Proactive ephemerality: How journalists use automated and manual tweet deletion to minimize risk and its consequences for social media as a public archive

Abstract

Despite their ephemeral constantly changing nature, social media constitute an archive of public discourse. In this study, we examine when, how, and why journalists practice proactive ephemerality, deleting their tweets either manually or automatically to consider the viability of social media as a public record. Based on interviews conducted with journalists in New York City, we find many journalists delete their tweets, and that software-aided mass deletion is common, damaging Twitter’s standing as an archive. Through deletion, journalists manipulate temporality, exposing the public to a brief tweeting window to reduce risks and regain control in a precarious labor market and a harassment-ridden public sphere in which employers leave them largely unprotected. When deleting tweets mechanically, journalists emulate platform logic by depending—as commercial platforms often do—on automatic procedures rather than on human expertise. This constitutes a surrender of the very qualities that make human judgment so valuable.

Introduction

The stakes involved in the deletion of social media discourse have grown as commercial Internet platforms have become a central infrastructure of the public sphere (Plantin et al., 2018) constituting together with news organizations a “platform press” (Bell et al., 2017). These platforms have also become crucial systems for the “production, distribution, and circulation of cultural content” (Nieborg and Poell, 2018: 1) more generally. This is a shift the COVID-19 crisis underscores as many social activities from political debates to classroom discussions occur online (Lorenz et al., 2020). Although sometimes overlooked, these platforms are also crucial for the preservation of online content and not only for its immediate distribution. Twitter is an especially important component of this public ecosystem presenting itself and operating as the world’s “public square” (Dorsey, 5 September 2018).

In many countries, a majority of journalists operate an account on Twitter for different purposes, from news gathering and reporting to attracting new audiences and maintaining a relationship with their readers (Hanusch and Bruns, 2017). Most studies of Twitter’s journalistic role have examined journalists’ tweeting practices and the ways in which Twitter serves as a branding tool (e.g. Lewis and Molyneux, 2018), studying commission rather than omission. In contrast, the deletion of tweets by journalists and its consequences for the public record have not been analyzed systematically. In this study, we ask how journalists, as a particular class of culture workers, balance professional commitments to promoting and preserving public debate with the risks inherent to maintaining a long-term presence online, and consider the collective consequences such a balance has for social media as a public archive. We do so based on semi-structured interviews with journalists.

We examine here how and why journalists delete some (or all) of their tweets because we argue these individual practices and discourses have collective consequences for archiving and journalism. Both archiving and journalism are central to societies’ ability to reflexively consider the past and present and democratically chart a future course. Because these enterprises operate in a public sphere and in media industries running as a layer on top of commercial Internet platforms, it is crucial to consider the implications of social media deletion practices for future collective memory. Through this analysis, we consider more broadly how individuals proactively use deletion—both manual and automatic—to control their reputation and livelihood drawing on studies of how humans attempt to use machines (Gunkel, 2017) and manipulate time (Fine, 1990) to be (or feel to be) masters of their own destiny in an opaque and distributed system. Our findings suggest that many of the journalists we talked to deleted most of the tweets they had published by using a third-party service that allows them to delete—once or periodically—the bulk of their tweets. In this article, we refer to this practice as using a “deletion service.” We argue that in deleting their tweets, they proactively contribute out of necessity to the ephemeral nature of social media platforms such as Twitter, further curbing their utility as stable public archives.

Literature review

Web archiving and proactive ephemerality

Web archiving is widely acknowledged as a solution for addressing the problem of the web’s ephemerality, and for providing evidence of the Web’s past (Brügger, 2018). In the past two decades, methods for web archiving have been institutionalized and standardized. The Internet Archive and dozens of national libraries around the world have developed software, policies, and best practices for preserving large portions of the Web (Ben-David, 2019; Ben-David and Amram, 2018). However, methods for web archiving, along with the justification and legitimacy of web archives as bearers of public evidence, may be suited for the open web but are ill-equipped for preserving social media content inside proprietary “walled gardens” (Littman et al., 2018).

