to change their behaviour. Often, personal disappointment ensues.

Resigning was the most prevalent tactic in the sample. In one recurrent reflection on the troubling matter of making sense of data collection, participants reasoned that this is necessary for the economic success of journalism, a fair trade if one wishes to receive free content. However, resigned to donating their data, the participants worried about their lack of power over data collection in general and over their personal data specifically. The problems of datafication they experienced through journalism produced resignation with regard to news companies’ data mining, as Jonna, 26, stated poignantly. Having accepted that her personal information is already everywhere on the Internet, she just clicks ‘Yes’, ‘Yes’ for all cookies when she accesses a site. Likewise, Arttu rationalised his decision to accept the cookies thus: ‘When I access the Internet, I must accept that this is how it works’. It is crucial to note that acceptance does not indicate agency – even if considered voluntary, it may be forced (Ehrlén et al., 2023) and lead to ‘apathy’ resulting from feelings of powerlessness (Møller Hartley and Schwartz, 2020). People might lack resources to resist the system, such as the necessary information, energy, or alternatives. For instance, 71-year-old Ilpo said that, while he would like to decline all the cookies, this is impossible: accepting ‘essential’ cookies is compulsory. Furthermore, even if every Web page presents a tick box pertaining to cookies, there is no practical way to know what one is agreeing to on social media or on datafied journalism (e.g. Ruckenstein, 2023: 107). Hence, whatever consent is given, anxieties and doubts still exist in many situations (e.g. Tanninen et al., 2022). The sense of powerlessness related to personal data collection, in general, also caused stress during discussions about datafied journalism.

However, informants associated their resignation with journalistic content as well, particularly with headlines but with charging for material. Nowadays, news outlets worldwide look to online subscribers as an ever more significant source for funding of journalism (International Centre for Journalists, 2019), and evidence is visible in the rising number of paywalls, interfaces requiring the user to pay for accessing the news content (Myllylahti, 2019). Some participants expressed sympathy related to funding, whereas others tried to resist paywalls, albeit often unsuccessfully. Participants expressed personal post-subscription disappointment in various terms. Some referenced ‘giving in’, others stated that they had subscribed to something by accident, and a few mused on multiple axes. For instance, Julia, 45, said a special offer for Iltalehti Plus had ‘tricked’ her into subscribing. Also, she wondered how a tabloid magazine had been able to convince her to pay for human-interest content when she remained unwilling to pay for the online news from a regional newspaper that she actually valued more. These ruminations highlight that resigning is a tactic rather than a lack of tactics: the participants tried to make sense of and reflect on their sometimes-contradictory actions for themselves and for group members. This crystallises well via the gloss of being ‘tricked’; informants articulated that they had been unsuccessful in resisting datafied journalism and that they had also supported the strategies of datafied journalism. This reflects a relatively high level of literacy regarding the functions of datafied journalism: participants could discern the intended user actions and evaluate whether their own actions aligned with them (see also Lomborg and Kapsch, 2020).

The notion of being tricked was even more frequently attached to online news headlines. Though the tabloids of yore became infamous for shocking headlines screaming from their printed front pages (e.g. Kivioja, 2018), the online environment has intensified the role of headlines even further: 1) not only tabloids but also ‘legacy media’ utilise clickbait headlines, 2) decisions on what to read online are based mainly on headlines (Heikkilä et al., 2023), and 3) many news media employ A/B-testing algorithms to optimise headlines for click counts (Hagar and Diakopoulos, 2019). When asked to define ‘clickbait’, many participants referred to headlines that lure one to click, with them citing shocking or appealing headlines as highly stimulating. For instance, one informant, who said that she experienced online news headlines as exhausting on account of them being ‘pulled out of proportion, making small things look big’, expressed disappointment with her seeming inability to stop clicking:

Now I stopped to wonder why I even click news even if it doesn’t tell me anything new. But I still do it. Why do I keep doing this to myself? (Elli, 29)

Many participants admitted following links to news about topics that they considered unnecessary to read about because curiosity prompted them to click, justifying their news consumption to themselves and other participants. Often, unconsidered clicking led to personal dissatisfaction: they ended up feeling powerless to resist clickbait headlines and that they had acted as directed by the structures of datafied journalism. One participant, 34-year-old Sanna, described her sense of unease after clicking as ‘an instant hangover’, then explained:

It is the feeling you get after reading a headline like ‘Check the pics! / Divorce! / Someone flashed [someone]!’ and you know all too well there is nothing important and that reading it is not good for you, but you still click it.

Resigning can be understood as participants’ steering toward restricted agency in interactions with datafied journalism they feel unable to resist.