Nevertheless, researchers have pointed to the value of social media platforms as a digital depository for individuals (Ruppert, Law and Savage, 2013) and to the ways in which social and mobile media documentation shape people’s experiences and memories (Özkul and Humphreys, 2015). Scholars are relying on social media to investigate contemporary cultural phenomena while arguing that published and distributed content on social media platforms may reflect people’s beliefs and behaviors (e.g. Bonilla and Rosa, 2015). Thus, social media are historically significant (Ankerson, 2012) for diverse memory agents including historians, sociologists, media scholars, journalists as well as lay attempts people make to memorialize the past. From such a societal perspective, it seems important to ask to what extent is social media being preserved and by whom.

Although of clear retrospective importance, researchers have demonstrated how social media users correctly perceive this medium as ephemeral, live, and constantly changing, where new stories are pushing back old ones, encouraging users to forget rather than remember (Bayer et al., 2016; Kaun and Stiernstedt, 2014) because platforms continuously redesign themselves and in the process past content is reshaped or even gets lost in translation (Brügger, 2018). While scholars have considered the inadvertent ephemeral nature of web content (such as linkrot and content migration) and the challenges in archiving social media content (Ben-David, 2016; Brügger, 2018; Milligan, 2016), this current research focuses on what we define as “proactive ephemerality,” a social media phenomenon that occurs when users intentionally remove their own content from their social media profiles manually or with the aid of mechanical tools.

Although journalists’ tweeting practices have received a fair amount of scholarly attention (Hermida, 2013), scholars consider the extent to which journalists treat Twitter and other social media as ephemeral only obliquely. Journalists use Twitter to post breaking news, and their employers encourage them to do so (Usher et al., 2018). Hence, Twitter is a promotional venue for both the news organization and the individual journalist who uses Twitter as a branding tool to present their work and engage with audiences (Molyneux et al., 2018). In addition, journalists use Twitter to retweet opinion, humor, and other content posted by others (Molyneux, 2015). Another typical way journalists use Twitter is as a space for public note-taking. Hence, Twitter is an outlet for journalists’ live-tweeting of various public events (Barnard, 2016). Twitter is also an arena for meta-journalistic discourse where journalists criticize and reflect on journalism as a field (Carlson, 2016). These findings suggest that most journalists’ practices on Twitter are oriented toward the present.

Work on journalism and memory has demonstrated how news production and products are involved in shaping collective memory (Zelizer and Tenenboim-Weinblatt, 2014). In posting journalistic content on social media, journalists shape collective memory (Han, 2020). If a newspaper is the first draft of history, then journalists’ professional tweets are “the first draft of the first draft of history”—as described by one of the interviewed journalists in this study. We argue that to fully understand how news producers shape collective memory through their use of social media, we must attend not only to what they post but also to what they proactively delete.

Entrepreneurship and journalism

The entrepreneur is future-oriented and ego-tropic, creating new products and ventures in pursuit of an individual vision (Gill, 2013). Such a temporal individualist orientation is oppositional to the archival impetus of preserving the past in the interest of the collective future needs of society. In general, social media platforms reflect the entrepreneurial ethos of their founders, legitimizing users’ intensive image management (Marwick, 2013). Workers feel a need, in a precarious labor market, to constantly craft their social media identity to serve their occupational goals (Gershon, 2017).

The entrepreneurial ethos is increasingly dominant in contemporary journalism as news ventures adopt technology sector practices and norms. In such an entrepreneurial labor market, image management might become more legitimate (Carlson and Usher, 2016), potentially supporting extensive deletion. As entrepreneurial journalists change employers, beats, or political orientations, contexts collapse and collide into one another (Davis and Jurgenson, 2014). Discourse and practices accepted in a past context conflict with what is acceptable or expected in the present, presenting the journalist as inconsistent or insincere and hence calling their occupational suitability into question. One possible way of avoiding such a collapse is to pre-empt it by editing past discourse to conform to the present and future. Moreover, the constant pressure to produce news content in an entrepreneurial industry (Bakker, 2012) might make journalists more error-prone while tweeting, driving more deletion later.

Journalists use Twitter for “professional brand development” (Molyneux, 2015). However, they could vary in their tendency and approach to deletion. While many have an entrepreneurial career whether by choice or compulsion (Davidson and Meyers, 2015) driving perhaps more retrospective editing of a journalist’s profile, others still enjoy stable employment and hence have less of a need to constantly reshape their online presence.

Temporality and work

As the future-oriented focus of entrepreneurship illustrates, time is a central dimension of work. Workers deal with objective temporal constraints and subjectively experience time in various ways. Manipulating time is a key mechanism for maintaining social control in the workplace. Workers can create “temporal niche[s] . . . a slice of time, cut from the rest of work time in which a worker or group of workers have autonomy” (Fine, 1990: 112). Restaurant cooks create such niches communally by covering for their colleagues, thereby allowing them to take time-off mid-shift. In a large technology firm, managers exercised “boundary control,” made up of cultural and bureaucratic tools to motivate and incentivize workers to spend more time at work and less at home. Some employees resisted such control by setting a rigid personal workday schedule as well as by sharing responsibilities with their team to regain temporal control over their lives, in the knowledge that it would limit career advancement (Perlow, 1998).

Commercial social media platforms prompt users, as unpaid media workers, to share historical content but use this content to create a constant flow of ephemeral posts to focus users on recently shared content rather than to create coherent historical narratives and a functional archive users could manipulate (Kaun and Stiernstedt, 2014). Hence, social media users have limited options to secure temporal control.

Turning to journalism, there is “an insufficient address to the temporality of news consumption” (Zelizer, 2018: 119), a tendency in both professional and scholarly discourse to assume that the lag between news production and consumption is brief to non-existent. Journalists’ decision to delete tweets involves a reflection that tweets can be consumed and redistributed well after their initial posting outside their original temporal context. Deleting tweets could be a novel means of creating a temporal niche by erasing much of a journalist’s public record preserving a limited continuously moving temporal window into a news worker’s online presence.

In this study, we investigate the manifestation of proactive ephemerality in social media. To do so, we therefore ask the following research questions:

• RQ1. How do journalists practice deletion?

• RQ2. How do journalists perceive deletion on Twitter?

• RQ3. How do journalists explain its prominence?

Method

To understand why and when journalists delete tweets, we conducted in-depth interviews with 17 journalists from New York City between March and August 2019. We focused on journalists based in New York City because it is a global hub of cultural production (Currid, 2007) where many national American and global news organizations have a presence with visibility in and normative influence on the field of journalism more generally.

We located the first interviewees with the help of research fellows at the Columbia Tow Center for Digital Journalism. We asked them to introduce us to journalist colleagues who are active on Twitter, have been verified by the platform, and have over 5000 followers. The Columbia Journalism Review Twitter account tweeted a request with similar criteria. These three criteria indicate a significant professional public presence on the platform. Follower numbers well below that number would suggest an account not used for public professional goals. Later, we snowballed from early contacts. We did not explain in advance that the research focuses on deletion practices since we wanted to understand their emic perspective on deletion.

We conducted the interviews in-person in a site the interviewee chose. The semi-structured Interviews lasted between 45 and 120 minutes. We promised all the informants complete anonymity. We recorded and fully transcribed the interviews. We started with general questions related to their news work and their Twitter use. These prompted an open conversation about tweeting practices. When the subject of deletion came up, we shifted the interview and focused on that issue asking the participant to describe the last time they deleted a tweet, what were the reasons for deleting, what kind of content they deleted and how they approached it in practice, and how they felt about the deletion they or others practiced.

We interviewed a diverse group of journalists. Nine identified as women and eight as men. In terms of employment status, nine were employed on a long-term basis by a news organization and eight were freelancers. Eight of them were reporters, four were editors, an additional four mainly wrote opinion pieces and one interviewee presented the news. The average number of years the interviewees had spent on Twitter was slightly more than 10 years, with a range of 7–13 years. We stopped interviewing after 17 interviews because we felt we had reached data saturation. We used Dedoose, a qualitative data analysis software, to focus-code key themes related to our theoretical framework (such as ephemerality, gendered harassment, and occupational considerations). We present here the findings according to these main themes.

Findings

Deletion practices

A majority of the journalists we interviewed said they delete tweets on a regular basis. Many interviewees suggested other journalists they knew also deleted tweets. The interviews indicate deletion can take many forms. Its extent varies between journalists, as do the tools they use, and the temporal patterns of their deletion practices. Nevertheless, a minority claim they delete rarely or never.