Curating

Many participants expressed wishes to resist the logic of datafied journalism but still actively consume news content. The core difference is that employing curation rather than resignation tactics entails hunting out and honing unconventional techniques of consuming news accordingly. The participants of this study expressed a desire to curate the content themselves, using intuition in combination with the lists of most recently released articles, rather than relying solely on personalised front pages. Via their non-traditional browsing habits, they aimed to minimise the influence of journalists and outlets algorithm’s on their news-exposure decisions. The objective is to alleviate stress and other problems associated with datafied journalism in a manner that enriches news consumption rather than encouraging disconnection.

In traditional nomenclature, it is the journalists who are the curators of information, gatekeepers ruling on what audiences ‘should’ know (Shoemaker and Vos, 2009); however, the analytics era has accorded audiences a place in gatekeeping practices (e.g. Salonen et al., 2023). This need not imply agency, though. Reader preferences get judged through analytics, with algorithms applying the resulting conclusions wholesale to modify online front pages’ layout (Tandoc, 2014) or providers tailoring the content to a user’s apparent preferences, for instance. Meanwhile, audience members in this study manifested distrust in editorial and algorithmic curation both. Although earlier studies offer evidence of some users preferring algorithmic over human gatekeeping (Thurman et al., 2019), the Iltalehti readers proved dubious about whether the personalisation algorithm’s output can reflect their genuine desires. As for the sample’s distrust in filtering by humans, concerns coalesced around how well this gauging of topics’ relevance might mesh with one’s own and whether the journalists even can interpret audience opinions correctly from data. To limit the risks, participants implemented curating informed by their intuition and conducted by untraditional browsing.

Intuition, which has been pinpointed as a practice through which people make sense of datafied systems, such as of news exposure in social media (Swart, 2021), strongly guided participants’ engagement with news too. It helped them decide whether a news item was worth reading or not. Guided by tacit knowledge, participants reached conclusions from the headline alone as to what kind of article lay behind the link which indicates a high level of news literacy. Ilpo opined that an ‘attention-seeking’ headline rarely points to an interesting story, and Sanna echoed his sentiments thus: ‘I feel that no real news has a clickbait headline, which is a good thing’. Many relied on intuition to resist the temptation of ‘accidental’ clicking or minimise risks of being ‘tricked’ into consuming shock-value news.

Atypical browsing habits contributed to some users’ curation tactics. Most often, this mechanism involved sorting the articles by publication time instead of following the front-page top-down. One reason cited for this method was its reduction in stimuli from visual elements. Kaisa, 48, explained that she likes to read the list of Iltalehti’s most recent articles via her mobile phone, which arranges the headlines chronologically, all in the same typeface and font size, without any photos or other visual elements. Her rationale was that this presentation ‘lacks some kind of noise information that I don’t need for anything’. Kaisa’s reasoning responds to the growth in popularity of visual elements in online journalism, a phenomenon arising from audience analytics (Fürst, 2020). Users targeted their curation at reducing information overload from non-visual stimuli too. For instance, by requesting to receive of push-notifications only for certain topics. Curation provided participants with a way to manage information overload without completely disconnecting.

This set of tactics extended to actively resisting algorithmic curation of news (items’ targeting etc.) via bespoke curating and browsing habits. For example, after Kaisa scrolls through the list of a news site’s most recently published articles, she consults the ‘top’ news on the front page. This reveals what the algorithms and journalists deem important. Ilpo described having kept nine tabs open in his Web browser for the last 3 years, to list the most recent pieces from nine news outlets. He described systematically scrolling through these Finnish and international sites’ listings until reaching an article he had already seen. Ilpo cited these reasons behind his action: doubting the journalists’ selection of what he deems the most interesting articles for the front page and not trusting the personalisation algorithm to function as he wished and that it would isolate him into a ‘filter bubble’. Ilpo’s reflection was consistent with work noting actions by which users attempt to see all the content on social media and filter it themselves (Swart, 2021) and fears of algorithm’s confinement feature (Ytre-Arne and Moe, 2021a).

Restraining

Restraining one from full engagement is a tactic reflecting efforts to maintain a distance from datafied journalism. In the study’s data, these tactics involve self-regulation by which participants strove to limit their news-consumption volume or restrict the trail of personal data left.