Many interviewees mentioned using a third-party service to delete tweets en masse. Some users (e.g. 4) noted the service they used was free, while others discussed using a paid service to delete large numbers of tweets at once. An examination of one service corroborates this claim as it found that a free service allowed up to 3200 tweets in total, while the paid version was unlimited (West, 2020b). ¹ Users report and our examination of the service corroborates that it allows users to mass delete “once only” or “automatically every few days.” The latter in combination with a setting of the “age of tweets to delete” creates a fixed temporal window (lasting between 1 week and 1 year) of published tweets, continuously trailing the present moment.

A smaller number of interviewees mentioned they delete some tweets from time to time for various reasons, with what seemed irregular frequency. When we asked interviewee 13 whether they deleted any tweets, their response was typical of a manual approach to deletion: “not in bulks but I delete some. Either something I kind of regret tweeting or just a tweet that got no activity at all . . . I just don’t want to clog up my timeline.” One female journalist (7) used Twitter’s own search function to identify what they viewed as problematic tweets:

Let me just double check that I didn’t like say anything totally out of character and crazy at some point . . . I searched like ‘slut’ and ‘whore’ and like a lot of like sexual terms . . . I like deleted some of the more like crazy stuff like just two or three tweets.

Those deleting large numbers of tweets at once, mentioned a particular service although some were vague about the name of the service and the extent to which they had shared their account data with the service. The following is a typical interaction between the interviewer and the interviewee regarding the service provider:

Q:

Do you know who they [company offering the service] are?

5:

No idea. It could be anyone.

This vagueness suggests the deleting journalists viewed the practice as commonplace. While we shall discuss deletion reasons, many journalists related their decision to mass delete to other journalists’ experiences rather than necessarily to their own: “seeing so many people get burned by having all their old tweets up. So, finally, I installed tweet-delete” (15). Several interviewees including 14 had learnt colleagues were using this service, prompting them to adopt it: “I had seen a couple of people like um, recommend them [the mass deletion services].”

Journalists noted that these external services were more efficient than manual deletion: “one of my coworkers I think was saying she . . . searches by year and then like was manually deleting them and I’m like, that seems like it’s such a waste of time” (15). Another journalist (6) noted, “I’d tweeted so much over the years that I just decided I would use tweetdelete and nuke everything.” As 4 put it succinctly: “you could manually delete all your tweets, but nobody has the time or patience to manually delete 80,000 tweets.”

Some interviewees suggested automatic deletion was more acceptable than manual deletion: “I responded to this guy who was accusing me of selectively deleting tweets I thought made [me] look bad and I wanted to clarify. No, it wasn’t selective. I just deleted them all” (4). One interviewee (1) added, “I’m not like a person who’s like gone back and deleted a bunch of tweets to erase like the positions I’ve taken.” Another interviewee (14) contrasted the proactive positive outcome of mass deletion (“you’re starting fresh”) with the defensive process involved in manual deletion (“it’s not like picking and choosing”). We now turn to consider the reasons interviewees assigned to their decisions to delete or preserve their tweets.

Ephemerality

One thing that stood out from the interviews is that the majority of the informants perceive Twitter as an ephemeral platform. Many interviewees claimed that tweets, in the words of one interviewee (5), are “not written on stone, forever.”

Many journalists talked about their tweets as belonging to the particular moment in which they were created and that they do not want them “to live in the public forever.” This understanding allows them to tweet more freely, without being afraid that what they are currently tweeting will “come back later and bite me” (14). In that sense, as explained by a few informants, treating Twitter as an ephemeral medium (with the help of an external software that mass deletes their old tweets) changes the ways in which they are tweeting: “you can say this thing and then like it’s not going to follow you forever [. . .] it’s just in the moment and then it’ll go away” (14). Some journalists described the deletion of old tweets as a “fresh start” allowing them to feel safer and more confident while tweeting. Journalists emphasized that the value of Twitter lay in immediate consumption:

like, oh, I was offline. I went on vacation, but I came, I came back to work. I’m going to read Twitter from four weeks ago? No one would do that. But you would do that with a magazine . . . Like there’s something inherently more ephemeral about Twitter is it’s about what’s happening now and like it, it loses its value. (1)

Many informants described tweeting as a conversation and therefore not meant to leave a trail as most conversations do not. Interviewees perceived tweeting as a “live experience” (5), “a cocktail party conversation” or like “giving a talk. It’s ephemeral but you’re reaching people” (4). Although many informants acknowledge the role of Twitter in their journalistic work, most informants do not think of these tweets as part of the public record worthy of preservation.

The journalists we talked to distinguished between the stories they wrote, which they hope will survive “for history,” and their tweets which will probably disappear. Compared with tweets, which they described as “random thoughts or jokes,” the published stories are referred to as “my work,” and as “evidence.” As noted by one journalist (5) who mass deleted most of their tweets: “What I want to survive is my considered edited, thoughtful view. That’s the thing that should live on, I think.”

Harassment

One crucial motivation for journalists to delete tweets is harassment. Many informants reported harassment on various levels, from hate comments to receiving offending private messages. Others reported extreme cases in which they received messages threatening their family. This is how one male journalist described it:

I got a lot of antisemitic comments, graphics, you know, a few threats, physical threats and things like that. Um, my parents actually got a phone call as well, so that was pretty troubling . . . Um, but you know, once you kind of deal with that . . . you kinda get hardened to it. (13)

Although the incidents described in this quote are shared by a few of the journalists we interviewed, we found that the reaction to these experiences and the ways the informants described them to us varied. While the male journalists quoted above portrayed it as “annoying” or “kind of sad and even humorous” other journalists, and in particular female journalists, described it as a “terrifying” experience which made them feel “under attack.” Female journalists said they received a deluge of sexist tweets: “like tweets after tweets after tweets, look at her, like you can tell she’s Jewish like look at her big nose and that sort of thing [. . .] like come to my legs or just things like that” (15). Journalists often connected their own harassment experiences or that of others to their decision to delete tweets as a way to pre-empt future harassment. Deletion is a protection mechanism allowing them to have an online presence with little record of their past tweets which may put them and their families at risk. Some journalists described how they deleted tweets that received “unwanted attention.” This includes comments that were not labeled as harassment but indicated that they could probably presage it. Deleting tweets that received unwanted attention was a way to prevent that. Many journalists described how Twitter is easily “weaponized.” Other journalists described Twitter as a platform that cultivates a “gotcha culture,” or even as a “war zone” reflecting popular and scholarly accounts of the American public sphere as deeply polarized (Benkler et al., 2018). Some interviewees related harassment to a loss of control and deletion as a means of regaining it:

I remember feeling absolutely terrified because it seemed I couldn’t control it. It goes back to this thing about control where I could not control what people were sending me and it was coming in these massive volumes so every time, I’d refresh there’d be more and more, more [. . .] And to me tweet delete represents like a taking back of some kind of ownership of my online presence. (2)

Our findings indicated that some journalists who believed their old tweets had public value argued this was trumped by their personal safety. This was the main justification for deleting their own tweets and they also sympathized with other journalists who did so. As one female journalist noted,

But like for me, and I’m sure for like a lot of other people like you know we have to worry about ourselves and our own like, you know, immediate safety and like livelihoods and sense of safety and that sort of trumps any concern about like maintaining some sort of like future record. Right? Like especially in a time like now, when everything is so volatile. (14)

The differences between female and male journalist’s deletion practices stood out from the interviews. It is obvious from the interviews that women are more concerned with safety issues when they tweet. Furthermore, the male journalists we interviewed also reported that they assumed women are more exposed to online harassment. Many of them justified deleting tweets because of cases they heard of women being bullied: “I think especially women are worried that they’re going to attract the wrong kind of attention” (6). A few male journalists had this notion. Other White male journalists understood why journalists hailing from disadvantaged groups resorted to deletion. Our analysis suggests that such differential deletion practices, if they are generalizable, have important consequences for the public record.

Occupational considerations

Many of the interviewees related their decision to delete tweets or that of others to occupational factors. They emphasized that in a dynamic entrepreneurial labor market in which journalists changed employers and considered leaving journalism altogether, journalists had to minimize the risk that their past tweets might attract unwanted attention (as mentioned above) and eventually damage their job security and prospects. While some claimed they managed these risks by tweeting less frequently and more carefully, many others chose to delete some or most of their tweets to contend with these risks.