The past few years’ crisis-oriented coverage of climate change, the SARS-CoV-2 pandemic, and Russia’s invasion of Ukraine has ushered in conditions of high-volume consumption and production of news. For instance, Digital News Report data for 2020 point to a pandemic-impelled rise in digital news consumption in that year in several countries, Finland among them (Newman et al., 2021). Prior research shows that the simultaneous burgeoning supply of news articles can leave audience members feeling overwhelmed (Ytre-Arne and Moe, 2021b). Accordingly, many participants remarked that, while at first, they consumed COVID-19 and Ukraine-war news intensely; they soon found themselves exhausted and anxious. Avoiding the news followed. Pirjo, 54, reported that she had blocked stress-inducing news and become a more selective user. She characterised this as a tactic for coping with the stress that the news caused her.

There is ample evidence that audience size plays a significant role in story selection (Fürst, 2020). Journalists are willing to produce more articles when the topics perform well by analytics metrics (Nelson and Tandoc, 2019). In a situation of item quantities and frequent online-news release pace by conclusions from audience analytics (Fürst, 2020), this study’s participants often said that their response to the flood of material was to take a break from news, an action consistent with those noted in other studies (e.g. Villi et al., 2022; Ytre-Arne and Moe, 2021b). That said, the news avoidance was not confined to moments of crisis. Participant Päivi, 42, pointed out that the amount of news these days is so enormous that missing some of it should not lead to huge damage. She said that, while she had even been addicted to news at one point, according to great value to ‘my own time’ led her to reduce her news consumption. Anni, a 65-year-old informant, portrayed a cycle in which her news consumption has been so intense that she sometimes has needed a break. After these pauses, she soon returned to her old habits, frequently using multiple devices and news services in parallel. Restraining, therefore, refers to digital disconnection – a break from digital devices – as pinpointed in several fields (Moe and Madsen 2021). However, in the context of journalism, this tactic may have broader societal consequences. If news users increasingly avoid the news due to its stress-inducing effects, newsrooms might interpret this as mere disinterest.

Additional means of self-restraint entailed limiting one’s digital traces, the personal data one leaves in one’s wake. Some participants accepted only ‘essential’ cookies, cleared their browsing history and cookie cache regularly, or decided not to register for news services, a similar tactic, noted by Møller Hartley and Schwartz (2020), involves people minimising the risks associates with data collection and mining. With this set of tactics, the resistance targeted mostly datafication, with datafied journalism being merely one manifestation faced. Participants’ reasoning for their actions focused on privacy and surveillance (cf. Syvertsen, 2020: 19). Many reported hoping to minimise the quantity of data collected from them. For instance, 50-year-old Jaana expressed distaste for the idea of being subject to surveillance. She gave this as her reason for not signing up to use Iltalehti’s online news service or any others that, while fee-free, require registration. Janne, in turn, stated that he wanted to be ‘as useless as possible for those actors who collect information’ and that declining cookies and clearing his browsing history were part of his endeavour to minimise the amount of information that can be gleaned from him. It is noteworthy that a correlation was visible between a critical attitude to data collection as a whole and active efforts by the participant to control the personal data shared with online news outlets specifically.

Among others, in contrast, refusal to store cookies was less consistent. Julia said that ‘usually I forget about [the data-collection issue] quickly, like I don’t even care that much but still I try to decline cookies, just not systematically. It doesn’t matter so much’. Sanna too, while stressing that she was aware of data collection and even considered herself conscientious about the issue, recognised that she might be deluding herself since her actions against the system are not systematic.

This extract encapsulates the struggles of those exercising only tactical power within the maze of datafied systems. Often, genuine resistance demands effort that is burdensome to maintain.

Conclusions

The four classes of tactics discussed, focussing on benefits, resigning oneself, curating, and restricting one’s actions, capture the variety of news users’ reflections and actions responding to datafied journalism. How permissive/resistive the tactics were hinged on how harshly the participants evaluated datafied journalism and their level of struggling against its established structures. Depending on the context and target, individuals’ various tactics for their everyday news-consumption influence which news outlets they favour, what articles they read, and the degree to which they allow the news companies to collect data pertaining to them. Besides being actions in the moment, tactics are a significant tool for making sense of datafied journalism: many participants negotiated their relationship with datafied journalism by simultaneously adopting a critical stance while seeking the most suitable ways to engage with it. These tactics align with many findings from prior studies on user tactics across various datafied systems (e.g. Lomborg and Kapsch, 2020; Møller Hartley and Schwartz, 2020; Ruckenstein, 2023; Ytre-Arne and Moe, 2021a). Datafied journalism employs similar user data-related practices and algorithms as other actors online, and users seem to employ similar tactics regardless of the context and target. This underscores the reflective nature of datafied journalism, in which user data and journalistic content are deeply intertwined. Thus, the same user tactics can target both data collection and journalistic content.