Journalists discussed occupational factors in relation to the full career cycle. Some suggested that individuals now contemplating a journalistic career who had a tweeting record stretching back into adolescence would probably delete their early tweets to hide immature behavior: “like some people started using Twitter when they were like, you know, 13, you know? . . . and thinking like I’m trying to start a career as a serious person and like I have all this like crap on my twitter feed” (16). While 16 was suggesting that other journalists might have used such a deletion strategy, another interviewee (1) noted that when he became a journalist, he deleted past posts which did not conform with his contemporary occupational identity.

For several of the interviewees, the reputational risks of past tweets did not recede as they established their careers because they perceived the journalistic labor market as unstable and competitive and believed hiring executives continuously monitored the online profiles of candidates and public reactions to these profiles:

the industry is so precarious right now and everybody is leaping from job to job in such a scary way that I think the market is just saturated with really talented people and you don’t want to give anybody even the slightest reason to not hire you. (6)

This interviewee (6) noted they had considered selective deletion of some tweets but realized “I’d never be able to go through cause I’d tweeted so much over the years that I just decided I would use tweetdelete and nuke everything.”

Another journalist (1) cited the case of a prominent journalist, Sarah Jeong, whose hiring to a prestigious position was severely criticized after right-wing activists re-posted tweets she had published years before (Cobb, 2018). She mentioned this case when discussing her decision to delete her own tweets in anticipation of future career developments:

I don’t want to have that happen to me. So it was a bit around that. I’m not saying that instigated it for me, but I’m saying that was definitely on my mind . . . that could prevent me from getting a job or make an editor thing.

This quote illustrates how journalists take into account the experiences of prominent journalists when they delete in a bid to avoid the difficulties those prominent figures encountered. A senior editor corroborated more junior journalists’ perceptions that employers used their twitter timeline to judge them: “It’s just like a sense of who they are and what they’re interested in” (10).

Journalists said they deleted tweets not only before or in anticipation of moves within or across news organizations but also as journalists considered exiting journalism to work in other fields. Interviewee 9 (and one other interviewee) highlighted the case of a well-known journalist turned political advisor who deleted a very large number of tweets a short time before his new position was announced. While some had publicly criticized this particular case, our interviewee suggested the decision to delete made sense:

it’s increasingly hard to remain in career-length journalism, very rare. And some of the best journalists I know, and the most opinionated by the way, have moved on, right? It [deletion] can give you freedom if you find the right kind of job.

Here, interviewee 9 suggested that having the individual freedom to delete tweets would allow journalists to safely pursue a career in a new field without being weighed down by past posts.

Many interviewees felt they had to manage risks alone and delete tweets they posted as journalists working for their employing news organization in the absence of any meaningful organizational support. While a senior editor (10) in a prominent news organization suggested her employer always supported her when she was attacked online, many more interviewees emphasized that their employers did not have the resources or inclination to support them. One journalist (9) noted,

when you’re under attack, you have no time to think about it. You need to defend yourself and you feel defenseless. And when you feel defenseless, you know, we’re not supposed to, you know, carry the burden of historical record on our own.

Journalists felt their employers “ill equipped and I have been dealing with this long enough that I know a lot better what to do and how to deal with it”(14), especially given that harassment was occurring online to workers working remotely. Interviewee 9 suggested that even powerful media companies did not defend their employees and, therefore, one should not expect weaker news organizations to be more protective.

Professional discourse as a resource for opposing deletion

A minority of journalists said they had not deleted in the past and would be wary of doing so in the future. Their explanations, we argue, illustrate how relatively powerful and autonomous individuals use professional discourse as a resource (Fine, 1996). Interviewee 17 emphasized their journalistic experience and frequent use of public records in their reporting and suggested this dependence meant “the idea of someone in my business or a journalist, like deleting the record is, I don’t, I can’t see doing that. And in fact like I think that it’s really sad.” They further emphasized how as a journalist they should be “held accountable” for what they publish. Interviewee 12, who emphasized his training at a prestigious news organization, also linked accountability with an opposition to deletion and minimized social media risks by mostly tweeting quotes of what other people were saying. Another journalist (10) who opposed deletion noted “the job of a journalist is to hold power to account,” even if they were attacked in response and argued deleting tweets would constitute caving in to “bad faith trolls.” Interviewee 8 argued that deletion contravened a journalistic commitment to transparency: “I think that transparency is such a huge part of journalism that deleting, once you delete something, you’re no longer being transparent.”

Why preserve?

Many interviewees discussed with us considerations for preserving their personal archives and motives to avoid deleting their entire tweeting history. Even journalists who eventually decided to use mass deletion services, for reasons detailed above, acknowledged the rationale for preserving tweets. One journalist even described how she had stopped using the mass deletion service for a while just to preserve a personal event she experienced and tweeted about in order not “to forget it.” Hence, one motivation to preserve tweets—or not to delete them—is to “not forget” meaningful events that were documented on Twitter. “I think it’s like an interesting record to have. I even like to go back sometimes, and I’ll be like, what was I tweeting six years ago today” (10). Another journalist (7) who did not use a mass deletion service noted, “I can use my own archives for not so much research, but like reference.”

In addition, since the journalists we talk to were screenshotting other tweets, they were aware that someone else might capture their own deleted tweets. Hence, one motivation to preserve tweets and avoid deleting them was the attention they might attract by the sheer act of deleting: “my worry is that if I delete it, it only brings further attention and then it can start a whole other round of uh ‘oh well she deleted it’” (10).

Discussion: proactive ephemerality as a form of individual platformization

We found, in response to our first research question (RQ1), that deletion, both manual and mechanical, is common among our interviewees. In response to the second research question (RQ2), we found that a majority of the interviewed journalists perceive the practice as common among their peers and legitimate. Journalists encounter numerous reputational, psychological, and physical risks inherent in managing an entrepreneurial ever-shifting career online in which occupational and broader cultural norms incessantly change. They felt under attack and in professional and physical danger and hence a deep lack of control. These feelings were felt most explicitly by female journalists. Such feelings align with American survey evidence (Lewis et al., 2020) that nearly all journalists experience online harassment but women, particularly young women, are more often subjected to the worst types of online abuse. In their struggle to reassert control, journalists created temporal niches by choosing to extensively and irreparably delete their Twitter timeline. Hence, journalists explain (RQ3) the common practice of tweet deletion by pointing to the numerous risks journalists now face.

Turning to the collective implications of these individual practices, the tools journalists use to reassert control necessarily involve a destruction of their record on Twitter as a public archive. While the temporal niches employees working in a physical workplace such as a restaurant (Fine, 1990) create involve the creation of an alternative safe space side by side with the continuing operation of the restaurant kitchen, journalists re-purpose Twitter and re-create it as a collection of temporal niches. The more temporally expansive arena of all or most published tweets no longer exists, neither in the present nor for future generations, after it is mass deleted.

From an archival perspective, the removal of published tweets is an ultimate expression of exclusion from future memory and historical discourse. The findings serve as a reminder to researchers and archivists, that what appears to be available now might be deleted later. While many stakeholders assume social media constitute a permanent and representative online record of social life (Zhao and Lindley, 2014), privileged social strata participate in more social media platforms (Hargittai, 2018). In addition to differential participation rates, our analysis demonstrates that the survival of digital representations also reflects social inequalities: the discourse of some people—minorities and women, for example—might be less likely to survive for posterity mirroring long-standing biases in analog archives that have traditionally preserved the discourses and deeds of the powerful.

In our conversations, participants imply an emergent temporal understanding of the public–private divide basic to much media work (Habermas, 1989). Their perception of social media as ephemeral (“a cocktail party”) and their use of mass deletion imply an understanding of the present as a public domain and the past as a private asset of the user. However, the same interviewees also note they screenshot the tweets of other prominent users and are aware that others might be archiving their own tweets. In addition, commercial platforms, such as Twitter that distribute digital media content and house some of it, surveille user behavior and retain extensive rights over much of the content (Zuboff, 2019). This suggests that the emergent temporal understanding of the present as public and the past as the user’s private asset is a comforting illusion more than it is an effective means of reducing occupational and physical risk.