However, some tactics were specific to datafied journalism. For instance, participants expressed satisfaction with the ‘benefiting’ tactic when they felt that datafied journalism helped them become better informed, and high consumption of journalism was considered a hallmark of a good citizen. The ‘resigning’ tactic was not only targeted at data collection but also at journalistic content. Participants reflected on how they implicitly gave their approval to content they did not appreciate, such as to clickbait headlines. The ‘curation’ tactic demonstrated how some participants applied curation learnt from other datafied platforms to manage their news consumption and to interact with journalism’s strategy-based frames in the harmony that suits the user best to minimise the need for using the ‘restraining’ tactic (such as avoiding news content). As traces of these tactics are reflected via audience data to the newsrooms and actively monitored, the tactics do not remain as individual experiences and actions, but also mould journalism and public discussion.

The scope of tactics raises questions of audience agency and the power imbalance in datafied journalism. By employing the ‘benefiting’ tactic, news users endorse datafied journalism and strengthen its existing structure. Although the ‘resigning’ tactic does not immediately shake the structure of datafied journalism, it indicates simmering resistance, a potential for future actions against datafied journalism. On the other hand, the ‘curating’ tactic shows that news users can influence their news exposure and engage in unconventional ways with datafied journalism. The ‘restraining’ tactic presents a choice of disengagement with the structure. In the case of datafied journalism, this tactic might have broader societal implications if many decide to distance themselves from datafied journalism by avoiding the news.

Tactics equip critical audience members with routes to living with datafied journalism despite its flaws and without complete disconnection. However, even audience members who are happy to contribute their data freely for use in datafied journalism have little influence over the extent of that use, the purposes to which the data are put, and the ultimate effects – outcomes that rest in the hands of those holding strategic power. Tellingly, the users in this study, while dissatisfied with or critical of datafied journalism, did not articulate a desire for it to change. They seemed to see the responsibility as theirs: though restricted in agency, they indicated that they should exercise better control over their data and media use. This appears consistent with the commonly stated view that in the current media landscape people are often expected to regulate their own actions if experiencing it as overly invasive rather than demanding the media itself to change (Moe and Madsen, 2021; Syvertsen, 2017). All the while, the strategic power resides in the newsrooms, leaving audiences with only tactics-level power which can hardly change the functions of datafied journalism, its fundaments, etc. If they want to receive news and operate as informed citizens, they must negotiate ways to live with datafied journalism, a practice revealed in studying tactics.

Since audiences’ relationship to journalism has not only personal but also societal consequences, the findings discussed here point to potentially troubling long-term implications embodied in current datafied journalism. Audience members experience this journalism as sparking negative emotions to such an extent that they must develop tactics for counteracting the stress. This is intimately connected with how they relate to journalism. Such phenomena illustrate the flow of changes between journalism’s strategies and corresponding adjustments in users’ tactics for engaging with journalism. For newsroom as holders of strategic power, it is vital for them to understand the consequences of their actions and, if they desire stable audience–journalism relations, to offer several ways of consuming news. This study’s findings highlight that data (demographic factors, cultural context, etc.) cannot capture the sum total of audience members’ news consumption. It is influenced also by factors invisible to any lens from data alone – and by the journalism itself.

The findings point to several avenues for further work. Firstly, the particularity of the Iltalehti context must be acknowledged. Although the participants, all readers of this tabloid, were encouraged to collect examples from every encounter with datafied journalism, the focus necessarily remained on Iltalehti. There is evidence that the context of news-reading influences the interpretation of an article, and, for instance, readers may be more critical of tabloids than of other such media (Kivioja, 2018). Hence, the reactions found in this study might be more negative than those of news audiences as a whole and the tactics more resistive in nature. Secondly, the methods’ collective nature simultaneously enabled natural avenues for sense-making and evaluation through group discussions. However, the group interviews may have led the participants to present a more polished or critical view of their news consumption and attitudes. This may have highlighted the role of resistive tactics in the findings. Nevertheless, the interaction between the researchers and participants provided a suitable approach for conducting research on opaque topics and great utility with which to interrelate research participation with the participants’ daily news consumption. Thirdly, the national context and Finland’s generally high trust in news – stemming from strong trust in the society, its democratic system, and therefore its national media (Ehrlén et al., 2023) – may have exerted a moderating influence on the tactics. Trust Finns experience in the news media might decrease the need felt for stricter resistive tactics, while other cultural contexts might manifest more expressions of distrust in datafied journalism and of stronger resistance mechanisms. Therefore, the study’s findings leave room for future research into this important subject. Work could fruitfully focus on other news outlets and cultural contexts, to enrich understanding of audience perspectives on datafied journalism.