Many journalists used automatic deletion services. In doing so, they consciously or unconsciously emulated platform rhetoric and practices exhibiting a form of ‘individualized platformization’: Just as platforms have for more than a decade championed the advantages of automatic algorithms for distributing and policing online discourse as an “objective approach” (Christin, 2017), journalists (among many other memory agents) understandably resort to automatic mechanical procedures as more efficient and legitimate within a system lacking in more nuanced deletion tools. The mathematician Norbert Wiener warned that when using “mechanical agency,” such as computers, society “had better be quite sure that the purpose put into the machine is the purpose which we really desire and not merely a colorful imitation of it” (Wiener, 1960: 1358). The massive, immediate, and irrevocable deletion of tweets is an illustration of the dangers Wiener warned us of when using “a mechanical agency.” When journalists mass delete, for many understandable reasons, they “individually platformize,” imitating the partially mechanical platforms they use and forsaking the very qualities of manual editing which are the unique hallmarks of journalists and many other cultural producers. The design recommendations we propose below might moderate the extent to which humans feel compelled to act like algorithmic platforms while urging the platforms to be more human.

In future studies, we hope to go beyond our empirical reliance on New York City journalists as we aim to further examine the extent of journalists’ tweet deletion as well as the relative contribution of various individual and organizational factors to the frequency of deletion. Moreover, journalists are not the only occupational group that practices deletion. Artists have for centuries destroyed their work for numerous reasons (Miller, 2019). Like journalists, they might also be proactively deleting their online presence with important potential implications for cultural heritage. Journalists themselves highlighted in our interviews that politicians and other public figures frequently delete tweets.

One data tool that aims to flag and display the deleted tweets of public officials and those running for office is Politwoops. This tool creates an archive of deleted tweets with the objective of increasing transparency and accountability in politics. As Meeks (2018) notes, such a tool provides scholars with an opportunity to study political impression management, and we would add that such a tool would allow scholars to consider the implications of social media deletion for the public record. In a study which examined the construction of Twitter archives of government agencies, McCammon (2020) demonstrates how deleted tweets create digital “memory holes” since government officials lack standards and policies regarding the practices of government employees on third-party social media platforms, such as Twitter. In a similar way, our findings suggest the construction of journalistic memory holes.

Conclusion: moving toward balanced deletion

In October 2019, Twitter’s head of product, Kayvon Beykpour noted, “I’m very interested in exploring how we might give customers more control” (Marino, 2019) over how their tweets are preserved and distributed. However, as of November 2020, Twitter provided only single-tweet deletion. Deletion is disincentivized by making it costly in terms of time and effort. As one journalist noted, “you could manually delete all your tweets, but nobody has the time or patience to manually delete 80,000 tweets.” Various third-party developers have stepped into the gap with easily accessible time-limited and keyword-driven mass deletion features. ² Together, the platform and third parties have turned a potentially nuanced action of retrospective editing into a dichotomous behavior—either one seldom or never deletes or one mass deletes a large number of tweets by assigning deletion responsibility to an automatic mechanism. Calls made in 2020 by museum curators to the public to preserve physical and digital ephemera regarding life with COVID-19 and amid the fight for and against social justice (Blume, 2020) illustrate the loss mass deletion could constitute for the public.

To achieve a better outcome that balances journalists’ wholly legitimate interest in protecting their physical and psychological well-being and occupational standing with the public’s interest in preserving portions of the social media record as a public archive, responsibility should be shared between the platform, news organizations, and journalists: The platform could create real-time and retrospective tools that would allow journalists to protect journalistic tweeting against permanent deletion by tagging them as preservation-worthy at the same time that they provide more efficient on-platform deletion options making nuanced deletion less costly. The final design of such features should be developed in consultation with the journalism community, given its crucial role in maintaining Twitter as a well-functioning “Public square” (Dorsey, 2018).

Furthermore, the creation of a more nuanced deletion toolkit demands not only technical solutions but also requires news organizations support their journalists’ social media activities as reflected in an internal report written by members of the Washington Post’s National desk recommending their employer assign full-time internal advocates to support their presence on social media (National Social Media Committee, 2020). The uneven relationship between Internet platforms and publishers seems to create a flow of constant product development, driving many publishers to conform to platform initiatives (Kleis Nielsen and Ganter, 2017). We echo the interviewees’ demands that news organizations and platforms collaborate as equal partners and assign trusted employees the responsibility to serve as liaisons to allow journalists to reach out to platforms when they are threatened. Such channels might convince journalists to resort less to massive deletion and preserve more of their online work for posterity.